JP2014128835A - Controller for power transmission device - Google Patents

Abstract

PROBLEM TO BE SOLVED: To control a power imparted to a secondary side element with high robustness, in a power transmission device provided with an elastic force generating mechanism capable of generating an elastic force and a viscous force between a primary side element and the secondary side element.SOLUTION: A secondary side power imparted to a secondary side element 3 is measured by using a measured value of a relative displacement between a primary side element 2 and the secondary side element 3 and a rigidity characteristic coefficient Ksp of an elastic force generating mechanism 4, and a control input is determined by sliding mode control so that the measured value converges on a target value. The inclination of a changeover hyperplane for sliding mode control is set variably according to a control value for the viscosity characteristic coefficient Kdp of the elastic force generating mechanism 4.

Description

  The present invention relates to a control device for a power transmission device for driving a joint or the like of a robot.

  In a power transmission device such as a joint of an industrial robot, position control (servo control) for controlling a displacement amount (rotation angle etc.) of a driven side element (secondary side element) to a target value is generally performed conventionally. Further, for example, Patent Document 1 proposes a technique in which sliding mode control is adopted as a control method for the position control and a switching function used for the sliding mode control is determined in the form of proportional / integral control. Has been.

JP-A-5-134758

  The position control of the driven-side element in the power transmission device is poor in flexibility in various external environments where it is difficult to specify or predict the position and shape of external objects or disturbances in advance. For example, when the driven element comes into contact with an unexpected external object, it becomes difficult to move the driven element appropriately, or an excessive external force acts on the driven element. Cheap.

  For this reason, in recent years, in order to realize a robot or the like that can operate flexibly in various external environments, it has been connected to the power transmission path of the power transmission device via an elastic force generation mechanism such as a spring member. Elasticity generated by an elastic force generating mechanism according to relative displacement between the primary side element and the secondary side element, provided with a primary side element (drive side element) and a secondary side element (driven side element) The inventors of the present application have studied that the power applied to the secondary side element by the force is controlled to the target value.

  In the operation control (power control) of the power transmission device having the structure in which the primary side element and the secondary side element are connected via the elastic force generation mechanism in this way, in a general control method such as PD control, In general, it is difficult to perform stable control that is unlikely to cause oscillation of the control system with respect to various condition fluctuations such as inertia fluctuations.

  For this reason, the inventor of the present application has attempted to adopt a sliding mode control method having a characteristic of high robustness against disturbance fluctuations or the like in the operation control of the power transmission device having an elastically deformable member. .

  This sliding mode control is a control method for trying to converge the state quantity of the controlled object to the target value on the switching hyperplane (hyperplane expressed in the form of switching function = 0) defined by the switching function. .

  The “hyperplane” is a generalized expression of a plane in a multi-dimensional phase space, and means a straight line in a two-dimensional phase space and a normal plane in a three-dimensional phase space.

  By the way, as described above, when the power transmission device including the elastic force generation mechanism is employed in a joint of a robot that can be operated in various environments, the elastic force generation mechanism includes the primary element and the secondary element. Viscosity characteristic coefficient (so-called viscosity coefficient) representing the ratio of the change in the viscous force to the change in the relative speed, while being able to generate a viscous force according to the relative speed between the two elements between the side elements It is considered desirable to be able to set variably.

  In this case, in order to control the power applied to the secondary side element to the target value by the sliding mode control, it is necessary to sequentially observe the power value and its temporal change rate.

  Here, the power applied to the secondary element basically includes a component caused by the elastic force generated by the elastic force generation mechanism (hereinafter referred to as an elastic force component) and a viscosity generated by the elastic force generation mechanism. It is the sum total of components due to force (hereinafter referred to as viscous force components). The elastic force component depends on the relative displacement amount between the primary side element and the secondary side element, and the viscous force component is the relative velocity between the primary side element and the secondary side element (temporal relative displacement amount). Change rate).

  Therefore, the value of the power applied to the secondary side element is determined by the elastic force component specified from the measured value of the relative displacement between the primary side element and the secondary side element, and the time of the measured value of the relative displacement. It is conceivable to measure the sum of the viscous force components specified from the dynamic change rate (differential value) and to calculate the temporal change rate of the measured power value as the measured value of the temporal change rate of the power. .

  However, the temporal change rate of the measured value of the relative displacement amount between the primary side element and the secondary side element is generally the temporal change rate of the actual relative displacement amount (that is, the actual relative speed between the two elements). Errors are likely to be included, and variations in the errors are likely to occur.

  For this reason, it is generally difficult to accurately specify the value of the viscous force component from the temporal change rate of the measured value of the relative displacement amount between the primary side element and the secondary side element. Furthermore, since it is difficult to specify the value of the viscous force component with high accuracy in this way, it is difficult to accurately acquire the measurement value of the temporal change rate of the power applied to the secondary side element. Become.

  Therefore, even if it is attempted to sequentially determine the control input for controlling the power applied to the secondary side element to the target value by the sliding mode control, the measurement value (especially the secondary value) necessary for determining the control input is determined. It is difficult to accurately obtain a measurement value of a temporal change rate of power applied to the side element. As a result, the control input determined by the sliding mode control frequently becomes inappropriate for controlling the actual power applied to the secondary side element, and the robustness of the control is increased. May not be sufficiently increased.

  The present invention has been made in view of such a background. In a power transmission device including an elastic force generation mechanism capable of generating a viscous force in addition to an elastic force between a primary side element and a secondary side element, An object of the present invention is to provide a control device capable of controlling the power applied to the secondary element with high robustness.

In order to achieve the above object, the control device for a power transmission device according to the present invention generates a primary side element that is displaced by a driving force of an actuator and generates an elastic force in the primary side element so as to be relatively displaceable with respect to the primary side element. A secondary element connected through a mechanism, and by an elastic force generated by the elastic force generation mechanism between the primary element and the secondary element according to the relative displacement between the primary element and the secondary element, In the power transmission device configured to transmit power between the two elements, the control device is configured to control the secondary power, which is the power applied to the secondary element by the power transmission, to a target value. And
The elastic force generation mechanism generates a viscous force according to a relative speed between the primary side element and the secondary side element, and a ratio of a change in the viscous force to a change in the relative speed between the two elements. The viscosity characteristic coefficient that represents is variably controllable,
The relative displacement amount between the primary side element and the secondary side element is measured, the measured value of the relative displacement amount, and the ratio of the change in the elastic force generated by the elastic force generation mechanism to the change in the relative displacement amount A secondary-side power measuring means for obtaining a measured value of the secondary-side power applied to the secondary-side element by the generated elastic force from the value of the stiffness characteristic coefficient representing
By control processing of sliding mode control using a switching function configured with a deviation between the measured value of the secondary power and the target value as a first variable component, and a temporal change rate of the deviation as a second variable component, Control input determining means for sequentially determining a control input for controlling the driving force of the actuator so that the first variable component converges to zero on a switching hyperplane defined by a switching function;
A switching hyperplane variable setting for setting the inclination of the switching hyperplane in a phase plane having the first variable component and the second variable component as two coordinate axis components according to the control value of the viscosity characteristic coefficient Means and
The control input determining means is configured to determine the control input using the switching function corresponding to the switching hyperplane having the set inclination (first invention).

  According to the first aspect of the invention, the measurement value acquired by the secondary power measurement unit is a measurement value of the secondary power applied to the secondary element by the elastic force generated by the elastic force generation mechanism. Therefore, the measurement value does not include a component due to the viscous force generated by the elastic force generation mechanism. And since the measured value of the secondary power is obtained using the measured value of the relative displacement amount between the primary side element and the secondary side element, the generated elasticity of the elastic force generating mechanism is obtained. The reliability as the actual value of the secondary power imparted to the secondary element by the force becomes high.

  In the first invention, since the measured value of the secondary side power does not include a component due to the viscous force generated by the elastic force generating mechanism as described above, the elastic force generating mechanism is not limited to the primary side element and the secondary side. In a situation where a viscous force is generated according to the relative speed between elements, the measured value of the secondary power has an error with respect to the actual value of the secondary power.

  However, in the first invention, the inclination of the switching hyperplane is variably set according to the control value of the viscosity characteristic coefficient of the elastic force generation mechanism. For this reason, even if the measured value of the secondary power has a transient error with respect to the actual value due to the influence of the viscous force generated by the elastic force generating mechanism, the control input determining means The control input can be determined so as to compensate for the influence of the error.

  That is, the first variable component is converged to zero with a stable time constant while preventing the first variable component and the second variable component from fluctuating excessively or oscillating. The power measurement value is converged to the target value).

  Therefore, according to 1st invention, the motive power provided to a secondary side element can be controlled by high robustness.

  Note that the control value of the viscosity characteristic coefficient in the first invention means a value realized by controlling the viscosity characteristic coefficient, and for example, a target value of the viscosity characteristic coefficient can be used. The same applies to other inventions described later.

  In the first aspect of the invention, the elastic force generation mechanism is configured such that, for example, the stiffness characteristic coefficient is a constant value. In this case, the switching hyperplane variable setting means converges on the switching hyperplane defined by the inclination of the switching hyperplane as a time constant of convergence of the first variable component to zero on the switching hyperplane. The inclination of the switching hyperplane is determined according to the control value of the viscosity characteristic coefficient so that the correlation between the time constant and the viscosity characteristic coefficient is represented by the following equation (A). It is preferable that it is comprised (2nd invention).

Tc = a + b · Kdp / Ksp (A)
However,
Tc: convergence time constant on the switching hyperplane a, b: predetermined constant value Ksp: value of the stiffness characteristic coefficient Kdp: value of the viscosity characteristic coefficient

That is, according to various experiments and examinations of the present inventors, when the value of the stiffness characteristic coefficient of the elastic force generation mechanism is constant, the convergence time constant on the switching hyperplane defined by the inclination of the switching hyperplane, By determining the inclination of the switching hyperplane according to the control value of the viscosity characteristic coefficient so that the correlation between the viscosity characteristic coefficient is represented by the equation (A), the secondary power The measured value can be suitably converged to the target value.

  In the first invention, the elastic force generation mechanism may be configured to variably control the viscosity characteristic coefficient and the rigidity characteristic coefficient. In this case, the switching hyperplane variable setting means includes a switching hyperplane defined by the inclination of the switching hyperplane as a time constant for convergence of the first variable component to zero on the switching hyperplane. The control value of the viscosity characteristic coefficient and the stiffness characteristic coefficient so that the correlation between the upper convergence time constant, the viscosity characteristic coefficient, and the stiffness characteristic coefficient is represented by the following equation (B): It is preferable that the inclination of the switching hyperplane is determined according to the control value (third invention).

Tc = (a2 / sqrt (Ksp)) + a1 * Ksp + a0 + b * Kdp / Ksp (B)
However,
Tc: convergence time constant on the switching hyperplane a2, a1, a0, b: constant values determined in advance Ksp: value of the stiffness characteristic coefficient Kdp: value of the viscosity characteristic coefficient

That is, according to various experiments and examinations by the inventors of the present application, when the viscosity characteristic coefficient and the rigidity characteristic coefficient of the elastic force generation mechanism are variably controlled, the switching hyperplane defined by the inclination of the switching hyperplane. The control value of the viscosity characteristic coefficient and the stiffness viscosity coefficient are set such that the correlation between the upper convergence time constant, the viscosity characteristic coefficient, and the stiffness characteristic coefficient is represented by the equation (B). By determining the inclination of the switching hyperplane according to the control value, it is possible to suitably converge the measured value of the secondary power to the target value.

  The control value of the stiffness characteristic coefficient in the third invention means a value realized by controlling the stiffness characteristic coefficient, and for example, a target value of the stiffness characteristic coefficient can be used. The same applies to other inventions described later.

  In the second invention or the third invention, more specifically, the control input determining means is, for example, the following expression as a control input component having a function of converging the first variable component to zero on the switching hyperplane: The first control input component calculated in (C) is combined with the second control input component determined according to the value of the switching function as a control input component having a function of converging the value of the switching function to zero. The value formed as described above is determined as the control input (fourth invention).

  The power transmission device according to the present invention is a linear motion type power transmission device that transmits a translational force between the primary side element and the secondary side element (the relative displacement between the primary side element and the secondary side element is A linear transmission power transmission device) and a rotary power transmission device that transmits rotational force (torque) between the primary and secondary elements (between the primary and secondary elements) The relative displacement of the power transmission device may be a rotational displacement. The Iin is the inertial mass on the input side in the direct acting power transmission device, and the inertia on the input side in the rotary power transmission device. In addition, the above-mentioned Iout is an output side inertial mass in the direct acting type power transmission device, and is an output side inertia in the rotary type power transmission device. The same applies to the 14th invention described later.

  According to the fourth aspect of the invention, after the set of the first variable component and the second variable component reaches the switching hyperplane, the first variable component is stably reduced to zero by the first control input component. It can be converged.

In the first invention, when the elastic force generation mechanism is configured such that the stiffness characteristic coefficient is a constant value, the switching hyperplane variable setting means is configured to determine the inclination of the switching hyperplane. A correlation between the viscosity characteristic coefficient values, a plurality of representative values of the viscosity characteristic coefficient, and a slope of the switching hyperplane determined in advance corresponding to each representative value of the viscosity characteristic coefficient According to the correlation specified in advance, and is configured to determine the inclination of the switching hyperplane according to the control value of the viscosity characteristic coefficient,
The inclination of the switching hyperplane corresponding to each representative value of the viscosity characteristic coefficient is such that the first variable component is changed from an arbitrary value to zero while the viscosity characteristic coefficient of the elastic force generation mechanism is controlled to the representative value. Specific response characteristics satisfying a predetermined requirement among a plurality of response characteristic data acquired in advance as data indicating a locus of transition of a set of the value of the first variable component and the value of the second variable component at the time of convergence Based on data,
A line indicating, on the phase plane, a first set value set according to the first allowable limit value between a first allowable limit value which is an allowable limit value of the magnitude of the value of the first variable component and zero. , A first set value line, a second set value set according to the second allowable limit value between zero and a second allowable limit value which is an allowable limit value of the value of the second variable component Is defined as a second set value line, the specific response characteristic data satisfying the predetermined requirement is the first variable component on the locus indicated by the specific response characteristic data. And the value of the second variable component fall within the first allowable limit value and the second allowable limit value, respectively, and the locus is the first set value line or the second set value. Response characteristic data that satisfies the requirement to cross the value line It is in,
The inclination of the switching hyperplane corresponding to each representative value of the viscosity characteristic coefficient is the locus indicated by each of the specific response characteristic data and the first set value line or the second set value line in the phase plane. The constraint that the time constant of convergence of the first variable component to zero on the line connecting the intersection and the origin of the phase plane is a value within a predetermined required range among It is preferable that it is determined so as to match or approximate the inclination of a line connecting the intersecting point and the origin of the phase plane (fifth invention).

  According to the fifth aspect, the correlation between the inclination of the switching hyperplane and the viscosity characteristic coefficient corresponds to a plurality of representative values of the viscosity characteristic coefficient and the representative values of the viscosity characteristic coefficient. It is specified in advance based on the inclination of the switching hyperplane determined in advance.

  In this case, the inclination of the switching hyperplane corresponding to each representative value of the viscosity characteristic coefficient is the plurality of response characteristic data acquired in advance in a state where the viscosity characteristic coefficient of the elastic force generation mechanism is controlled to the representative value. Of these, it is determined based on specific response characteristic data satisfying the predetermined requirement.

  Accordingly, of the response characteristic data, the magnitude of the value of the first variable component on the locus indicated by the response characteristic data becomes excessively larger than the first allowable limit value, or The response characteristic data such that the magnitude of the value of the second variable component exceeds the second allowable limit value is the switching hyperplane corresponding to each representative value of the viscosity characteristic coefficient. It is not used as specific response characteristic data for determining the inclination.

  In the response characteristic data, the value of the first variable component on the locus indicated by the response characteristic data is kept smaller than the first set value, and the Response characteristic data in which the magnitude of the value of the second variable component on the locus is kept smaller than the second set value, that is, the first variable component at an arbitrary point on the locus The response characteristic data in which the values of the second variable component are both too small are specific response characteristic data for determining the inclination of the switching hyperplane corresponding to each representative value of the viscosity characteristic coefficient. Not used.

  That is, the magnitude of the value of the first variable component at the point on the trajectory exceeds the first allowable limit value, or the value of the second variable component at the point on the trajectory. The size of the first variable component does not become excessively large beyond the second allowable limit value, and the first variable component value is greater than or equal to the first set value (a size that is not too small). Or only the response characteristic data including a point on the trajectory where the magnitude of the value of the second variable component is greater than or equal to the second set value (a magnitude that is not too small). Are used as specific response characteristic data for determining the inclination of the switching hyperplane corresponding to each of the representative values.

  Therefore, the response characteristic data having high reliability is obtained as the first variable component can be favorably converged to zero, and the specific response for determining the inclination of the switching hyperplane corresponding to each representative value of the viscosity characteristic coefficient. It can be used as characteristic data.

  The inclination of the switching hyperplane corresponding to each representative value of the viscosity characteristic coefficient is the trajectory indicated by each of the specific response characteristic data and the first set value line or the second set value line in the phase plane. The constraint that the time constant of convergence of the first variable component to zero on the line connecting the intersection and the origin of the phase plane is a value within a predetermined required range among It is determined so as to coincide with or approximate the inclination of a line connecting the intersecting point and the origin of the phase plane.

  In this case, the intersection point is a point at which the value of the first variable component matches the first set value or the value of the second variable component matches the second set value. It is a point that the magnitude of the component value or the magnitude of the value of the second variable component is an appropriate magnitude that is not too large or too small.

  Of these intersections, the time constant for convergence of the first variable component to zero on the line connecting the intersection and the origin of the phase plane is a value within a predetermined required range. Using the intersection satisfying the constraint conditions (hereinafter sometimes referred to as the requirement-matching intersection), each representative of the viscosity characteristic coefficient is matched or approximated to the slope of the line connecting the requirement-matching intersection and the origin of the phase plane. The slope of the switching hyperplane corresponding to the value is determined.

  Thus, since the inclination of the switching hyperplane corresponding to each representative value of the viscosity characteristic coefficient is determined, it is possible to efficiently and easily determine the inclination of the switching hyperplane from a plurality of acquired response characteristic data. be able to.

  In this case, the control method for acquiring the response characteristic data is not limited to a specific control method, and can be easily acquired using a general-purpose control method such as PD control.

  Furthermore, the inclination of the switching hyperplane corresponding to each representative value of the viscosity characteristic coefficient can be determined by using the required fitting intersection, for example, by a known statistical identification method such as a least square method.

  In the fifth aspect of the invention, based on the plurality of representative values of the viscosity characteristic coefficient and the inclination of the switching hyperplane determined as described above corresponding to each of the representative values, the inclination of the switching hyperplane and the viscosity The correlation with the characteristic coefficient (correlation used by the control input determining means to determine the inclination of the switching hyperplane according to the control value of the viscosity characteristic coefficient) is specified.

  For this reason, the control input determination means performs the control input by the control process of the sliding mode control so that the measured value of the secondary power can be controlled to the target value with high robustness and desired convergence characteristics. Can be determined.

  The fifth invention may be combined with the second invention or the fourth invention. When the fifth invention and the second invention are combined, the correlation is expressed by the formula (A). In this case, based on a plurality of representative values of the viscosity characteristic coefficient and the inclination of the switching hyperplane determined as described above corresponding to each of the representative values, by a statistical identification method such as the least square method, The correlation is determined by specifying the values of the variables a and b in the formula (A).

  However, the correlation in the fifth invention is determined, for example, by using a plurality of representative values of the viscosity characteristic coefficient and the inclination of the switching hyperplane determined as described above corresponding to each of the representative values as map data. It may be a correlation.

  In the fifth invention, the first permissible limit value is set so that the displacement speed and displacement acceleration of the primary side element due to the driving force of the actuator do not exceed predetermined permissible limit values, respectively. It is preferable that the value is set according to the allowable limit value of the displacement speed, the allowable limit value of the displacement acceleration of the primary side element, and the value of the stiffness characteristic coefficient (the sixth invention).

  According to the sixth aspect, for each representative value of the viscosity characteristic coefficient, the displacement speed and displacement acceleration of the primary element out of the plurality of response characteristic data exceed a predetermined allowable limit value. By using such response characteristic data, the inclination of the switching hyperplane corresponding to each representative value of the viscosity characteristic coefficient can be determined.

  Therefore, during the operation control of the power transmission device, the operation of the power transmission device is controlled so that the displacement speed and the displacement acceleration of the primary side element do not exceed predetermined allowable limit values, respectively. Regardless of the control state of the viscosity characteristic coefficient of the generating mechanism, it can be performed with high reliability. As a result, regardless of the control state of the viscosity characteristic coefficient of the elastic force generation mechanism, the operation of the power transmission device is stabilized in a region where the displacement speed or displacement acceleration of the primary element is close to the allowable limit value. It is possible to do so.

  In the fifth or sixth aspect of the invention, the second allowable limit value prevents occurrence of vibration of the power transmission system according to the natural vibration of the power transmission system from the actuator to the secondary element. Thus, it is preferable that the value is set in accordance with the value of the rigidity characteristic coefficient (seventh invention).

  According to the seventh aspect of the present invention, for each representative value of the viscosity characteristic coefficient, out of the plurality of response characteristic data, using response characteristic data that does not generate vibration of the power transmission system according to the natural vibration, The inclination of the switching hyperplane corresponding to each representative value of the viscosity characteristic coefficient can be determined.

  For this reason, the power transmission device is operated so that the vibration of the power transmission system according to the natural vibration does not occur or hardly occurs regardless of the control state of the viscosity characteristic coefficient of the elastic force generation mechanism. Control can be performed with high reliability.

In the first invention, when the elastic force generation mechanism is configured to be able to variably control the viscosity characteristic coefficient and the rigidity characteristic coefficient, the switching hyperplane variable setting means includes: A correlation between the inclination of the switching hyperplane, the viscosity characteristic coefficient, and the stiffness characteristic coefficient, wherein a plurality of representative values of the viscosity characteristic coefficient, a plurality of representative values of the stiffness characteristic coefficient, A control value of the viscosity characteristic coefficient according to a correlation determined in advance based on a slope of the switching hyperplane determined in advance corresponding to each of the representative value sets of the viscosity characteristic coefficient and the rigidity characteristic coefficient; The inclination of the switching hyperplane is determined according to the control value of the stiffness characteristic coefficient,
The inclination of the switching hyperplane corresponding to each set of representative values of the viscosity characteristic coefficient and the rigidity characteristic coefficient is in a state where the viscosity characteristic coefficient and the rigidity characteristic coefficient of the elastic force generation mechanism are controlled to the representative values, respectively. , A plurality of response characteristics acquired in advance as data indicating a locus of transition of a set of the value of the first variable component and the value of the second variable component when the first variable component is converged from an arbitrary value to zero Determined based on specific response characteristics data that meets certain requirements of the data,
A line indicating, on the phase plane, a first set value set according to the first allowable limit value between a first allowable limit value which is an allowable limit value of the magnitude of the value of the first variable component and zero. , A first set value line, a second set value set according to the second allowable limit value between zero and a second allowable limit value which is an allowable limit value of the value of the second variable component Is defined as a second set value line, the specific response characteristic data satisfying the predetermined requirement is the first variable component on the locus indicated by the specific response characteristic data. And the value of the second variable component fall within the first allowable limit value and the second allowable limit value, respectively, and the locus is the first set value line or the second set value. Response characteristic data that satisfies the requirement to cross the value line It is in,
The inclination of the switching hyperplane corresponding to each set of representative values of the viscosity characteristic coefficient and the rigidity characteristic coefficient is the trajectory indicated by each of the specific response characteristic data and the first setting on the phase plane. Among the intersections with the value line or the second set value line, the time constant for convergence of the first variable component to zero on the line connecting the intersection and the origin of the phase plane is within a predetermined required range. Preferably, it is determined so as to coincide with or approximate the inclination of a line connecting the intersection satisfying the constraint condition of the value and the origin of the phase plane (eighth invention).

  According to the eighth aspect of the present invention, the correlation among the inclination of the switching hyperplane, the viscosity characteristic coefficient, and the rigidity characteristic coefficient includes a plurality of representative values of the viscosity characteristic coefficient, and a plurality of rigidity characteristic coefficients. And a slope of the switching hyperplane determined in advance corresponding to each set of representative values of the viscosity characteristic coefficient and the rigidity characteristic coefficient.

  In this case, the inclination of the switching hyperplane corresponding to each set of representative values of the viscosity characteristic coefficient and the rigidity characteristic coefficient is determined by changing the viscosity characteristic coefficient and the rigidity characteristic coefficient of the elastic force generation mechanism to the representative value (arbitrary value) In this case, it is determined based on specific response characteristic data satisfying the predetermined requirement among the plurality of response characteristic data acquired in advance in a controlled state.

  Accordingly, as in the case of the fifth invention, the magnitude of the value of the first variable component on the locus indicated by the response characteristic data exceeds the first allowable limit value in the response characteristic data. The response characteristic data that is excessively large such that the magnitude of the value of the second variable component exceeds the second allowable limit value is the viscosity characteristic coefficient and It is not used as specific response characteristic data for determining the inclination of the switching hyperplane corresponding to each set of representative values of the stiffness characteristic coefficient.

  In the response characteristic data, the value of the first variable component on the locus indicated by the response characteristic data is kept smaller than the first set value, and the Response characteristic data in which the magnitude of the value of the second variable component on the locus is kept smaller than the second set value, that is, the first variable component at an arbitrary point on the locus The response characteristic data whose magnitudes of the second variable component and the second variable component are both too small determine the inclination of the switching hyperplane corresponding to each of the representative value pairs of the viscosity characteristic coefficient and the rigidity characteristic coefficient. It is not used as specific response characteristic data.

  That is, as in the case of the fifth invention, the magnitude of the value of the first variable component at the point on the trajectory exceeds the first allowable limit value, or the trajectory The magnitude of the value of the second variable component at the upper point does not become excessive so as to exceed the second allowable limit value, and the magnitude of the value of the first variable component is not less than the first set value. Or a point on the locus where the value of the second variable component is not less than the second set value (not too small). Only the response characteristic data that is included is used as specific response characteristic data for determining the inclination of the switching hyperplane corresponding to each of the sets of representative values of the viscosity characteristic coefficient and the stiffness characteristic coefficient.

  Accordingly, since the first variable component can be favorably converged to zero, highly reliable response characteristic data is obtained from the inclination of the switching hyperplane corresponding to each of the representative values of the viscosity characteristic coefficient and the rigidity characteristic coefficient. Can be used as specific response characteristic data for determining.

  The inclination of the switching hyperplane corresponding to each of the representative value pairs of the viscosity characteristic coefficient and the rigidity characteristic coefficient is the trajectory indicated by each of the specific response characteristic data and the first setting on the phase plane. Among the intersections with the value line or the second set value line, the time constant for convergence of the first variable component to zero on the line connecting the intersection and the origin of the phase plane is within a predetermined required range. It is determined so as to match or approximate the inclination of a line connecting the intersection satisfying the constraint condition that is the value of and the origin of the phase plane.

  In this case, the intersection point is a point at which the value of the first variable component matches the first set value or the value of the second variable component matches the second set value. It is a point that the magnitude of the component value or the magnitude of the value of the second variable component is an appropriate magnitude that is not too large or too small.

  Of these intersections, the time constant for convergence of the first variable component to zero on the line connecting the intersection and the origin of the phase plane is a value within a predetermined required range. Using the intersection satisfying the constraint condition (that is, the requirement adaptation intersection), each of the viscosity characteristic coefficient and the stiffness characteristic coefficient is matched or approximated to the slope of the line connecting the requirement adaptation intersection and the origin of the phase plane. The slope of the switching hyperplane corresponding to each of the representative value sets is determined.

  In this way, since the inclination of the switching hyperplane corresponding to each of the representative values of the viscosity characteristic coefficient and the stiffness characteristic coefficient is determined, the inclination of the switching hyperplane is determined from a plurality of acquired response characteristic data. Can be done efficiently and easily.

  In this case, the control method for acquiring the response characteristic data is not limited to a specific control method, and can be easily acquired using a general-purpose control method such as PD control.

  Further, determining the inclination of the switching hyperplane corresponding to each of the representative value pairs of the viscosity characteristic coefficient and the stiffness characteristic coefficient using the required fitting intersection is, for example, a known statistical method such as a least square method. This can be done by an identification technique.

  In the eighth invention, the inclination of the switching hyperplane determined as described above corresponding to each of a plurality of representative values of the viscosity characteristic coefficient, a plurality of representative values of the stiffness characteristic coefficient, and a set of these representative values. Based on the above, the correlation between the inclination of the switching hyperplane, the viscosity characteristic coefficient, and the stiffness characteristic coefficient (the control input determining means depends on the control value of the viscosity characteristic coefficient and the control value of the stiffness characteristic coefficient). The correlation used to determine the slope of the switching hyperplane.

  For this reason, the control input determination means performs the control input by the control process of the sliding mode control so that the measured value of the secondary power can be controlled to the target value with high robustness and desired convergence characteristics. Can be determined.

  The eighth invention may be combined with the third invention or the fourth invention. When the eighth invention and the third invention are combined, the correlation is represented by the formula (B). In this case, statistical identification such as the least square method is performed based on a plurality of representative values of the stiffness characteristic coefficient and the inclination of the switching hyperplane determined as described above corresponding to each of the representative value sets. The correlation is determined by specifying the values of the variables a2, a1, a0, b in the formula (B) by the method.

  However, the correlation in the eighth invention is determined as described above, for example, corresponding to each of a plurality of representative values of the viscosity characteristic coefficient, a plurality of representative values of the rigidity characteristic coefficient, and a set of these representative values. The correlation determined by using the inclination of the switching hyperplane as map data may be used.

  In the eighth aspect of the invention, the first allowable limit value is a value set corresponding to each representative value of the rigidity characteristic coefficient, and the first allowable limit value corresponding to each representative value of the rigidity characteristic coefficient. In a state where the stiffness characteristic coefficient of the elastic force generation mechanism is controlled to the representative value, the displacement speed and displacement acceleration of the primary element due to the driving force of the actuator do not exceed predetermined allowable limit values, respectively. Further, it is preferable that the value is set according to the allowable limit value of the displacement speed of the primary side element, the allowable limit value of the displacement acceleration of the primary side element, and the representative value of the stiffness characteristic coefficient (the ninth). invention).

  According to the ninth aspect of the present invention, the displacement speed and the displacement acceleration of the primary side element of the plurality of response characteristic data are respectively determined in advance for each set of representative values of the viscosity characteristic coefficient and the rigidity characteristic coefficient. The inclination of the switching hyperplane can be determined using response characteristic data that does not exceed the allowable limit value.

  Therefore, during the operation control of the power transmission device, the operation of the power transmission device is controlled so that the displacement speed and the displacement acceleration of the primary side element do not exceed predetermined allowable limit values, respectively. Regardless of the control state of the viscosity characteristic coefficient and the rigidity characteristic coefficient of the generating mechanism, it can be performed with high reliability. As a result, the power transmission device in a region where the displacement speed or displacement acceleration of the primary side element is close to the allowable limit value regardless of the control state of the viscosity characteristic coefficient and the rigidity characteristic coefficient of the elastic force generation mechanism. It is possible to perform the operation in a stable manner.

  In the eighth or ninth invention, the second allowable limit value is a value set corresponding to each representative value of the stiffness characteristic coefficient, and corresponds to each representative value of the stiffness characteristic coefficient. The second permissible limit value is the power transmission system corresponding to the natural vibration of the power transmission system from the actuator to the secondary element in a state where the stiffness characteristic coefficient of the elastic force generation mechanism is controlled to the representative value. It is preferable that the value is determined in accordance with the representative value of the stiffness characteristic coefficient so as to prevent the occurrence of vibration (tenth invention).

  According to the tenth aspect, vibration of the power transmission system corresponding to the natural vibration is generated among the plurality of response characteristic data for each set of representative values of the viscosity characteristic coefficient and the rigidity characteristic coefficient. By using such response characteristic data, the inclination of the switching hyperplane corresponding to each set of representative values of the viscosity characteristic coefficient and the rigidity characteristic coefficient can be determined.

  Therefore, regardless of the control state of the viscosity characteristic coefficient and the rigidity characteristic coefficient of the elastic force generation mechanism, the power transmission system is designed so that the vibration of the power transmission system according to the natural vibration does not occur or hardly occurs. It is possible to control the operation of the apparatus with high reliability.

  In the fifth to tenth inventions, the value of the first variable component is changed stepwise from zero to the first allowable limit value, and the convergence of the value of the first variable component to zero is controlled by the actuator. When the time constant realized when it is assumed that the displacement acceleration of the primary side element due to the driving force is set to the allowable limit value of the displacement acceleration is defined as the specific time constant, the constraint condition is the first condition. It is preferable that the time constant for the convergence of the variable component to zero is a condition that limits the time constant to a value in the range equal to or greater than the specific time constant (11th invention).

  That is, when it is assumed that the value of the first variable component is changed stepwise from zero to a certain value not more than the first allowable limit value, the value of the first variable component is changed from zero to the first allowable limit value. In this case, the relative displacement amount between the primary side element and the secondary side element is maximized.

  In this case, since the displacement of the primary side element with respect to the secondary side element is limited to the displacement at the displacement acceleration equal to or less than the allowable limit value, the time constant for convergence of the value of the first variable component to zero is the specified time. It cannot be smaller than a constant.

  Accordingly, in the eleventh aspect of the invention, the constraint condition that defines the requirement-matching intersection is a condition that limits the time constant of convergence of the first variable component to zero to a value in a range greater than or equal to the specific time constant. Set.

  Therefore, according to the eleventh aspect, at the time of controlling the operation of the power transmission device, it is possible to stably control the secondary power to the target value in a wide range of operation of the power transmission device with high robustness.

  In the first to eleventh aspects of the invention, a first variable component measurement value that is a value of the first variable component calculated from the measurement value of the secondary power and the target value of the secondary power; An estimated value of the first variable component obtained by reducing the influence of disturbance from a second variable component measured value that is a value of the second variable component calculated as a rate of temporal change of the first variable component measured value; The observer further includes an observer that sequentially calculates the estimated value of the second variable component, and the control input determining means uses the observer instead of the first variable component measured value and the second variable component measured value. While calculating the value of the switching function using the calculated estimated value of the first variable component and the estimated value of the second variable component, the control input is sequentially generated using the value of the switching function. It is preferable to be configured (the twelfth invention).

  According to this twelfth aspect, the value of the switching function is calculated while calculating the value of the switching function using the estimated value of the first variable component and the estimated value of the second variable component calculated by the observer. By using the control input of the sliding mode control to sequentially generate the control input, the influence of disturbance is suppressed.

  For this reason, the stability of the control input sequentially determined by the control input determining means is increased. As a result, the robustness of the control of the secondary power can be further enhanced.

  The control device for the power transmission device of the present invention is not limited to the first to eleventh aspects, and may employ the following aspects.

That is, in order to achieve the above object, the control device for a power transmission device according to the present invention includes a primary side element that is displaced by a driving force of an actuator, and a primary side element that is displaceable relative to the primary side element. Elasticity generated by the elastic force generation mechanism between the primary side element and the secondary side element according to relative displacement between the primary side element and the secondary side element. In a power transmission device configured to transmit power between both elements by force, control for controlling secondary power, which is power applied to the secondary element by the power transmission, to a target value A device,
The elastic force generation mechanism generates a viscous force according to a relative speed between the primary side element and the secondary side element, and a ratio of a change in the viscous force to a change in the relative speed between the two elements. The viscosity characteristic coefficient that represents is variably controllable,
The relative displacement amount between the primary side element and the secondary side element is measured, the measured value of the relative displacement amount, and the ratio of the change in the elastic force generated by the elastic force generation mechanism to the change in the relative displacement amount A secondary-side power measuring means for obtaining a measured value of the secondary-side power applied to the secondary-side element by the generated elastic force from the value of the stiffness characteristic coefficient representing
By control processing of sliding mode control using a switching function configured with a deviation between the measured value of the secondary power and the target value as a first variable component, and a temporal change rate of the deviation as a second variable component, Control input determining means for sequentially determining a control input for controlling the driving force of the actuator so that the first variable component converges to zero on a switching hyperplane defined by a switching function;
The control input determination means is a phase plane having a correction value obtained by correcting the control value of the viscosity characteristic coefficient with a predetermined correction coefficient, and the first variable component and the second variable component as two coordinate axis components. The control input is determined using the switching function determined so that the inclination of the switching hyperplane becomes an inclination independent of a change in the control value of the viscosity characteristic coefficient. (13th invention).

  According to the thirteenth aspect, the measurement value acquired by the secondary-side power measuring means is applied to the secondary-side element by the elastic force generated by the elastic force generating mechanism, as in the first invention. Since it is the measured value of the secondary power, the measured value does not include a component due to the viscous force generated by the elastic force generating mechanism. And since the measured value of the secondary power is obtained using the measured value of the relative displacement amount between the primary side element and the secondary side element, the generated elasticity of the elastic force generating mechanism is obtained. The reliability as the actual value of the secondary power imparted to the secondary element by the force becomes high.

  In the thirteenth aspect, as described above, the measured value of the secondary side power does not include a component due to the viscous force generated by the elastic force generation mechanism. In a situation where viscous force is generated according to the relative speed between the primary side element and the secondary side element, the measured value of the secondary side power has an error with respect to the actual value of the secondary side power. Become.

  Here, in the thirteenth aspect, the inclination of the switching hyperplane defined by the switching function used by the control input determining means to determine the control input does not depend on a change in the control value of the viscosity characteristic coefficient. On the other hand, a correction value obtained by correcting the control value of the viscosity characteristic coefficient with a predetermined correction coefficient is used to determine the control input.

  For this reason, even if the measured value of the secondary power has a transient error with respect to the actual value due to the influence of the viscous force generated by the elastic force generating mechanism, the control input determining means The control input can be determined so as to compensate for the influence of the error.

  That is, the first variable component is converged to zero with a stable time constant while preventing the first variable component and the second variable component from fluctuating excessively or oscillating. The power measurement value is converged to the target value).

  Therefore, according to the thirteenth aspect, the power applied to the secondary side element can be controlled with high robustness.

  In the thirteenth aspect of the invention, more specifically, the control input determining means is calculated by the following equation (D) as a control input component having a function of converging the first variable component to zero on the switching hyperplane, for example. And a second control input component determined according to the value of the switching function as a control input component having a function of converging the value of the switching function to zero. A control device for a power transmission device, wherein the control input is determined as the control input.

  Note that whether Iin is input side inertia or input side inertial mass, and whether Iout is output side inertia or output side inertial mass is the same as in the case of the fourth invention. is there.

  According to the fourteenth aspect of the present invention, the correction value (= b · Kdp) obtained by correcting the control value of the viscosity characteristic coefficient with a predetermined correction coefficient is appropriately reflected in the first control input component. it can. For this reason, after the set of the first variable component and the second variable component reaches the switching hyperplane, the first variable component can be stably converged to zero by the first control input component. .

In the thirteenth aspect or the fourteenth aspect, the elastic force generation mechanism is configured such that, for example, the stiffness characteristic coefficient is a constant value. In this case, there is a correlation between an appropriate time constant for converging the first variable component to zero and the viscosity characteristic coefficient, and a plurality of representative values of the viscosity characteristic coefficient and the viscosity characteristic In the correlation specified in advance based on the value of the appropriate time constant determined in advance corresponding to each representative value of the coefficient, the value of the appropriate time constant when the value of the viscosity characteristic coefficient is zero The inclination of the switching hyperplane is determined so that the convergence time constant on the switching hyperplane that is the time constant of the convergence of the first variable component to zero on the switching hyperplane matches, In the correlation, the value of the appropriate time constant corresponding to the value of the arbitrary viscosity characteristic coefficient is different from the value of the appropriate time constant when the value of the viscosity characteristic coefficient is zero. The coefficient value is corrected by the correction coefficient. Said correction coefficient to be a function value of becomes the correction value has been determined,
The appropriate time constant value corresponding to each representative value of the viscosity characteristic coefficient converges the first variable component from an arbitrary value to zero with the viscosity characteristic coefficient of the elastic force generation mechanism controlled to the representative value. Specific response characteristic data satisfying a predetermined requirement among a plurality of response characteristic data acquired in advance as data indicating the trajectory of transition of the value of the first variable component and the value of the second variable component at the time of the combination Based on
A line indicating, on the phase plane, a first set value set according to the first allowable limit value between a first allowable limit value which is an allowable limit value of the magnitude of the value of the first variable component and zero. , A first set value line, a second set value set according to the second allowable limit value between zero and a second allowable limit value which is an allowable limit value of the value of the second variable component Is defined as a second set value line, the specific response characteristic data satisfying the predetermined requirement is the first variable component on the locus indicated by the specific response characteristic data. And the value of the second variable component fall within the first allowable limit value and the second allowable limit value, respectively, and the locus is the first set value line or the second set value. Response characteristic data that satisfies the requirement to cross the value line It is in,
The value of the appropriate time constant corresponding to each representative value of the viscosity characteristic coefficient is the trajectory indicated by each of the specific response characteristic data and the first set value line or the second set value line on the phase plane. The constraint that the time constant of convergence of the first variable component to zero on the line connecting the intersection and the origin of the phase plane is a value within a predetermined required range among Preferably, it is determined so as to coincide with or approximate the time constant of convergence of the first variable component to zero on a line connecting the intersection point to be satisfied and the origin of the phase plane (15th invention).

  According to the fifteenth aspect of the present invention, the correlation between the appropriate time constant for converging the first variable component to zero and the viscosity characteristic coefficient includes a plurality of representative values of the viscosity characteristic coefficient and the viscosity. It is specified in advance based on the value of the appropriate time constant determined in advance corresponding to each representative value of the characteristic coefficient.

  In this case, the value of the appropriate time constant corresponding to each representative value of the viscosity characteristic coefficient is the plurality of response characteristic data acquired in advance with the viscosity characteristic coefficient of the elastic force generation mechanism controlled to the representative value. Of these, it is determined based on specific response characteristic data satisfying the predetermined requirement.

  Accordingly, of the response characteristic data, the magnitude of the value of the first variable component on the locus indicated by the response characteristic data becomes excessively larger than the first allowable limit value, or The response characteristic data such that the magnitude of the value of the second variable component exceeds the second allowable limit value is an appropriate time constant corresponding to each representative value of the viscosity characteristic coefficient. It is not used as specific response characteristic data for determining a value.

  In the response characteristic data, the value of the first variable component on the locus indicated by the response characteristic data is kept smaller than the first set value, and the Response characteristic data in which the magnitude of the value of the second variable component on the locus is kept smaller than the second set value, that is, the first variable component at an arbitrary point on the locus And the response characteristic data in which the magnitudes of the values of the second variable component are both too small are specific response characteristic data for determining the value of the appropriate time constant corresponding to each representative value of the viscosity characteristic coefficient. Not used.

  That is, the magnitude of the value of the first variable component at the point on the trajectory exceeds the first allowable limit value, or the value of the second variable component at the point on the trajectory. The size of the first variable component does not become excessively large beyond the second allowable limit value, and the first variable component value is greater than or equal to the first set value (a size that is not too small). Or only the response characteristic data including a point on the trajectory where the magnitude of the value of the second variable component is greater than or equal to the second set value (a magnitude that is not too small). This is used as specific response characteristic data for determining the value of the appropriate time constant corresponding to each representative value.

  Therefore, the response characteristic data having high reliability is obtained as the first variable component can be favorably converged to zero, and the specific response for determining the value of the appropriate time constant corresponding to each representative value of the viscosity characteristic coefficient It can be used as characteristic data.

  Then, the value of the appropriate time constant corresponding to each representative value of the viscosity characteristic coefficient is the trajectory indicated by each of the specific response characteristic data and the first set value line or the second set value line on the phase plane. The constraint that the time constant of convergence of the first variable component to zero on the line connecting the intersection and the origin of the phase plane is a value within a predetermined required range among It is determined so as to coincide with or approximate the time constant of convergence of the first variable component to zero on the line connecting the satisfying intersection and the origin of the phase plane.

  In this case, the intersection point is a point at which the value of the first variable component matches the first set value or the value of the second variable component matches the second set value. It is a point that the magnitude of the component value or the magnitude of the value of the second variable component is an appropriate magnitude that is not too large or too small.

  Of these intersections, the time constant for convergence of the first variable component to zero on the line connecting the intersection and the origin of the phase plane is a value within a predetermined required range. Using the intersection satisfying the constraint condition (that is, the requirement adaptation intersection), the time constant of the convergence of the first variable component to zero on the line connecting the requirement adaptation intersection and the origin of the phase plane or As an approximation, the value of the appropriate time constant corresponding to each representative value of the viscosity characteristic coefficient is determined.

  Since the appropriate time constant value corresponding to each representative value of the viscosity characteristic coefficient is determined in this way, it is possible to easily and efficiently determine the appropriate time constant value from a plurality of acquired response characteristic data. be able to.

  In this case, the control method for acquiring the response characteristic data is not limited to a specific control method, and can be easily acquired using a general-purpose control method such as PD control.

  Furthermore, it is possible to determine a value of an appropriate time constant corresponding to each representative value of the viscosity characteristic coefficient by using the required fitting intersection, for example, by a known statistical identification method such as a least square method.

  In the fifteenth aspect, based on the plurality of representative values of the viscosity characteristic coefficients and the values of the appropriate time constants determined as described above corresponding to the representative values, the appropriate time constant and the viscosity characteristic coefficient The correlation between them has been identified. This correlation corresponds to the correlation between the inclination of the switching hyperplane and the viscosity characteristic coefficient in the fifth invention.

  In the fifteenth invention, in this correlation, the value of the appropriate time constant when the value of the viscosity characteristic coefficient is zero is changed to the convergence of the first variable component to zero on the switching hyperplane. The inclination of the switching hyperplane is determined so that the time constants of convergence on the switching hyperplane coincide with each other.

  In the correlation, the value of the appropriate time constant corresponding to the value of the arbitrary viscosity characteristic coefficient is different from the value of the appropriate time constant when the value of the viscosity characteristic coefficient is zero. The correction coefficient is determined so as to be a function value of a correction value obtained by correcting the value of the viscosity characteristic coefficient by the correction coefficient.

  For this reason, the control input determination means can target the measured value of the secondary power with high robustness and desired convergence characteristics without changing the inclination of the switching hyperplane according to the control value of the viscosity characteristic coefficient. The control input can be determined by the control process of the sliding mode control using a correction value obtained by correcting the control value of the viscosity characteristic coefficient by an appropriate correction coefficient so that the value can be controlled to a value.

  In the fifteenth aspect, the correlation is, for example, a relation represented by the following formula (E) (the sixteenth aspect).

Tc = a + b · Kdp / Ksp (E)
However,
Tc: the appropriate time constant Ksp: the value of the stiffness characteristic coefficient Kdp: the value of the viscosity characteristic coefficient a: a predetermined value determined in advance b: a predetermined value determined in advance as the correction coefficient

That is, according to various experiments and examinations of the present inventors, when the value of the stiffness characteristic coefficient of the elastic force generation mechanism is constant, an appropriate time constant for causing the first variable component to converge to zero, and the viscosity The correlation between the characteristic coefficients can be approximated by equation (E).

  In this case, in the correlation of the equation (E), the value a is the value of the appropriate time constant when the value of the viscosity characteristic coefficient is zero. It is determined as the convergence time constant on the hyperplane.

  Further, in the correlation of the equation (E), the value b is determined as the correction coefficient. That is, in the correlation of the equation (E), the value obtained by multiplying the difference between the value of the arbitrary viscosity characteristic coefficient and the value a by the value of the rigidity characteristic coefficient of the elastic force generation mechanism is the viscosity characteristic coefficient. The value of the correction coefficient b is determined so that the value obtained by correcting the value of the above by the coefficient b (= b · Kdp) matches.

  By determining the inclination of the switching hyperplane and the correction coefficient in this way, it is possible to suitably converge the measured value of the secondary power to the target value.

  In the fifteenth aspect or the sixteenth aspect, the first permissible limit value is set so that the displacement speed and displacement acceleration of the primary side element due to the driving force of the actuator do not exceed predetermined permissible limit values. It is preferably a value determined according to the allowable limit value of the displacement speed of the primary side element, the allowable limit value of the displacement acceleration of the primary side element, and the value of the stiffness characteristic coefficient (17th invention).

  According to the seventeenth aspect, for each representative value of the viscosity characteristic coefficient, among the plurality of response characteristic data, the displacement speed and displacement acceleration of the primary side element may exceed predetermined allowable limit values. By using such response characteristic data, the appropriate time constant value corresponding to each representative value of the viscosity characteristic coefficient can be determined.

  Therefore, during the operation control of the power transmission device, the operation of the power transmission device is controlled so that the displacement speed and the displacement acceleration of the primary side element do not exceed predetermined allowable limit values, respectively. Regardless of the control state of the viscosity characteristic coefficient of the generating mechanism, it can be performed with high reliability. As a result, regardless of the control state of the viscosity characteristic coefficient of the elastic force generation mechanism, the operation of the power transmission device is stabilized in a region where the displacement speed or displacement acceleration of the primary element is close to the allowable limit value. It is possible to do so.

  In the fifteenth to seventeenth aspects of the present invention, the second permissible limit value prevents the occurrence of vibration of the power transmission system in accordance with the natural vibration of the power transmission system from the actuator to the secondary side element. Further, it is preferable that the value is determined according to the value of the stiffness characteristic coefficient (18th invention).

  According to the eighteenth aspect of the present invention, for each representative value of the viscosity characteristic coefficient, among the plurality of response characteristic data, using response characteristic data that does not cause vibration of the power transmission system according to the natural vibration, The appropriate time constant value corresponding to each representative value of the viscosity characteristic coefficient can be determined.

  For this reason, the power transmission device is operated so that the vibration of the power transmission system according to the natural vibration does not occur or hardly occurs regardless of the control state of the viscosity characteristic coefficient of the elastic force generation mechanism. Control can be performed with high reliability.

In the thirteenth or fourteenth aspect, the elastic force generation mechanism is configured to be able to variably control the viscosity characteristic coefficient and the rigidity characteristic coefficient,
Switching hyperplane variable setting means for setting the inclination of the switching hyperplane to change according to the control value of the stiffness characteristic coefficient,
The switching hyperplane variable setting means is a correlation between an appropriate time constant for converging the first variable component to zero, the viscosity characteristic coefficient, and the stiffness characteristic coefficient, wherein the viscosity characteristic coefficient A plurality of representative values, a plurality of representative values of the stiffness characteristic coefficient, and a value of the appropriate time constant determined in advance corresponding to each set of representative values of the viscosity characteristic coefficient and the stiffness characteristic coefficient. Based on the correlation specified in advance, the value of the viscosity characteristic coefficient is set to zero, and the value of the appropriate time constant when the value of the rigidity characteristic coefficient matches the control value of the rigidity characteristic coefficient, Depending on the control value of the stiffness characteristic coefficient, the switching hyperplane has a convergence time constant on the switching hyperplane that is a time constant for convergence of the first variable component to zero on the switching hyperplane. Configured to determine the tilt Cage,
Further, in the correlation, the difference between the value of the appropriate time constant corresponding to the value of the arbitrary viscosity characteristic coefficient and the value of the appropriate time constant when the value of the viscosity characteristic coefficient is zero is the arbitrary time constant. The correction coefficient is determined in advance to be a function value of a correction value obtained by correcting the value of the viscosity characteristic coefficient with the correction coefficient,
The value of the appropriate time constant corresponding to each set of representative values of the viscosity characteristic coefficient and the rigidity characteristic coefficient is the state in which the viscosity characteristic coefficient and the rigidity characteristic coefficient of the elastic force generation mechanism are controlled to the representative values, respectively. , A plurality of response characteristics acquired in advance as data indicating a locus of transition of a set of the value of the first variable component and the value of the second variable component when the first variable component is converged from an arbitrary value to zero Determined based on specific response characteristics data that meets certain requirements of the data,
A line indicating, on the phase plane, a first set value set according to the first allowable limit value between a first allowable limit value which is an allowable limit value of the magnitude of the value of the first variable component and zero. , A first set value line, a second set value set according to the second allowable limit value between zero and a second allowable limit value which is an allowable limit value of the value of the second variable component Is defined as a second set value line, the specific response characteristic data satisfying the predetermined requirement is the first variable component on the locus indicated by the specific response characteristic data. And the value of the second variable component fall within the first allowable limit value and the second allowable limit value, respectively, and the locus is the first set value line or the second set value. Response characteristic data that satisfies the requirement to cross the value line It is in,
The value of the appropriate time constant corresponding to each set of representative values of the viscosity characteristic coefficient and the stiffness characteristic coefficient is the trajectory indicated by each of the specific response characteristic data and the first setting on the phase plane. Among the intersections with the value line or the second set value line, the time constant for convergence of the first variable component to zero on the line connecting the intersection and the origin of the phase plane is within a predetermined required range. It is determined so as to match or approximate the time constant of convergence of the first variable component to zero on a line connecting the intersection satisfying the constraint condition of the value of the value and the origin of the phase plane. Preferred (19th invention).

  According to the nineteenth aspect of the present invention, a correlation between an appropriate time constant for converging the first variable component to zero, the viscosity characteristic coefficient, and the rigidity characteristic coefficient is a plurality of viscosity characteristic coefficients. Based on a representative value, a plurality of representative values of the stiffness characteristic coefficient, and a value of the appropriate time constant determined in advance corresponding to each of the representative value sets of the viscosity characteristic coefficient and the stiffness characteristic coefficient. Have been identified.

  In this case, the values of the appropriate time constants corresponding to the respective sets of the representative values of the viscosity characteristic coefficient and the rigidity characteristic coefficient are controlled so that the viscosity characteristic coefficient and the rigidity characteristic coefficient of the elastic force generation mechanism are respectively set to the representative values. It is determined based on specific response characteristic data satisfying the predetermined requirement among the plurality of response characteristic data acquired in advance.

  Accordingly, of the response characteristic data, the magnitude of the value of the first variable component on the locus indicated by the response characteristic data becomes excessively larger than the first allowable limit value, or The response characteristic data such that the magnitude of the value of the second variable component exceeds the second allowable limit value is a set of representative values of the viscosity characteristic coefficient and the rigidity characteristic coefficient. It is not used as specific response characteristic data for determining the value of the appropriate time constant corresponding to each.

  In the response characteristic data, the value of the first variable component on the locus indicated by the response characteristic data is kept smaller than the first set value, and the Response characteristic data in which the magnitude of the value of the second variable component on the locus is kept smaller than the second set value, that is, the first variable component at an arbitrary point on the locus And response characteristic data in which the magnitudes of the values of the second variable component are both too small are for determining appropriate time constant values corresponding to the respective sets of representative values of the viscosity characteristic coefficient and the rigidity characteristic coefficient. It is not used as specific response characteristic data.

  That is, the magnitude of the value of the first variable component at the point on the trajectory exceeds the first allowable limit value, or the value of the second variable component at the point on the trajectory. The size of the first variable component does not become excessively large beyond the second allowable limit value, and the first variable component value is greater than or equal to the first set value (a size that is not too small). Or only the response characteristic data including a point on the trajectory where the magnitude of the value of the second variable component is greater than or equal to the second set value (a magnitude that is not too small). And appropriate response characteristic data for determining a value of an appropriate time constant corresponding to each set of representative values of stiffness characteristic coefficients.

  Therefore, the response characteristic data with high reliability is obtained as a value that can successfully converge the first variable component to zero, and the value of the appropriate time constant corresponding to each of the representative value sets of the viscosity characteristic coefficient and the rigidity characteristic coefficient. Can be used as specific response characteristic data for determining.

  Then, the values of the appropriate time constants corresponding to the respective representative value sets of the viscosity characteristic coefficient and the stiffness characteristic coefficient are the trajectory indicated by each of the specific response characteristic data and the first setting on the phase plane. Among the intersections with the value line or the second set value line, the time constant for convergence of the first variable component to zero on the line connecting the intersection and the origin of the phase plane is within a predetermined required range. It is determined so as to coincide with or approximate the time constant of convergence of the first variable component to zero on a line connecting the intersection satisfying the constraint condition of the value and the origin of the phase plane.

  In this case, the intersection point is a point at which the value of the first variable component matches the first set value or the value of the second variable component matches the second set value. It is a point that the magnitude of the component value or the magnitude of the value of the second variable component is an appropriate magnitude that is not too large or too small.

  Of these intersections, the time constant for convergence of the first variable component to zero on the line connecting the intersection and the origin of the phase plane is a value within a predetermined required range. Viscosity characteristics so as to coincide with or approximate the time constant of convergence of the first variable component to zero on a line connecting the intersection satisfying the constraint condition (that is, the required matching intersection) and the origin of the phase plane. A value of an appropriate time constant corresponding to each set of representative values of the coefficient and the stiffness characteristic coefficient is determined.

  As described above, the appropriate time constant value corresponding to each representative value set of the viscosity characteristic coefficient and the stiffness characteristic coefficient is determined, and therefore, the appropriate time constant value is determined from a plurality of acquired response characteristic data. Can be done efficiently and easily.

  In this case, the control method for acquiring the response characteristic data is not limited to a specific control method, and can be easily acquired using a general-purpose control method such as PD control.

  Further, the determination of the appropriate time constant value corresponding to each of the representative value pairs of the viscosity characteristic coefficient and the stiffness characteristic coefficient using the required fitting intersection is, for example, a known statistical method such as a least square method. This can be done by an identification technique.

  In the nineteenth invention, a plurality of representative values of the viscosity characteristic coefficient, a plurality of representative values of the rigidity characteristic coefficient, and a value of the appropriate time constant determined as described above corresponding to each of the set of the representative values. Based on this, the correlation among the appropriate time constant, the viscosity characteristic coefficient, and the rigidity characteristic coefficient is specified. This correlation corresponds to the correlation between the inclination of the switching hyperplane and the viscosity characteristic coefficient in the eighth invention.

  In the nineteenth aspect of the invention, during the operation control of the power transmission device, the switching hyperplane variable setting means sets the value of the viscosity characteristic coefficient to zero in the correlation and sets the value of the rigidity characteristic coefficient to the rigidity. The value of the appropriate time constant when matched with the control value of the characteristic coefficient matches the convergence time constant on the switching hyperplane that is the time constant of convergence of the first variable component to zero on the switching hyperplane. Thus, the inclination of the switching hyperplane is determined according to the control value of the stiffness characteristic coefficient.

  Further, in the correlation, a difference between the value of the appropriate time constant corresponding to the value of the arbitrary viscosity characteristic coefficient and the value of the appropriate time constant when the value of the viscosity characteristic coefficient is zero is the arbitrary time constant. The correction coefficient is determined in advance so as to be a function value of a correction value obtained by correcting the value of the viscosity characteristic coefficient with the correction coefficient.

  For this reason, the control input determination means can target the measured value of the secondary power with high robustness and desired convergence characteristics without changing the inclination of the switching hyperplane according to the control value of the viscosity characteristic coefficient. The control input can be determined by the control process of the sliding mode control using a correction value obtained by correcting the control value of the viscosity characteristic coefficient by an appropriate correction coefficient so that the value can be controlled to a value.

  In the nineteenth aspect, the correlation is, for example, a relation represented by the following equation (F) (20th aspect).

Tc = (a2 / sqrt (Ksp)) + a1 · Ksp + a0 + b · Kdp / Ksp (F)
However,
Tc: the appropriate time constant Ksp: the value of the stiffness characteristic coefficient Kdp: the value of the viscosity characteristic coefficient a2, a1, a0: a predetermined value determined in advance b: a predetermined value determined in advance as the correction coefficient

That is, according to various experiments and examinations of the inventors of the present application, when the viscosity characteristic coefficient and the rigidity characteristic coefficient of the elastic force generation mechanism are variably controlled, the first variable component is converged to zero. The correlation between the appropriate time constant, the viscosity characteristic coefficient, and the rigidity characteristic coefficient can be approximated by Expression (F).

  In this case, in the correlation of the formula (F), the value of (a2 / sqrt (Ksp)) + a1 · Ksp + a0 is the value of the appropriate time constant when the value of the viscosity characteristic coefficient is zero. The value of (a2 / sqrt (Ksp)) + a1 · Ksp + a0 is determined according to the control value of the stiffness characteristic coefficient as the convergence time constant on the switching hyperplane in the twentieth invention.

  Further, in the correlation of the formula (F), the value b is determined as the correction coefficient. That is, in the correlation of the formula (F), the value of the stiffness characteristic coefficient of the elastic force generation mechanism is set to the difference between the value of an arbitrary viscosity characteristic coefficient and the value of (a2 / sqrt (Ksp)) + a1 · Ksp + a0. The value of the correction coefficient b is determined so that the value obtained by correcting the value of the viscosity characteristic coefficient by the coefficient b (= b · Kdp) matches the value obtained by multiplication.

  By determining the inclination of the switching hyperplane and the correction coefficient in this way, it is possible to suitably converge the measured value of the secondary power to the target value.

  In the nineteenth or twentieth invention, the first allowable limit value is a value set corresponding to each representative value of the stiffness characteristic coefficient, and the first limit value corresponding to each representative value of the stiffness characteristic coefficient. 1 allowable limit value is a predetermined allowable limit value in which the displacement speed and the displacement acceleration of the primary element due to the driving force of the actuator are respectively determined in a state where the stiffness characteristic coefficient of the elastic force generation mechanism is controlled to the representative value. So as not to exceed the allowable limit value of the displacement speed of the primary side element, the allowable limit value of the displacement acceleration of the primary side element, and the representative value of the stiffness characteristic coefficient. Preferred (21st invention).

  According to the twenty-first aspect, for each of the sets of representative values of the viscosity characteristic coefficient and the rigidity characteristic coefficient, a displacement speed and a displacement acceleration of the primary side element among the plurality of response characteristic data are respectively determined in advance. The value of the appropriate time constant can be determined using response characteristic data that does not exceed the allowable limit value.

  Therefore, during the operation control of the power transmission device, the operation of the power transmission device is controlled so that the displacement speed and the displacement acceleration of the primary side element do not exceed predetermined allowable limit values, respectively. Regardless of the control state of the viscosity characteristic coefficient and the rigidity characteristic coefficient of the generating mechanism, it can be performed with high reliability. As a result, power transmission in a region where the displacement speed or displacement acceleration of the primary element is close to the allowable limit value, regardless of the control state of the viscosity characteristic coefficient and the rigidity characteristic coefficient of the elastic force generation mechanism. It becomes possible to perform the operation of the apparatus stably.

  In the nineteenth to twenty-first inventions, the second allowable limit value is a value set corresponding to each representative value of the stiffness characteristic coefficient, and the second value corresponding to each representative value of the stiffness characteristic coefficient. The second allowable limit value is a value of the power transmission system corresponding to the natural vibration of the power transmission system from the actuator to the secondary element in a state where the stiffness characteristic coefficient of the elastic force generation mechanism is controlled to the representative value. It is preferable that the value is determined according to the representative value of the stiffness characteristic coefficient so as to prevent the occurrence of vibration (22nd invention).

  According to the twenty-second aspect, vibration of the power transmission system corresponding to the natural vibration is generated among the plurality of response characteristic data for each set of representative values of the viscosity characteristic coefficient and the rigidity characteristic coefficient. By using such response characteristic data, an appropriate time constant value corresponding to each set of representative values of the viscosity characteristic coefficient and the rigidity characteristic coefficient can be determined.

  Therefore, regardless of the control state of the viscosity characteristic coefficient and the rigidity characteristic coefficient of the elastic force generation mechanism, the power transmission system is designed so that the vibration of the power transmission system according to the natural vibration does not occur or hardly occurs. It is possible to control the operation of the apparatus with high reliability.

  In the fifteenth to twenty-second inventions, the value of the first variable component is changed stepwise from zero to the first allowable limit value, and the convergence of the value of the first variable component to zero is controlled by the actuator. When the time constant realized when it is assumed that the displacement acceleration of the primary side element due to the driving force is set to the allowable limit value of the displacement acceleration is defined as the specific time constant, the constraint condition is the first condition. It is preferable that the time constant for convergence of the variable component to zero is a condition that limits the time constant to a value in the range equal to or greater than the specific time constant (23rd invention).

  That is, when it is assumed that the value of the first variable component is changed stepwise from zero to a certain value not more than the first allowable limit value, the value of the first variable component is changed from zero to the first allowable limit value. In this case, the relative displacement amount between the primary side element and the secondary side element is maximized.

  In this case, since the displacement of the primary side element with respect to the secondary side element is limited to the displacement at the displacement acceleration equal to or less than the allowable limit value, the time constant for convergence of the value of the first variable component to zero is the specified time. It cannot be smaller than a constant.

  Therefore, in the twenty-third aspect of the invention, the constraint condition that defines the requirement-matching intersection is a condition that limits a time constant for convergence of the first variable component to zero to a value in a range equal to or greater than the specific time constant. Set.

  Therefore, according to the twenty-third aspect, at the time of controlling the operation of the power transmission device, it is possible to stably control the secondary power to the target value in a wide range of operation of the power transmission device with high robustness.

  In the thirteenth to twenty-third inventions, a first variable component measurement value that is a value of the first variable component calculated from the measurement value of the secondary power and the target value of the secondary power; An estimated value of the first variable component obtained by reducing the influence of disturbance from a second variable component measured value that is a value of the second variable component calculated as a rate of temporal change of the first variable component measured value; The observer further includes an observer that sequentially calculates the estimated value of the second variable component, and the control input determining means uses the observer instead of the first variable component measured value and the second variable component measured value. While calculating the value of the switching function using the calculated estimated value of the first variable component and the estimated value of the second variable component, the control input is sequentially generated using the value of the switching function. Preferably, it is configured (24th invention).

  According to this twenty-fourth aspect, the value of the switching function is calculated while calculating the value of the switching function using the estimated value of the first variable component and the estimated value of the second variable component calculated by the observer. By using the control input of the sliding mode control to sequentially generate the control input, the influence of disturbance is suppressed.

  For this reason, the stability of the control input sequentially determined by the control input determining means is increased. As a result, the robustness of the control of the secondary power can be further enhanced.

The figure which shows the system configuration | structure in 1st Embodiment of this invention. The figure which shows the structure of the elastic force generation mechanism of the power transmission device shown in FIG. The figure which shows the principal part structure of the elastic force generation mechanism shown in FIG. IV-IV sectional view taken on the line of FIG. The figure which shows the switching hyperplane of sliding mode control. The graph for demonstrating the determination process of the inclination of a switching hyperplane. FIGS. 7A to 7C are graphs for explaining the determination process of the inclination of the switching hyperplane. 8A is an enlarged view of part A in FIG. 6, and FIG. 8B is an enlarged view of part B in FIG. The block diagram which shows the process of the control input determination part of the control apparatus shown in FIG. The block diagram which shows the process of the control input determination part in 2nd Embodiment of this invention. The block diagram which shows the process of the control input determination part in 3rd Embodiment of this invention. The block diagram which shows the process of the control input determination part in 4th Embodiment of this invention. The block diagram which shows the process of the control input determination part in 5th Embodiment of this invention.

[First Embodiment]
A first embodiment of the present invention will be described below with reference to FIGS.

  As shown in FIG. 1, the power transmission device 1 according to this embodiment includes a driving pulley 2 as a primary side element, a driven pulley 3 as a secondary side element, and an elastic force between these pulleys 2 and 3. An elastic force generating mechanism 4 for generating a viscous force, an electric motor 5 as an actuator for applying a rotational driving force to the driving pulley 2, and a load fixed to the driven pulley 3 so as to rotate integrally with the driven pulley 3. And a member 6.

  Although the load member 6 is illustrated as an integral structure in FIG. 1, the load member 6 may not be an integral structure. For example, the load member 6 may be a link mechanism including one or more joints.

  The drive pulley 2 is connected to the rotary drive shaft 5 a of the electric motor 5 via the speed reducer 7. The drive pulley 2 rotates in conjunction with the rotation of the rotation drive shaft 5a of the electric motor 5 by the rotation drive force (torque) applied from the rotation drive shaft 5a of the electric motor 5 via the speed reducer 7. It has become.

  Note that the speed reducer 7 may have an arbitrary structure, and for example, a speed reducer constituted by a harmonic drive (registered trademark) or a plurality of gears may be employed. Alternatively, the speed reducer 7 may include a mechanism that converts a linear motion into a rotational motion. In that case, as the actuator, for example, a linear motion actuator constituted by an electric motor and a ball screw, or an electric linear motor may be employed.

  In FIG. 1, the electric motor 5 and the drive pulley 2 are coaxially arranged, but their rotational axes may not be coaxial.

  The driven pulley 3 is juxtaposed on the side of the drive pulley 2 so that the rotation axis thereof is parallel to the rotation axis of the drive pulley 2.

  The elastic force generation mechanism 4 includes a wire 11 laid between the driving pulley 2 and the driven pulley 3 and a stiffness / viscosity variable mechanism 12 for changing the stiffness and viscosity between the pulleys 2 and 3.

  As shown in FIG. 2, the wire 11 has two extending portions 11 a and 11 b extending between the pulleys 2 and 3, and portions other than the extending portions 11 a and 11 b are both pulleys 2 and 3. Of the outer periphery of the drive pulley 2 so that it does not slip to any part except for the portion of the outer periphery of the drive pulley 2 that faces the driven pulley 3 and the portion of the outer periphery of the driven pulley 3 that faces the drive pulley 2. It is wrapped around. Note that the wire 11 has some elasticity.

  The stiffness / viscosity variable mechanism 12 is configured as shown in FIGS. That is, the stiffness / viscosity variable mechanism 12 includes a rotating bar 14 having rollers 13a and 13b pivotally attached to both ends. The rotary bar 14 can rotate integrally with the rotary shaft 15 around the axis of the rotary shaft 15 fixed at the center thereof. The rotating shaft 15 is disposed in a position parallel to the rotational axis of the pulleys 2 and 3 at a position between the driving pulley 2 and the driven pulley 3.

  The rollers 13 a and 13 b at both ends of the rotating bar 14 have their respective rotation axes oriented in a direction parallel to the rotation axes of the drive pulley 2 and the driven pulley 3.

  And the outer peripheral part of the inner end side (side facing the other roller 13b) of one roller 13a of the rollers 13a and 13b is one extending portion of the two extending portions 11a and 11b of the wire 11 The outer peripheral portion on the inner end side (the side facing one roller 13a) of the other roller 13b is in pressure contact with the other stretched portion 11b of the wire 11. In this case, the extending portions 11a and 11b of the wire 11 are curved at the press contact portions of the rollers 13a and 13b, respectively.

  The variable stiffness / viscosity mechanism 12 is further engaged with a gear (spur gear) 16 that is connected to the rotary bar 14 via a rotary shaft 15 so as to be rotatable integrally with the rotary bar 14. A spring worm 17, an electric motor 18 that rotationally drives the spring worm 17, and a cylinder 19 in which viscous oil is sealed.

  The spring worm 17 is a spring member formed in a coil spring shape so as to be able to function as a worm gear, and is externally attached to the rotation drive shaft 18 a of the electric motor 18. One end of the spring worm 17 near the main body of the electric motor 18 is fixed to a spring seat member 20a fixed to the rotation drive shaft 18a. Accordingly, the spring worm 17 rotates integrally with the rotation drive shaft 18a of the electric motor 18, and the gear 16 rotates as the spring worm 17 rotates.

  The cylinder 19 has a cylindrical portion 21 disposed coaxially with the rotation drive shaft 18 a on the other end side of the spring worm 17. A rotation drive shaft 18a of the electric motor 18 passes through the inside of the cylinder portion 21, and the cylinder portion 21 is slidable in the axial direction along the rotation drive shaft 18a. The other end of the spring worm 17 is fixed to a spring seat member 20b fixed to the end face of the cylindrical portion 21 on the spring worm 17 side. Therefore, as the spring worm 17 expands and contracts, the cylinder portion 21 of the cylinder 19 slides in the axial direction of the rotation drive shaft 18 a of the electric motor 18.

  In addition, a piston 22 fixed to the rotary drive shaft 18 a is provided inside the cylinder portion 21, and an outer peripheral surface of the piston 22 is in sliding contact with an inner peripheral surface of the cylinder portion 21.

  Then, viscous oil is enclosed in two oil chambers 23 a and 23 b defined by the piston 22 inside the cylinder portion 21. These oil chambers 23 a and 23 b are communicated by a communication pipe 25 having an orifice portion 24. In this case, the orifice portion 24 can change the opening area by a valve mechanism or the like (not shown).

  Here, the operation of the elastic force generation mechanism 4 having the above configuration will be described. When the spring worm 17 is rotationally driven by the electric motor 18, the rotating bar 14 is rotated via the gear 16 meshed with the spring worm 17. Therefore, the rotation angle (phase angle) of the rotating bar 14 can be controlled by the servo control of the electric motor 18. Here, in the following description, as shown in FIG. 2, the phase angle of the rotating bar 14 is the direction in which the rotating bar 14 extends (the interval direction between the rollers 13 a and 13 b). Is defined as a rotation angle φ of the rotating bar 14 from a state perpendicular to the rotation bar 14.

  In a state where power transmission between the pulleys 2 and 3 (transmission of rotational driving force) is not performed, the phase angle φ of the rotary bar 14 is controlled to a predetermined angle value (for example, φ0 in FIG. 2), and in this state Assume a state in which the rotation of the rotation drive shaft 18a of the electric motor 18 (and hence the rotation of the spring worm 17) is stopped (hereinafter, this state is referred to as a reference state).

  In this reference state, when a rotational driving force (torque) is applied from the electric motor 5 to the driving pulley 2, a tension proportional to the rotational driving force is generated in one of the stretched portions 11a and 11b of the wire 11, A rotational driving force is transmitted from the driving pulley 2 to the driven pulley 3 via tension.

  At the same time, due to the tension generated in one of the stretched portions 11a and 11b of the wire 11, both pulleys 2 are attached to the roller 13a or 13b of the rollers 13a and 13b that are in contact with the stretched portion 11a or 11b. A translational force in a direction substantially perpendicular to the three spacing directions acts.

  For example, as shown in FIG. 2, when a torque τd in the counterclockwise direction is applied to the drive pulley 2, a tension Te proportional to the torque τd (= τd / effective rotation radius of the drive pulley 2) is applied to the stretched portion 11a of the wire 11. ) Occurs, and the translation force F acts on the roller 13a by the tension Te. The magnitude of the translational force F is substantially proportional to the torque τd. Further, the tension Te in FIG. 2 indicates the tension acting on the roller 13a.

  When the phase angle of the rotating bar 14 in the reference posture state is not zero (for example, the situation shown in FIG. 2), due to the translational force acting on the roller 13a or 13b (hereinafter referred to as F), A rotational driving force (torque) is applied to the rotating bar 14. As a result, the drive pulley 2 rotates relative to the driven pulley 3 and the rotating bar 14 rotates. As a result, the gear 16 meshing with the spring worm 17 rotates integrally with the rotating bar 14.

  In this case, the torque (hereinafter referred to as τa) acting on the rotating bar 14 due to the translational force F acting on the roller 13a or 13b is related to the translational force F by the following equation (1). Have As shown in FIG. 2, φ0 is the value of the phase angle φ of the rotating bar 14 in the reference posture state, and Ra is the rotational radius of the shaft center portion of the rollers 13a and 13b around the shaft center of the rotating shaft 15. is there.

τa = F ・ sin (φ0) ・ Ra (1)

Since the spring worm 17 does not rotate in the situation where the torque τa is applied to the rotary bar 14 as described above, a part of the spring worm 17 (specifically, the meshed portion of the gear 16 and the electric motor 18 are rotated by the rotation of the gear 16. The portion between the side spring seat member 19a) is extended or shortened, and the spring worm 17 generates an elastic force corresponding to the amount of expansion / contraction.

  In this case, the amount of expansion and contraction of the spring worm 17 from the reference state, and hence the amount of rotation of the rotating bar 14 from the reference state (the amount of change in phase angle) is applied to the gear 16 by the elastic force (translation force) of the spring worm 17. The applied torque balances with the torque acting on the rotating bar 14 (= the torque acting on the gear 16) due to the translational force F acting on the roller 13 a or 13 b due to the tension of the wire 11. In this equilibrium state, the torque τd applied to the drive pulley 2 is transmitted to the driven pulley 3 via the elastic force generating mechanism 4.

  The rotation amount of the rotating bar 14 from the reference state in the equilibrium state is Δφ [rad], the rotation radius of the gear 16 is Rb, the rigidity of the spring worm 17 (the elasticity generated per unit change amount of the expansion / contraction amount of the spring worm 17). If the force change amount is Ksp_w, the torque (hereinafter referred to as τb) acting on the gear 16 by the elastic force of the spring worm 17 in the equilibrium state is given by the following equation (2).

τb = Ksp_w · sin (Δφ) · Rb≈Ksp_w · Δφ · Rb (2)

From this equation (2) and the above equation (1), the relationship between the translational force F in the equilibrium state and the rotation amount Δφ from the reference state of the rotating bar 14 is given by the following equation (3).

F = (Ksp_w · Rb / (sin (φ0) · Ra) · Δφ (3)

Therefore, the translational force F acting on the roller 13a or 13b by the tension of the wire 11 is proportional to the rotation amount Δφ of the rotating bar 14 from the reference state.

  Here, the translational force F acting on the roller 13a or 13b by the tension of the wire 11 increases as the torque applied to the drive pulley 2 (and thus the torque transmitted to the driven pulley 3) increases. Further, the relative rotation amount (relative rotation amount from the reference state) of the drive pulley 2 with respect to the driven pulley 3 increases as the rotation amount from the reference state of the rotary bar 14 in the equilibrium state increases.

  Accordingly, the torque transmitted from the drive pulley 2 to the driven pulley 3 (torque applied to the driven pulley 3) is expressed as τsp in a steady state where the relative rotation amount between the pulleys 2 and 3 is kept constant. When the relative rotation amount between the pulleys 2 and 3 is expressed as Δθ, a proportional relationship of the following equation (4) is approximately established between τsp and Δθ.

τsp = Ksp · Δθ (4)

Therefore, the elastic force generation mechanism 4 functions as a spring member that transmits power between the driving pulley 2 and the driven pulley 3. The torque τsp corresponds to torque transmitted by the elastic force generated between the pulleys 2 and 3 by the elastic force generating mechanism 4 (hereinafter referred to as elastic force torque τsp). In this case, Ksp in equation (4) is the ratio of the change in elastic force torque τsp to the change in the relative rotation amount Δθ between the pulleys 2 and 3 (hereinafter referred to as the pulley rotation angle difference Δθ). It is called a characteristic coefficient Ksp.

  The stiffness characteristic coefficient Ksp indicates the stiffness between the pulleys 2 and 3, and the greater the value of Ksp, the higher the stiffness between the pulleys 2 and 3 (the difference in rotational amount between the pulleys 2 and 3 is greater). It is difficult to occur). The value of the stiffness characteristic coefficient Ksp of the elastic force generation mechanism 4 in this embodiment basically corresponds to the phase angle φ0 of the rotating bar 14 in the reference state, and the value of Ksp increases as φ0 increases. Becomes smaller.

  Further, in the elastic force generation mechanism 4 of the present embodiment, when the rotary bar 14 rotates from the reference state, the cylinder portion 21 of the cylinder 19 slides relative to the piston 22 as the spring worm 17 expands and contracts. .

  At this time, when viscous oil flows between the oil chambers 23a and 23b through the communication pipe 25 having the orifice portion 24, a viscous force serving as a resistance force against the sliding of the cylindrical portion 21 is generated. As a result, a rotation of the rotating bar 14 from the reference state, and thus a viscous force that is a resistance force against the relative rotation of the driving pulley 2 with respect to the driven pulley 3 occurs between the pulleys 2 and 3. And this viscous force will change by changing the opening area of the orifice part 24. FIG.

  In this case, the viscous force generated in the cylinder 19 when the opening area of the orifice portion 24 is kept constant is proportional to the moving speed of the cylinder portion 21 with respect to the piston 22, and consequently the expansion / contraction speed of the spring worm 17. The expansion / contraction speed of the spring worm 17 is substantially proportional to the rotation speed of the rotary bar 14.

  Further, the temporal change rate of the relative rotation amount between the pulleys 2 and 3, that is, the relative rotation speed between the pulleys 2 and 3 (difference between the rotation angles of the pulleys 2 and 3), The relative rotation speed between the pulleys 2 and 3 increases as the rotation speed of the rotation bar 14 increases.

  Therefore, the relative rotational speed between the pulleys 2 and 3 is expressed as Δω [rad / s], and the torque applied to the driven pulley 3 by the viscous force between the pulleys 2 and 3 (hereinafter referred to as the viscous force torque) When expressed as τdp, the relationship of the following equation (5) is approximately established between Δω and τdp.

τdp = Kdp · Δω (5)

Therefore, the elastic force generation mechanism 4 also has a function of generating a viscous force between the driving pulley 2 and the driven pulley 3. In this case, Kdp in equation (5) is the ratio of the change in the viscous force torque τdp to the change in the relative rotational speed Δω between the pulleys 2 and 3 (hereinafter referred to as the inter-pulley rotational speed difference Δω). It is called a characteristic coefficient Kdp.

  The viscosity characteristic coefficient Kdp indicates the degree of viscosity between the pulleys 2 and 3, and the larger the value of Kdp, the higher the viscosity between the pulleys 2 and 3 (viscosity generated between the pulleys 2 and 3). It means that power is likely to increase). The value of the viscosity characteristic coefficient Kdp of the elastic force generation mechanism 4 in the present embodiment basically corresponds to the opening area of the orifice portion 24, and the value of Kdp decreases as the opening area increases. .

  The above is the mechanical configuration of the power transmission device 1 of the present embodiment.

  In the present embodiment, as a configuration for controlling the operation of the power transmission device 1, a control device 30 and angle detectors 31 and 32 for detecting the rotation angle θin of the drive pulley 2 and the rotation angle θout of the driven pulley 3, respectively. And an angle detector 33 for detecting the rotation angle θw of the rotational drive shaft 18a of the electric motor 18 of the stiffness / viscosity variable mechanism 12.

  The angle detectors 31, 32, and 33 are configured by, for example, a rotary encoder. The angle detectors 31 and 32 are provided to face the driving pulley 2 and the driven pulley 3, respectively, and the angle detector 33 is attached to the electric motor 18. Note that each of the angle detectors 31, 32, and 33 may be configured by an angle sensor other than a rotary encoder, such as a potentiometer.

  The control device 30 is configured by an electronic circuit unit including a CPU, RAM, ROM, interface circuit, and the like, and receives output signals (detection signals) from the angle detectors 31, 32, and 33. Further, the control device 30 has a target value τ_cmd (hereinafter referred to as “secondary torque τ”) of a torque (hereinafter referred to as “secondary torque τ”) applied to the driven pulley 3 by transmission of the rotational driving force from the drive pulley 2 to the driven pulley 3. Secondary target torque τ_cmd), target value Ksp_cmd of the stiffness characteristic coefficient Ksp (hereinafter referred to as target stiffness characteristic coefficient Ksp), target value Kdp_dmp of the viscosity characteristic coefficient Kdp (hereinafter referred to as target viscosity characteristic coefficient Kdp), and Are sequentially input from other external control devices or servers.

  The secondary target torque τ_cmd, the target stiffness characteristic coefficient Ksp, and the target viscosity characteristic coefficient Kdp are target values for performing a desired operation of the load member 6. In the present embodiment, the secondary target torque τ_cmd and the target viscosity characteristic coefficient Kdp are target values that are arbitrarily variably set within a predetermined range. On the other hand, the target stiffness characteristic coefficient Ksp_cmd is a predetermined fixed value (fixed value) target value in the present embodiment.

  The target stiffness characteristic coefficient Ksp in the present embodiment is more specifically the target value of the stiffness characteristic coefficient Ksp in the reference state of the elastic force generation mechanism 4.

  The control device 30 uses the detection signals input from the angle detectors 31, 32, and 33, the secondary target torque τ_cmd, the target stiffness characteristic coefficient Ksp, and the target viscosity characteristic coefficient Kdp to generate the electric motor 5 (hereinafter, referred to as the electric motor 5). The power source motor 5), the electric motor 18 of the elastic force generation mechanism 4 (hereinafter, referred to as a stiffness variable motor 18), and the orifice portion 24 are controlled.

  More specifically, the control device 30 uses the actual secondary torque τ (observation) as a function realized by executing a program to be implemented (function realized by software) or a function realized by a hardware configuration. Value) to follow the secondary target torque τ_cmd, through a control input determining unit 34 for sequentially determining a control input for controlling the operation of the power source motor 5 and a motor drive circuit (not shown) according to the control input. And a motor control unit 35 that controls the energization current (and consequently the output torque) of the power source motor 5.

  In the present embodiment, the control input is, for example, a target torque (target value of output torque) of the power source motor 5. However, the control input is not limited as long as it can define the target torque of the power source motor 5, and may be other than the target torque. For example, as the control input, a target value of torque applied to the drive pulley 2 itself from the power source motor 5 side, or a target value of the energization current of the power source motor 5 can be used.

  Although details will be described later, the control input determination unit 34 sequentially determines the control input for causing the actual secondary side torque τ (observed value) to follow the secondary side target torque τ_cmd by the control process of the sliding mode control. .

  Further, the motor control unit 35 controls the energization current of the power source motor 5 so that the target torque defined by the control input is output to the power source motor 5. For example, the target value of the energization current of the power source motor 5 is determined according to the target torque, and the actual energization current (observed value) is feedback-controlled to the target value of the energization current.

  Further, the control device 30 controls, as its function, a stiffness control unit 36 that controls the stiffness variable motor 18 according to the target stiffness characteristic coefficient Ksp_cmd (and thus controls the stiffness between the pulleys 2 and 3), and According to the target viscosity characteristic coefficient Kdp_cmd, there is provided a viscosity control section 37 that controls the orifice section 24 (and thus controls the viscosity between the pulleys 2 and 3).

  The rigidity control unit 36 sets a target value of the rotation angle of the rotation drive shaft 18a of the rigidity variable motor 18 for setting the phase angle φ of the rotation bar 14 in the reference state to an angle value corresponding to the target rigidity characteristic coefficient Ksp_cmd. , Ksp_cmd is determined by a predetermined map or arithmetic expression. Then, the rigidity control unit 36 controls the actual rotation angle of the rotation drive shaft 18a (observed value indicated by the output of the angle detector 33) to the target value determined from Ksp_cmd by servo control.

  The viscosity control unit 37 determines a target value of the opening area of the orifice unit 24 corresponding to the target viscosity characteristic coefficient Kdp_cmd from Kdp_cmd using a predetermined map or arithmetic expression. And the viscosity control part 37 controls the actual opening area of the orifice part 24 to the determined target value via the valve mechanism etc. which are not shown in figure.

  Supplementally, since the stiffness characteristic coefficient Ksp of the elastic force generation mechanism 4 is defined according to the value of the phase angle φ of the rotation bar 14 in the reference state, instead of the target stiffness characteristic coefficient Ksp_cmd, The target value of the phase angle φ (the value of the phase angle φ of the rotary bar 14 corresponding to Ksp_cmd), or the target value of the rotational angle of the rotary drive shaft 18a of the variable stiffness motor 18 (of the rotary drive shaft 18a corresponding to Ksp_cmd) The value of the rotation angle may be input to the control device 30.

  Furthermore, when the target stiffness characteristic coefficient Ksp_cmd used during the operation of the power transmission device 1 is a constant value as in this embodiment, the stiffness variable motor 18 and the stiffness control unit 36 are omitted, and one end of the spring worm 17 ( One end on the opposite side of the cylinder 19) is fixed to a member (not shown) that rotatably supports the driving pulley 2 and the driven pulley 3 and the like, and power is not transmitted between the pulleys 2 and 3 The gear 16 may be meshed with the spring worm 17 so that the phase angle φ of the rotating bar 14 in the reference state becomes a phase angle corresponding to the target stiffness characteristic coefficient Ksp_cmd. In this case, the spring worm 17 functions as a simple spring having no function as a worm gear.

  Further, since the viscosity characteristic coefficient Kdp of the elastic force generating mechanism 4 is defined according to the opening area of the orifice part 24, the target value (Kdp_cmd) of the opening area of the orifice part 24 is used instead of the target viscosity characteristic coefficient Kdp_cmd. The value of the opening area of the corresponding orifice portion 24) may be input to the control device 30.

  Further, the secondary target torque τ_cmd, the target stiffness characteristic coefficient Ksp, and the target viscosity characteristic coefficient Kdp may be sequentially determined by the control device 30.

  Next, basic matters regarding the control processing in the control input determination unit 34 will be described.

  The behavior of the power transmission device 1 as a system to be controlled in the present embodiment is modeled by a state equation of the following equation (10) in a discrete system.

  Here, θin is the rotation angle of the drive pulley 2, θout is the rotation angle of the driven pulley 3, dθin is the time change rate of θin (that is, the rotation angular velocity of the drive pulley 2), and dθout is the time change rate of θout (ie, , Rotation angular velocity of the driven pulley 3), DT is a control processing cycle, Ksp is the stiffness characteristic coefficient, Kdp is the viscosity characteristic coefficient, and Iin is a system on the drive pulley 2 side (here, the drive pulley 2, the speed reducer 7, and the electric motor) Inertia (input-side inertia) of the system constituted by the motor 5, Iout is an inertia (output-side inertia) of a system on the driven pulley 3 side (here, a system constituted by the driven pulley 3 and the load member 6), u is the input torque on the side of the drive pulley 2 (for example, the output torque of the power source motor 5). The subscripts n and n-1 in parentheses are numbers representing discrete time.

  On the other hand, the secondary torque τ is a torque component caused by the elastic force generated between the pulleys 2 and 3 by the elastic force generating mechanism 4 (that is, the elastic force torque τsp expressed by the above equation (4)), and The resultant torque is a combined torque with a torque component (that is, the viscous force torque τdp expressed by the above equation (5)) caused by the viscous force generated between the pulleys 2 and 3 by the elastic force generating mechanism 4.

  Therefore, the following expressions (11a) and (11b) are established with respect to the secondary torque τ and its temporal change rate dτ (hereinafter referred to as secondary torque change rate dτ).

τ (n) = Ksp · Δθ (n) + Kdp · Δω (n) (11a)
dτ (n) = Ksp · Δω (n) + Kdp · dΔω (n) (11b)
However,
Δθ (n) ≡θin (n) −θout (n) (12a)
Δω (n) ≡ωin (n) −ωout (n) = dθin (n) −dθout (n) (12b)
dΔω (n) ≡dωin (n) −dωout (n) (12c)

Note that dΔω in the equation (11b) is the difference between the pulley rotation speeds Δω expressed by the expression (12b) (the difference between the rotation angular velocities ωin (= dθin) and ωout (= dθout) of the pulleys 2 and 3). The rate of change over time, where dωin in equation (12c) is the rate of change over time (= rotation angular acceleration) of the rotational angular velocity ωin (= dθin) of the drive pulley 2, and dωout is the rotational angular velocity ωout (= dθout) is a temporal change rate (= rotational angular acceleration).

  Here, in order to feedback-control the secondary side torque τ, it is considered to measure (observe) the secondary side torque τ and the secondary side torque change rate dτ which is a temporal change rate thereof. In this case, if Δθ is measured, the measured value of Δω can be calculated as a temporal change rate (differential value) of the measured value of Δθ, and the measured value of dΔω is the time of the measured value of Δω. It can be calculated as a dynamic change rate (differential value).

  Therefore, in principle, τ and dτ are respectively expressed by the equations (11a) and (11b) based on the measured values of Δθ obtained from the measured values of θin and θout indicated by the outputs of the angle detectors 31 and 32. It can be measured.

  However, in this case, since the measured value of Δω is calculated as a temporal change rate (differential value) of the measured value of Δθ, in general, an error is likely to occur with respect to an actual value (true value). As a result, the measurement value of dΔω calculated as the temporal change rate (differential value) of the measurement value of Δω is more likely to cause an error. Further, since the viscosity of the viscous oil in the cylinder 19 of the elastic force generating mechanism 4 is easily affected by the environmental temperature or the like, the actual value of the viscosity characteristic coefficient Kdp in the above equation (5) is relative to the target viscosity characteristic coefficient Kdp_cmd. Error.

  From these facts, in particular, the measured value of the second term (= viscous force torque τdp) of the equation (11a) and the measured value of the second term (= time change rate of τdp) of the equation (11b) It is difficult to accurately obtain As a result, even if it is going to measure (tau) and (dτ) using said Formula (11a) and (11b) as it is, it is difficult to acquire a reliable measured value. Further, even if an attempt is made to control the actual secondary torque τ to the secondary target torque τ_cmd using such measured values of τ and dτ, it is difficult to appropriately perform highly stable control.

  On the other hand, the actual value of the stiffness characteristic coefficient Ksp in the equation (4) (value in the reference state) can be controlled to the target stiffness characteristic coefficient Ksp_cmd with relatively high accuracy. Therefore, the measured value of the first term (= elastic force torque τsp) of the equation (11a) and the measured value of the first term (= time change rate of τsp) of the equation (11b) are the viscous force torque. Compared to the measured value of τdp and its temporal change rate, it can be obtained with higher accuracy. Further, in a steady state where the inter-pulley rotation angle difference Δθ is kept constant, the viscous force torque τdp is zero, so the actual secondary torque τ matches the elastic force torque τsp.

  Therefore, in the present embodiment, the following equation is obtained by removing the second term on the right side of equation (11a) from the measured value of Δθ obtained as the difference between the measured values of θin and θout indicated by the outputs of the angle detectors 31 and 32. Based on (13a) and the following equation (13b) obtained by removing the second term on the right side of equation (11b), the measured values of τ and dτ are acquired.

  In other words, assuming that the actual viscous force torque τdp between the pulleys 2 and 3 and its rate of change dτdp are zero (assuming that τ = τsp), the measured values of τ and dτ Are obtained based on the following equations (13a) and (13b), respectively.

τ (n) = Ksp · Δθ (n) = Ksp · (θin (n) −θout (n)) (13a)
dτ (n) = Ksp · Δω (n) = Ksp · (dθin (n) −dθout (n)) (13b)

Note that the target stiffness characteristic coefficient Ksp_cmd is used as the value of Ksp required to obtain the measured values of τ and dτ by these equations (13a) and (13b).

  In this embodiment, while using the measured values of τ and dτ acquired in this way, the viscous force between the pulleys 2 and 3 that is not reflected in these measured values (by the elastic force generation mechanism 4). The control input for feedback control of the actual secondary torque τ to the secondary target torque τ is determined by executing the sliding mode control process so as to compensate for the influence of the viscous force generated by .

  In the present embodiment, the calculation process for determining the control input by the sliding mode control method assumes that the viscosity characteristic coefficient Kdp of the elastic force generation mechanism 4 is zero, that is, between the pulleys 2 and 3. It is constructed based on a model that expresses the behavior of the power transmission device 1 when it is assumed that no viscous force is generated.

  The behavior of the power transmission device 1 in this case is expressed by a state equation in which Cin and Cout in the equation (10) are zero, that is, a state equation in the following equation (14).

  When this equation of state is rearranged using the above equations (13a) and (13b), as a model expressing the behavior regarding τ and dτ when the viscosity characteristic coefficient Kdp of the elastic force generation mechanism 4 is assumed to be zero. The following equation (15) is obtained.

  In the sliding mode control in the present embodiment, a control process for determining a control input is constructed based on the state equation (model) of the equation (15).

  More specifically, in this embodiment, as a state variable to be controlled in the sliding mode control, as shown by the following equation (16), the deviation between the measured value of the secondary torque τ and the target secondary torque τ_cmd. A state variable X (vertical vector of 2 rows and 1 column) composed of a secondary torque deviation τ_err and a secondary torque deviation speed dτ_err which is a temporal change rate (differential value) of the deviation τ_err is two components. Used. Note that the superscript “T” in equation (16) means transposition.

X = [τ_err, dτ_err] ^T (16)
However, τ_err = τ−τ_cmd, dτ_err = τ_err temporal change rate (differential value)

In this case, the target torque τm_cmd (new target torque determined for each control processing cycle) of the power source motor 5 as the control input by the sliding mode control is the matrix A and the column vector defined by the proviso of the equation (15). Using B and the switching function σ expressed by the following equation (18), for example, it can be determined by the following equation (17).

τm_cmd (n) =-(S ・ B) ^-1・ (S ・ A ・ X (n)
+ Ksld · (σ (n) / (| σ (n) | + δ)))
...... (17)
σ (n) = S · X (n) = s 1 · τ_err (n) + s 2 · dτ_err (n) (18)
However, S = [s1, s2] (: 1 row by 2 column row vector)

These formulas (17) and (18) are basic formulas for determining the target torque τm_cmd of the power source motor 5 as a control input in this embodiment.

  In this case, the measured values of τ and dτ necessary for calculating τ_err and dτ_err in Expression (18) are measured values acquired based on Expressions (13a) and (13b) as described above.

  In addition, each component of the matrix A and the column vector B necessary for the calculation of the right side of the equation (17) is a value calculated based on the proviso definition equation of the equation (15). The values of the control processing period DT, the inertia on the driving pulley 2 side (input side inertia) Iin, and the inertia on the driven pulley 3 side (output side inertia) Iout, which are required to calculate each component of A and B, are as follows: In the embodiment, each is a predetermined value (a constant value) determined in advance.

  Further, as the value of the stiffness characteristic coefficient Ksp, the value of the target stiffness characteristic coefficient Ksp_cmd (a constant value in the present embodiment) is used.

  On the other hand, each component of S (coefficient components constituting the switching function σ) s1, s2 is a predetermined value variably determined according to the target viscosity characteristic coefficient Kdp_cmd, as will be described later.

  Here, the technical meaning of the equations (17) and (18) will be described. The first term on the right side of the equation (17) is a state in which a pair of values of τ_err and dτ_err exists on the switching hyperplane. , Τ_err means a control input component that functions to converge to zero.

  The switching hyperplane is represented by the equation σ = 0. Therefore, the inclination of the switching hyperplane σ = 0 (in this case, a straight line) in the phase plane having τ_err and dτ_err as two coordinate axis components is defined by the ratio of the S components s1 and s2.

  For example, as shown in FIG. 5, assuming a phase plane in which the coordinate axis of the secondary torque deviation speed dτ_err is the vertical axis and the coordinate axis of the secondary torque deviation τ_err is the horizontal axis, the switching hyperplane σ = The slope of 0 (straight line) is -s1 / s2. In addition, assuming a phase plane in which the coordinate axis of the secondary torque deviation speed dτ_err is the horizontal axis and the coordinate axis of the secondary torque deviation τ_err is the vertical axis, the gradient of the switching hyperplane σ = 0 in the phase plane Becomes -s2 / s1.

  On the switching hyperplane σ = 0, dτ_err = (− s1 / s2) · τ_err. Therefore, if (−s1 / s2) is set to a negative value, the switching hyperplane σ = 0. Τ_err will converge to zero. In this case, the time constant Tc of convergence of τ_err on the switching hyperplane σ = 0 (hereinafter referred to as the convergence time constant Tc) is given by the following equation (19). Further, the relationship between τ_err and dτ_err on the switching hyperplane σ = 0 is expressed by the following equation (20) using the convergence time constant Tc of equation (19).

Tc = s2 / s1 (19)
τ_err = −Tc · dτ_err (20)

Thus, the inclination of the switching hyperplane σ = 0 or the convergence time constant Tc is defined by the ratio of the coefficient components s1, s2.

  Supplementally, assuming a phase plane in which the coordinate axis of the secondary torque deviation speed dτ_err is the vertical axis and the coordinate axis of the secondary torque deviation τ_err is the horizontal axis, the slope of the switching hyperplane σ = 0 in the phase plane is Since -s1 / s2, the relationship between the slope and the convergence time constant Tc is slope = -1 / Tc.

  On the other hand, assuming a phase plane in which the coordinate axis of the secondary torque deviation speed dτ_err is the horizontal axis and the coordinate axis of the secondary torque deviation τ_err is the vertical axis, the gradient of the switching hyperplane σ = 0 in the phase plane is − Since s 2 / s 1, the relationship between this slope and the convergence time constant Tc is slope = −Tc.

  Thus, the inclination of the switching hyperplane σ = 0 and the convergence time constant Tc on the switching hyperplane σ = 0 have a one-to-one correspondence. Therefore, determining the inclination of the switching hyperplane σ = 0 is equivalent to determining the value of the convergence time constant Tc on the switching hyperplane σ = 0.

  In the present embodiment, since the measured values of τ and dτ used for determining the target torque τm_cmd by the equations (17) and (18) are measured values based on the equations (13a) and (13b), the elasticity In a situation where a viscous force is generated between the pulleys 2 and 3 by the force generating mechanism 4, the measured values of τ and dτ have an error from the actual values.

  In the state equation (15) that is the basis of the equation (17), the viscosity between the pulleys 2 and 3 is ignored, and the matrix A depends on the viscosity characteristic coefficient Kdp of the elastic force generation mechanism 4. Contains no ingredients. Furthermore, in this embodiment, the viscosity characteristic coefficient Kdp of the elastic force generation mechanism 4 is variably controlled.

  Therefore, in a situation where viscous force is generated in the elastic force generating mechanism 4, suppose that the gradient of the switching hyperplane σ = 0 (or the convergence time constant Tc on the switching hyperplane σ = 0) is the viscosity of the elastic force generating mechanism 4. If it is set not to depend on the characteristic coefficient Kdp, τ_err cannot be stably converged to zero with an appropriate time constant. For example, vibration of τ_err is likely to occur.

  Therefore, in this embodiment, the viscous force (between the pulleys 2 and 3) of the elastic force generation mechanism 4 that is not reflected in the measured value based on the equations (13a) and (13b) and the state equation of the equation (15). In order to compensate for the influence of the viscous force of the elastic force generating mechanism 4, the gradient of the switching hyperplane σ = 0 (or the convergence time constant Tc on the switching hyperplane σ = 0) and the control value of the viscosity characteristic coefficient Kdp of the elastic force generating mechanism 4 The correlation between the target viscosity characteristic coefficient Kdp_cmd as a value (realized by control) is determined in advance, and the gradient (or the switching hyperplane σ = 0 according to the target viscosity characteristic coefficient Kdp_cmd is determined according to the correlation (or The convergence time constant Tc) on the switching hyperplane σ = 0 is variably determined.

  Here, according to various experiments and examinations by the inventors of the present application, when the stiffness characteristic coefficient Ksp of the elastic force generation mechanism 4 is kept constant or substantially constant, τ_err and dτ_err when τ_err stably converges to zero. Is approximated by the relationship of the following equation (21).

τ_err = − (a + (b · Kdp_cmd) / Ksp_cmd) · dτ_err (21)

Therefore, in the present embodiment, the slope on the switching hyperplane σ = 0 (or the switching hyperplane so that the relationship between τ_err and dτ_err on the switching hyperplane σ = 0 becomes the relationship of Expression (21). The convergence time constant Tc) over σ = 0 is determined according to the target viscosity characteristic coefficient Kdp_cmd.

  In other words, as the correlation between the convergence time constant Tc on the switching hyperplane σ = 0 and Kdp_cmd is expressed by the following equation (22), the slope (or switching) of the switching hyperplane σ = 0: The convergence time constant Tc) on the hyperplane σ = 0 is determined according to the target viscosity characteristic coefficient Kdp_cmd.

Tc = (a + (b · Kdp_cmd) / Ksp_cmd) (22)

The variable a in the equations (21) and (22) is the convergence time constant Tc on the switching hyperplane σ = 0 when the viscosity characteristic coefficient Kdp of the elastic force generation mechanism 4 is zero (when Kdp_cmd = 0). Means the value of

  Further, since the variable b in the equations (21) and (22) is likely to cause an error in the actual viscosity characteristic coefficient Kdp with respect to the target viscosity characteristic coefficient Kdp_cmd, the target viscosity characteristic coefficient Kdp_cmd is set to the elastic force generation mechanism 4. Is a correction coefficient for correcting to an estimated value (= b · Kdp_cmd) of the actual viscosity characteristic coefficient Kdp.

  Here, the relationship between the viscous force torque τdp and the temporal change rate dτsp of the elastic force torque τsp is expressed by the following equation (23) using the equations (4) and (5).

τdp = (Kdp / Ksp) · dτsp (23)

Therefore, the right side of the equation (21) shows the actual secondary when the actual viscosity characteristic coefficient Kdp of the elastic force generating mechanism 4 is zero (when the viscous force between the pulleys 2 and 3 is zero). It has a meaning as a deviation between the elastic force torque τsp as the side torque τ and the secondary target torque τ_cmd. When the actual viscosity characteristic coefficient Kdp of the elastic force generating mechanism 4 is not zero, the actual secondary torque τ and the secondary target torque obtained by combining the elastic force torque τsp and the viscous force torque τdp. It has a meaning as a deviation from τ_cmd.

  In this embodiment, the values of the variables a and b in the equation (22) are determined in advance as will be described later, so that the convergence time constant Tc on the switching hyperplane σ = 0 and the target of the elastic force generation mechanism 4 are obtained. The correlation between the viscosity characteristic coefficient Kdp_cmd (and hence the correlation between the inclination of the switching hyperplane σ = 0 and Kdp_cmd) is determined in advance as the relationship of Expression (22). Then, the values of the coefficient components s1, s2 necessary for the calculations of the equations (17), (18) are determined corresponding to the convergence time constant Tc determined according to Kdp_cmd based on this correlation.

  In this case, one value of s1, s2 may be a constant value, for example, s1 = 1 (or s2 = 1). In that case, the value of the coefficient component s2 (or coefficient component s1) is uniquely determined by determining the slope of the switching hyperplane σ = 0 (or the convergence time constant Tc on the switching hyperplane σ = 0). Will be.

  The second term on the right side of the equation (17) converges the value of the switching function σ to zero (in other words, converges the set of values of τ_err and dτ_err on the switching hyperplane σ = 0). Means a functioning control input component. Ksld and δ are parameters that define the convergence characteristics of the value of the switching function σ.

  In this case, in this embodiment, δ is a predetermined value (constant value) set in advance based on experiments or the like.

  Further, the value of Ksld is determined to be a predetermined value (a constant value), for example, according to the following guidelines, or is determined according to the value of the switching function σ.

That is, in order to converge the value of the switching function σ to zero, the differential value (temporal change rate) of the Lyapunov function σ ² of σ needs to be a negative value.

This necessary condition is equivalent to the condition that (σ (n)) ² − (σ (n−1)) ² <0 in the discrete system. From this condition and the expressions (17) and (18), the condition of the following expression (24) regarding Ksld is obtained.

| Ksld | <| σ (n) | + δ (24)

Therefore, the value of Ksld only needs to be set so as to satisfy the condition of the above formula (24).

  In the present embodiment, the value of Ksld is variably determined according to the value of σ so that the magnitude of Ksld increases as the absolute value of the switching function σ for each control processing cycle increases.

  For example, Ksld is determined to be a value proportional to | σ (n) | as shown in the following equation (25) at each control processing cycle.

Ksld = (1 / K0) · | σ (n) | (25)

K0 in Expression (25) is a constant value (for example, an integer value of 3 or more) set in advance so as to satisfy the condition of Expression (24) within the range of actual values that can be taken by σ (n).

  The value of Ksld may be a constant value. Alternatively, the value of Ksld may be set so that | Ksld | <δ.

  Next, a preparatory process for determining the correlation between the gradient of the switching hyperplane σ = 0 (or the convergence time constant Tc on the switching hyperplane σ = 0) and the target viscosity characteristic coefficient Kdp_cmd will be described. . In this embodiment, this advance preparation process is performed in the following procedure.

  (Procedure 1) First, in a state in which the opening area of the orifice portion 24 of the elastic force generating mechanism 4 is controlled according to one representative value selected from a plurality of predetermined representative values of the target viscosity characteristic coefficient Kdp_cmd, as appropriate. The secondary torque τ of the power transmission device 1 is changed from a variety of arbitrary initial states of the power transmission device 1 to various target values (steps) by the control method (for example, PD control). The experiment (or simulation) for performing control to converge to the target value) is performed.

  In this experiment, according to the deviation between the target value of the secondary torque τ and the measured value (measured value based on the equation (13a)), the deviation is converged to zero by a general control law such as a PD control law. The output torque of the power source motor 5 is manipulated so that In this case, the rotation angle of the rotation drive shaft 18a of the stiffness variable motor 18 (= the rotation angle of the spring worm 17) is constant at a rotation angle corresponding to the target stiffness characteristic coefficient Ksp_cmd (a constant value in this embodiment). Retained.

  The operation control of the power source motor 5 and the stiffness variable motor 18 may be performed using an appropriate computer or the like.

  Then, the value of the secondary torque deviation τ_err (the value calculated using the measured value of τ based on the equation (13a)) and the value of the secondary torque deviation speed dτ_err (the above equation) in each experiment of the convergence control. The transition (temporal change) of the pair of values calculated using the measured value of dτ based on (13b) is measured. Further, by plotting the measurement data on a phase plane having τ_err and dτ_err as two coordinate axis components, response characteristic data indicating the transition trajectory of a set of values of τ_err and dτ_err is created.

  In this way, a plurality of response characteristic data (response characteristic data corresponding to representative values during selection of the target viscosity characteristic coefficient Kdp_cmd) in which the conditions such as the initial state and the target value of the secondary torque τ are variously varied are obtained. Created.

  Some typical examples of the response characteristic data created in this way are shown in FIG. Each of the trajectories marked with reference signs a1 to a6 in FIG. 6 shows an example of response characteristic data.

  (Procedure 2) Next, the response characteristic data satisfying a predetermined requirement among the plurality of response characteristic data created as described above is an inclination of the switching hyperplane σ = 0 in the phase plane (the target viscosity characteristic coefficient Kdp_cmd It is selected as specific response characteristic data (hereinafter referred to as inclination determination response characteristic data) for determining (slope corresponding to the representative value being selected).

  The inclination determination response characteristic data selected here is such that the magnitude (absolute value) of the secondary torque deviation τ_err on the locus indicated by the inclination determination response characteristic data is a predetermined first allowable limit. Requirement that the magnitude (absolute value) of the secondary torque deviation speed dτ_err on the locus is equal to or smaller than a predetermined second allowable limit value dτ_err_lim (hereinafter, (Referred to as selection requirement 1) and a line in which the magnitude (absolute value) of τ_err coincides with a preset first set value τ_err_a in the phase plane having τ_err and dτ_err as two coordinate axis components (in the phase plane) τ_err = + τ_err_a or τ_err = −τ_err_a) and a line in which the magnitude (absolute value) of dτ_err matches a preset second setting value dτ_err_a (in the phase plane) The response characteristic data satisfying the requirement that the trajectory intersects one of the lines (hereinafter, referred to as selection requirement 2) of dτ_err = + dτ_err_a or dτ_err = −dτ_err_a).

  In other words, the selection requirement 1 is that the values of τ_err and dτ_err at any point on the trajectory indicated by the response characteristic data satisfy the condition that −τ_err_lim ≦ τ_err ≦ + τ_err_lim and −τ_err_lim ≦ dτ_err ≦ dτ_err_lim. It is a requirement.

  The selection requirement 2 is, in other words, a requirement that there is at least one point where | τ_err | = | τ_err_a | or | dτ_err | = | dτ_err_a | on the trajectory indicated by the response characteristic data.

  Specifically, in the procedure 2, first, a first allowable limit value τ_err_lim (> 0) and a second allowable limit value dτ_err_lim (> 0) related to the selection requirement 1 are determined. (Procedure 2-1).

  The first allowable limit value τ_err_lim means the allowable limit value of the value of the secondary torque deviation τ_err, and the second allowable limit value dτ_err_lim is the allowable limit of the value of the secondary torque deviation speed dτ_err. Mean value. .

  Here, if the magnitude of τ_err is too large, the rotational angular velocity or the rotational angle of the drive pulley 2 (primary element) when the output torque of the power source motor 5 is controlled so as to converge the τ_err to zero. The acceleration may exceed the respective allowable limit value. The first allowable limit value τ_err_lim is a limit value of the magnitude of τ_err for preventing the rotational angular velocity and rotational angular acceleration of the drive pulley 2 (primary element) from exceeding the respective allowable limit values.

  The permissible limit values of the rotational angular velocity and the rotational angular acceleration of the drive pulley 2 are determined by design in advance under conditions such as the capability of the power source motor 5 or the mechanical constraints of the power transmission device 1. Value. The value may allow for some margin with respect to actual tolerance limits.

  In the present embodiment, the first permissible limit value τ_err_lim includes the permissible limit value ω1_lim of the rotational angular velocity of the drive pulley 2 (hereinafter referred to as the primary side speed limit value ω1_lim) and the permissible limit value dω1_lim of the rotational angular acceleration of the drive pulley 2 ( Hereinafter, it is determined as follows according to the primary acceleration limit value dω1_lim) and the stiffness characteristic coefficient Ksp of the elastic force generation mechanism 4.

  The amount of change in the inter-pulley rotation angle difference Δθ required to converge τ_err to zero from the initial state where the secondary torque deviation τ_err is a certain value τ_err_0 (≠ 0) is τ_err_0 / Ksp.

  Here, as shown in FIG. 7C, the time required to change the inter-pulley rotation angle difference by τ_err — 0 / Ksp is set to ta. Then, for example, as shown in FIG. 7A, in the first half period (period from 0 to ta / 2) of the time ta, driving is performed with a positive rotational angular acceleration having the magnitude of the primary acceleration limit value dω1_lim. The pulley 2 is rotated relative to the driven pulley 3 so as to increase the speed. Subsequently, in the second half of the time ta (period from ta / 2 to ta), a negative value having a primary acceleration limit value dω1_lim is obtained. A case is assumed in which the drive pulley 2 is rotated relative to the driven pulley 3 so as to decelerate at a rotational angular acceleration in the direction of.

  In other words, the rotational angular velocity of the driving pulley 2 is sequentially increased and decelerated at the rotational angular acceleration of the primary acceleration limit value dω1_lim, and the driving pulley 2 is moved relative to the driven pulley 3 by the rotational angle of τ_err_0 / Ksp. It is assumed that the relative rotation of the drive pulley 2 with respect to the driven pulley 3 is stopped when the relative rotation angle of the drive pulley 2 with respect to the driven pulley 3 reaches τ_err_0 / Ksp.

  In this case, the relationship among τ_err_0 / Ksp, dω1_lim, and ta is expressed by the following equation (26).

(1/4) · dω1_lim · ta ² = τ_err_0 / Ksp (26)

In this case, the maximum rotational angular velocity of the drive pulley 2 is (1/2) · dω1_lim · ta as shown in FIG. 7B, so the condition of the following equation (27) needs to be satisfied. .

(1/2) ・ dω1_lim ・ ta ≦ ω1_lim (27)

From the above equations (26) and (27), the condition of the following equation (28) regarding the magnitude (absolute value) of τ_err_0 is obtained.

| Τ_err_0 | ≦ (ω1_lim ² / dω1_lim) · Ksp (28)

Therefore, in the present embodiment, the first allowable limit value τ_err_lim, which is the allowable limit value of the value of the secondary torque deviation τ_err, is determined by the following equation (29).

τ_err_lim = (ω1_lim ² / dω1_lim) · Ksp (29)

In this case, the target stiffness characteristic coefficient Ksp_cmd (a constant value determined in advance in the present embodiment) is used as the value of Ksp.

  Τ_err_lim determined in this way becomes a larger value as the stiffness characteristic coefficient Ksp is larger. Note that τ_err_lim may be a value slightly smaller than the value determined by Expression (29).

  Further, if the secondary torque deviation speed dτ_err is too large, the driven pulley from the electric motor 5 is controlled when the output torque of the electric motor 5 is controlled to converge the secondary torque deviation τ_err to zero. Due to the resonance phenomenon caused by the natural vibration of the power transmission system 3, oscillation of the power transmission system is likely to occur. The second allowable limit value dτ_err_lim is a limit value of the magnitude of dτ_err for preventing the occurrence of such oscillation.

  In this case, if the natural frequency of the power transmission system from the electric motor 5 to the driven pulley 3 (the natural frequency in the dimension of the angular frequency) is ω_vb, the rotation angle of the drive pulley 2 is driven by this natural frequency ω_vb. The secondary side torque change speed dτ_vb when vibrating relative to the pulley 3 is given by the following equation (30).

dτ_vb = Ksp · ωvb (30)

Therefore, in the present embodiment, the second allowable limit value dτ_err_lim, which is the allowable limit value of the magnitude of the secondary torque deviation speed dτ_err, is determined by the following equation (31).

dτ_err_lim = Ksp · ωvb (31)

The dτ_err_lim determined in this way becomes a larger value as the stiffness characteristic coefficient Ksp is larger.

  In this case, the target stiffness characteristic coefficient Ksp_cmd (predetermined constant value in the present embodiment) is used as the value of Ksp used for the calculation of the right side of Expression (31).

  The value of the natural frequency ωvb can be specified based on, for example, experiments and measurements. Alternatively, the value of ωvb may be approximately determined by the following equation (32).

ωvb = sqrt ((Ksp / (Iin + Iout)) (32)

In the above equation (32), sqrt () is a square root function. Further, Iin and Iout are the inertia of the system on the side of the driving pulley 2 and the inertia of the system on the side of the driven pulley 3 shown in the above formula (10), respectively.

  Note that the second allowable limit value dτ_err_lim may be determined to be slightly smaller than the value determined by the equation (31).

  In the phase plane of FIG. 6, examples of positive and negative lines defined by the first allowable limit value τ_err_lim (lines indicated by the expression τ_err = + τ_err_lim and τ_err = −τ_err_lim, respectively) are shown as lines L1p, Examples of positive and negative lines defined by the second allowable limit value dτ_err_lim (lines indicated by dτ_err = + dτ_err_lim and dτ_err = −dτ_err_lim, respectively) are lines L2p and L2n. .

  In the procedure 2, next, the first set value τ_err_a (> 0) and the second set value dτ_err_a (> 0) related to the selection requirement 2 are determined (procedure 2-2).

  The first set value τ_err_a is a value between zero and the first allowable limit value τ_err_lim (a value satisfying 0 <τ_err_a <τ_err_lim) and is determined according to the first allowable limit value τ_err_lim so as not to approach zero. Is done. For example, τ_err_a is determined to be a value Ma1 times τ_err_lim (= Ma1 · τ_err_lim). However, Ma1 is a positive constant smaller than 1, for example, a value in the range of about 0.25 to 0.75.

  Similarly, the second set value dτ_err_a is a value between zero and the second allowable limit value dτ_err_lim (a value satisfying 0 <dτ_err_a <dτ_err_lim), and is set to the second allowable limit value dτ_err_lim so as not to approach zero. Will be decided accordingly. For example, dτ_err_a is determined to be a value Ma2 times dτ_err_lim (= Ma2 · dτ_err_lim). However, Ma2 is a positive constant smaller than 1, for example, a value within a range of about 1/4 to 1/6.

  In the phase plane of FIG. 6, examples of positive and negative lines defined by the first set value τ_err_a (lines indicated by the expressions τ_err = + τ_err_a and τ_err = −τ_err_a, respectively) are shown as lines LL1p and LL1n. Examples of the second set value dτ_err_a (lines indicated by an expression dτ_err = + dτ_err_a and an expression dτ_err = −dτ_err_a, respectively) are lines LL2p and LL2n.

  In the procedure 2, next, the selection requirements 1 and 2 are determined using the first allowable limit value τ_err_lim and the second allowable limit value dτ_err_lim, the first set value τ_err_a and the second set value dτ_err_a determined as described above. The response characteristic data for determining inclination that satisfies is selected (procedure 2-3).

  For example, in the example shown in FIG. 6, the response characteristic data given the reference signs a1 to a6 satisfy the selection requirement 1. On the other hand, the response characteristic data with the reference signs a1 to a5 satisfy the selection requirement 2, but the response characteristic data with the reference sign a6 does not satisfy the selection requirement 2.

  Accordingly, response characteristic data (response characteristic data of a1 to a5) excluding the response characteristic data of a6 is selected as the response characteristic data for determining the inclination.

  As described above, in the procedure 2, the response characteristic data for inclination determination is selected.

  (Procedure 3) Next, the intersection of the trajectory of the response characteristics data for slope determination selected as described above and the lines LL1p, LL1n, LL2p, LL2n indicated by the first set value τ_err_a and the second set value dτ_err_a, respectively. Among them, the inclination of the switching hyperplane σ = 0 is determined using an intersection satisfying a predetermined requirement condition for the intersection (hereinafter referred to as an intersection requirement condition).

  The intersection requirement condition is a condition that the slope of a line connecting the intersection and the origin of the phase plane (point where τ_err = 0, dτ_err = 0) is within a predetermined slope requirement range.

  Here, since the inclination of the switching hyperplane σ = 0 and the convergence time constant Tc on the switching hyperplane σ = 0 correspond one-to-one, the required inclination range is the time of convergence on the switching hyperplane σ = 0. This is the required range of the inclination of the switching hyperplane σ = 0 corresponding to the required range of the constant Tc.

  When constructing a system for controlling the secondary torque τ in the power transmission device 1 to the target value (target secondary torque τ_cmd), normally the convergence of the secondary torque τ to the target value (the convergence of τ_err to zero) There is a design requirement (target) that the convergence time constant Tc, which is the time constant of), should be set to a value within a certain range. Therefore, as an example, the required tilt range is set in correspondence with the required design range of the value of the convergence time constant Tc.

  For example, the design requirement range of the value of the convergence time constant Tc is a range of Tc_L ≦ Tc ≦ Tc_H, and the relationship between the gradient of the switching hyperplane σ = 0 and the convergence time constant Tc is If the relationship is −1 / Tc, the required tilt range may be set as a range of − (1 / Tc_L) ≦ slope ≦ − (1 / Tc_H). When the relationship between the gradient of the switching hyperplane σ = 0 and the convergence time constant Tc is a relationship of gradient = −Tc, the required gradient range is a range of −Tc_H ≦ tilt ≦ −Tc_L. Can be set as

  However, since the rotational angular acceleration of the drive pulley 2 needs to be limited to the primary acceleration limit value dω1_lim or less, the value of the convergence time constant Tc cannot be reduced without limitation.

  In the present embodiment, the torque applied from the power source motor 5 to the drive pulley 2 from the initial state where the magnitude of the secondary torque deviation τ_err matches the first allowable limit value τ_err_lim is the maximum allowable limit. When the specific convergence time as the shortest convergence time constant Tc that can be realized when τ_err is converged to zero by sequentially increasing and decreasing the relative rotational speed of the driving pulley 2 with respect to the driven pulley 3 A constant Tcx is determined in advance. In this case, the specific convergence time constant Tcx is determined in advance for each representative value of the target viscosity characteristic coefficient Kdp_cmd selected in the procedure 1.

  In step 3, the convergence time constant Tc on the switching hyperplane σ = 0 is greater than or equal to the specific convergence time constant Tcx (Tc ≧ Tcx) as a constraint condition (hereinafter referred to as a time constant constraint condition). In the range satisfying this time constant constraint condition, the tilt required range of the switching hyperplane σ = 0 is set.

  The specific convergence time constant Tcx is determined as follows.

  From the initial state where the magnitude of the secondary torque deviation τ_err matches the first allowable limit value τ_err_lim in a state where no viscous force is generated between the pulleys 2, 3 (assuming that Kdp = 0), As shown in FIGS. 7A to 7C, it is assumed that τ_err is converged to zero by increasing and decreasing the rotational angular velocity of the driving pulley 2 with respect to the driven pulley 3 for the same time. In this case, the convergence required time ta, which is the time required to converge τ_err from the first allowable limit value τ_err_lim to zero, is set as tax0, and the convergence time constant Tc corresponding to the required convergence time tax0 is set as Tcx0.

  In this case, the magnitude of the torque applied to the drive pulley 2 during the period of increasing the rotational angular velocity of the drive pulley 2 and the magnitude of the torque applied to the drive pulley 2 during the period of decreasing the rotational angular velocity are the power The maximum allowable torque that can be applied from the source motor 5 to the drive pulley 2 is set.

  The relationship between the convergence required time tax0 and the convergence time constant Tcx0 is approximately expressed by the following equation (33).

Tcx0 = tax0 / 3 (33)

Since tax0 is the value of ta when τ_err_lim is substituted for τ_err_0 in the above equation (26), the following equation (34) is established.

(1/4) · dω1_lim · tax0 ² = τ_err_lim / Ksp (34)

From the above equations (33) and (34), the following equation (35) is obtained.

Tcx0 = (2/3) · sqrt ((τ_err_lim / dω1_lim) / Ksp) (35)

Next, in a state where a viscous force is generated between the pulleys 2 and 3 (a state where Kdp ≠ 0), from an initial state where the magnitude of the secondary torque deviation τ_err matches the first allowable limit value τ_err_lim, It is assumed that τ_err is converged to zero by sequentially increasing and decreasing the rotational angular velocity of the driving pulley 2 with respect to the driven pulley 3.

  Even in this case, the magnitude of the torque applied to the drive pulley 2 during the period of increasing the rotational angular velocity of the drive pulley 2 and the magnitude of the torque applied to the drive pulley 2 during the period of decreasing the rotational angular speed are: The maximum allowable torque that can be applied from the power source motor 5 to the drive pulley 2 is set.

  In this case, if the time period during which the rotational angular velocity of the drive pulley 2 is increased and the time period during which the rotational angular velocity is decelerated are set to half the time of tax0, FIGS. As shown by the broken line in c), the rotational angular acceleration, rotational angular velocity, and inter-pulley rotational angle difference of the drive pulley 2 change, so that τ_err cannot be converged to zero within the time of tax0.

  However, in a state where the viscous force between the pulleys 2 and 3 does not increase so much (a state where Kdp is relatively small), the rotational angular velocity of the driving pulley 2 is indicated by a two-dot chain line in FIGS. 7 (a) and 7 (b). Τ_err can be converged to zero within the time of tax0 by increasing the time of the period during which the speed is increased from tax0 / 2.

  On the other hand, in a state where the viscous force between the pulleys 2 and 3 is relatively large (a state where Kdp is relatively small), τ_err may not be converged to zero within the time of tax0.

  Therefore, in this embodiment, for each representative value of the target viscosity characteristic coefficient Kdp_cmd, τ_err is converged from the first allowable limit value τ_err_lim to zero within the convergence required time tax0 defined by the above equation (34). If it is possible, the convergence time constant Tcx0 calculated by the equation (35) is determined as the specific time constant Tcx corresponding to the representative value.

  Further, for each representative value of the target viscosity characteristic coefficient Kdp_cmd, τ_err cannot be converged from the first allowable limit value τ_err_lim to zero within the convergence required time tax0 defined by the above equation (34). Includes determining a required convergence time tax (> tax0) that allows τ_err to converge to zero from the first allowable limit value τ_err_lim, and a convergence time constant Tc (= tax / 3) corresponding to the required convergence time tax. Is determined as the specific time constant Tcx.

  In this case, whether or not it is possible to converge τ_err from the first allowable limit value τ_err_lim within the period of the convergence required time tax0 defined by the equation (34) is determined as follows. The

  Assume a state in which a torque τ1_max having a maximum allowable magnitude is applied to the drive pulley 2 from the power source motor 5. At this time, with respect to the relative rotational motion of the drive pulley 2 with respect to the driven pulley 3, a dynamic relationship (relationship in a discrete system) of the following equation (36) is established.

I · dΔω (n) = τ1_max−Kdp · Δω (n-1) (36)

Here, I is the inertia of the rotating system including the driving pulley 2, Δω is the relative rotational angular velocity of the driving pulley 2 with respect to the driven pulley 3, as defined by the above equation (12b), and dΔω is the above equation (12c). The rotational angular acceleration of the drive pulley 2 relative to the driven pulley 3 (= time change rate of Δω). The second term on the right side of the equation (36) represents the viscous force torque τdp.

  From this equation (36), the following equation (37) is obtained.

dΔω (n) = dΔω_max−Mu · Δω (n−1) (37)
However,
dΔω_max≡τ1_max / I
Mu≡Kdp / I

Note that dΔω_max coincides with the primary acceleration limit value dω1_lim.

  Further, regarding the relative rotational angular velocity Δω of the driving pulley 2 with respect to the driven pulley 3 and the relative rotational angle Δθ of the driving pulley 2 with respect to the driven pulley 3 (that is, the inter-pulley rotational angle difference Δθ), respectively, The recurrence formulas (38) and (39) are established.

Δω (n) = Δω (n-1) + DT · dΔω (n) (38)
Δθ (n) = Δθ (n-1) + DT · Δω (n) (39)

From Expression (37) and Expression (38), the following Expression (40) is obtained.

Δω (n) = (dΔω_max / Mu) · (1-Edp n) + Edp n · Δω (0) ...... (40)
However, Edp≡1-DT · Mu

Furthermore, the following equation (41) is obtained from the equation (40) and the equation (39).

Δθ (n) = Δθ (0)
+ (DΔω_max / Mu) · n · DT
+ DT · (Δω (0) - (dΔω_max / Mu)) · Edp · (1-Edp n) / (1-Edp)
...... (41)

Here, when Δω (0) = 0 and Δθ (0) = 0, the driving pulley 2 is moved with respect to the driven pulley 3 at a rotational angular acceleration of dΔω_max from the initial time 0 to a time period of N1 · DT. Assume that the drive pulley 2 is rotated relative to the driven pulley 3 at −dΔω_max rotation angular acceleration in the period of time N2 · DT. N1 and N2 are integer values.

  In this case, Δθ (= Δθ (N1 + N2)) when the time period N1 · DT + N2 · DT elapses is given by the following equation (42) by the above equations (40) and (41).

Δθ (N1 + N2) = (dΔω_max / Mu) ・ N1 ・ DT
−DT · (dΔω_max / Mu) · Edp · (1-Edp ^N1 ) / (1-Edp)
-(DΔω_max / Mu) ・ N2 ・ DT
+ DT · (dΔω_max / Mu) · (2-Edp ^N1 ) · Edp · (1-Edp ^N2 ) / (1-Edp)
...... (42)

Here, when (N1 + N2) · DT = tax0 and Δθ (N1 + N2) = τ_err_lim / Ksp, τ_err is the first allowable limit value during the convergence required time tax0 defined by the above equation (34). It will converge from τ_err_lim to zero. Since dΔω_max = dω1_lim, the following equation (43) is obtained from the equation (34).

τ_err_lim / Ksp = (1/4) · dΔω_max · tax0 ² (43)

Accordingly, the following equation (44) is satisfied as a condition for allowing τ_err to converge from the first allowable limit value τ_err_lim to zero during the convergence required time tax0 defined by the equation (34). N1 (<tax0 / DT) exists.

(1 / Mu) ・ N1 ・ DT-DT ・ (1 / Mu) ・ Edp ・ (1-Edp ^N1 ) / (1-Edp)
-(1 / Mu) / (N-N1) / DT
+ DT ・ (1 / Mu) ・ (2-Edp ^N1 ) ・ Edp ・ (1-Edp ^N-N1 ) / (1-Edp)
= (1/4) · tax0 ² …… (44)

Therefore, in the present embodiment, when determining the specific convergence time constant Tcx corresponding to each representative value of the target viscosity characteristic coefficient Kdp_cmd, τ_err is set to the first value within the period of the convergence required time tax0 defined by the equation (34). Judgment as to whether or not it is possible to converge from one allowable limit value τ_err_lim to zero is made by using each representative value of the target viscosity characteristic coefficient Kdp_cmd and the necessary convergence time tax0 defined by the equation (43). This is done by determining whether or not there is N1 that satisfies (44).

  That is, if there is N1 such that Equation (44) holds for one representative value of the target viscosity characteristic coefficient Kdp_cmd, τ_err is set to the first allowable limit value within the period of convergence required time tax0. It is determined that τ_err_lim can be converged to zero. In this case, the convergence time constant Tcx0 determined by the equation (35) corresponding to the convergence required time tax0 is set as the specific time constant Tcx0 corresponding to the one representative value of the target viscosity characteristic coefficient Kdp_cmd. It is determined.

  Further, if there is no N1 that satisfies the formula (44) for one representative value of the target viscosity characteristic coefficient Kdp_cmd, τ_err is set to the first allowable limit value within the period of convergence required time tax0. It is determined that τ_err_lim cannot be converged to zero. In this case, tax0 'is determined in an exploratory manner such that there exists N1 in which the formula obtained by replacing the convergence required time tax0 in formula (44) with the unknown number tax0' exists. In this case, tax0 'is determined to be as short as possible within the time range longer than the convergence required time tax0 according to the equation (34).

  Then, with tax0 ′ determined in this way as the necessary convergence time corresponding to the one representative value of the target viscosity characteristic coefficient Kdp_cmd, the convergence time constant (= tax ′ / 3) corresponding to the necessary convergence time tax0 ′ is specified. It is determined as the convergence time constant Tcx.

  In this embodiment, the convergence time constant Tc on the switching hyperplane σ = 0 is not less than the specific time constant Tcx determined as described above for each representative value of the target viscosity characteristic coefficient Kdp_cmd (Tc ≧ Tcx). )) Is set as the time constant constraint condition, and the tilt required range of the switching hyperplane σ = 0 for each representative value of the target viscosity characteristic coefficient Kdp_cmd is set within the range satisfying the time constant constraint condition.

  In this case, when the lower limit value Tc_L of the design requirement range of the value of the convergence time constant Tc satisfies the time constant constraint condition (when Tc_L ≧ Tcx), the inclination hyperplane σ = 0 is requested. The range is set to a range corresponding to the required range of the convergence time constant Tc such that Tc_L ≦ Tc ≦ Tc_H. For example, if the relationship between the gradient of the switching hyperplane σ = 0 and the convergence time constant Tc is the relationship of gradient = −1 / Tc, the required gradient range is − (1 / Tc_L) ≦ tilt ≦ It is set as a range of − (1 / Tc_H).

  Further, when the lower limit value Tc_L of the design requirement range of the convergence time constant Tc does not satisfy the time constant constraint condition (when Tc_L <Tcx), the inclination requirement range of the switching hyperplane σ = 0. Is set to a range corresponding to the required range of the time constant Tc, Tcx ≦ Tc ≦ Tc_H. For example, when the relationship between the inclination of the switching hyperplane σ = 0 and the time constant Tc is the relationship of inclination = −1 / Tc, the required inclination range is − (1 / Tcx) ≦ inclination ≦ −. It is set as a range of (1 / Tc_H).

  The intersection requirement condition in the procedure 3 is that the slope of the line connecting the intersection and the origin of the phase plane corresponds to one representative value of the target viscosity characteristic coefficient Kdp_cmd selected in the procedure 1 as described above in advance. This is a condition that it is within a predetermined tilt required range.

  In the phase plane of FIG. 6, the lines Lb_max and Lb_min indicated by the two-dot chain lines are examples of the maximum inclination line and the minimum inclination line in the inclination required range set as described above, respectively. Show. In this case, since the relationship of slope = −1 / Tc is established in the phase plane of FIG. 6, the magnitude of the slope of the line Lb_max is 1 / max (Tc_L, Tcx), and the magnitude of the slope of the line Lb_min is 1 / Tc_H.

  In step 3, an intersection satisfying the intersection requirement defined by the inclination requirement range is extracted. For example, in the example shown in FIG. 6, the intersection of the white circles shown in FIG. 8A as an enlarged view of the broken line frame portion A in FIG. 6 and FIG. 8B as an enlarged view of the broken line frame portion B in FIG. Are extracted as intersections satisfying the intersection requirement condition.

  Then, using the set of values of τ_err and dτ_err at the plurality of extracted intersections, an inclination that matches or approximates the inclination of the line connecting each intersection and the origin as much as possible is calculated by the method of least squares. Is determined as an appropriate inclination of the switching hyperplane σ = 0 corresponding to the representative value being selected for the target viscosity characteristic coefficient Kdp_cmd.

  For example, in the example shown in FIG. 6, the slope of the line Lc indicated by the alternate long and short dash line in the figure is determined as an appropriate slope of the switching hyperplane σ = 0.

  Note that the gradient of the switching hyperplane σ = 0 may be determined by a statistical identification method other than the method of least squares. For example, the average value of the slopes of the lines connecting the intersections satisfying the intersection requirement condition and the origin may be determined as the slope of the switching hyperplane σ = 0.

  In step 3, the convergence time constant Tc corresponding to the inclination of the switching hyperplane σ = 0 (the convergence time constant Tc corresponding to each representative value of the target viscosity characteristic coefficient Kdp_cmd) may be determined.

  The processes of Procedure 1 to Procedure 3 described above are executed for each of all representative values of the target viscosity characteristic coefficient Kdp_cmd. As a result, an appropriate gradient of the switching hyperplane σ = 0 corresponding to each of all the representative values of the target viscosity characteristic coefficient Kdp_cmd (or an appropriate convergence time constant Tc on the switching hyperplane σ = 0) is determined.

  (Procedure 4) Next, a set of each representative value of the target viscosity characteristic coefficient Kdp_cmd and a convergence time constant Tc corresponding to the inclination of the switching hyperplane σ = 0 determined as described above corresponding to the representative value ( The values of the variables a and b in the correlation of the equation (22) are determined by using, for example, the least square method, using the number of sets of representative values. As a result, the correlation between the target viscosity characteristic coefficient Kdp_cmd and the convergence time constant Tc on the switching hyperplane σ = 0, and hence, between the target viscosity characteristic coefficient Kdp_cmd and the inclination of the switching hyperplane σ = 0. The correlation is finally determined.

  The preparatory process for determining the correlation between the gradient of the switching hyperplane σ = 0 (or the convergence time constant Tc on the switching hyperplane σ = 0) and the target viscosity characteristic coefficient Kdp_cmd has been described above.

  In the present embodiment, the inclination requirement range that defines the intersection requirement condition in step 3 is set so as to reflect the time constant restriction condition. However, the inclination requirement range includes the time constant restriction condition. Without consideration, it may be determined simply according to the design requirement of the convergence time constant Tc.

  In that case, the inclination of the switching hyperplane σ = 0 is tentatively determined by a method such as the least square method based on the intersection satisfying the intersection requirement condition defined by this inclination requirement range. When the value of the convergence time constant Tc defined by the provisional inclination satisfies the time constant constraint condition, the provisional inclination is determined as it is as the inclination of the switching hyperplane σ = 0. May be.

  In addition, when the value of the convergence time constant Tc defined by the provisional inclination does not satisfy the time constant restriction condition, the intersection requirement condition defined by the inclination requirement range reflecting the time constant restriction condition as described above is set. The inclination of the switching hyperplane σ = 0 may be determined using only the intersections that satisfy the condition.

  Next, the operation of the control device 30 when the output torque of the power source motor 5 is controlled so that the secondary torque τ of the power transmission device 1 follows the secondary target torque τ_cmd will be described.

  In the following description, the subscript “_act” is attached to a reference sign indicating an actual value of an arbitrary state quantity (angle, torque, etc.) or an observed value (measured value or estimated value).

  In the present embodiment, the control device 30 sequentially determines the target torque τm_cmd of the power source motor 5 as a control input by executing the processing shown by the block diagram of FIG.

  In each control processing cycle of the control device 30, the control input determination unit 34 has the actual rotation angle θin_act (measured value) of the driving pulley 2 indicated by the output signals of the angle detectors 31 and 32, and the driven pulley 3. The actual rotation angle θout_act (measured value) is sequentially input, and the target secondary torque τ_cmd is sequentially input.

  In this situation, the actual rotation angle θw_act of the rotation drive shaft 18a of the stiffness variable motor 18 (and hence the phase angle φ of the rotation bar 14 in the reference state) is the target stiffness characteristic coefficient Ksp_cmd (in this embodiment). The rotation of the rotary drive shaft 18a is stopped in this state.

  Then, the control input determination unit 34 uses the secondary torque measurement unit 34f to determine the values of θin_act and θout_act (current values) in the current control processing cycle and the value of the actual stiffness characteristic coefficient Ksp of the elastic force generation mechanism 4. The measured value (estimated value) of the actual secondary torque τ_act is calculated by calculating the right side of the equation (13a) using the value of the target stiffness characteristic coefficient Ksp_cmd.

  Note that since the value of the target stiffness characteristic coefficient Ksp_cmd is a constant value in the present embodiment, it may be stored in advance in a memory (not shown).

  The control input determination unit 34 corrects τ_cmd by adding the output value of the low-pass filter 34b to the input target secondary torque τ_cmd (current value) by the calculation unit 34a. Hereinafter, this corrected τ_cmd is referred to as a corrected target secondary torque τ_cmd_c.

  This correction is a correction for compensating for the influence of the offset component (steady error component) included in the secondary torque τ_act (measured value). In this case, the output value of the calculation result of the calculation unit 34c is sequentially input to the low-pass filter 34b.

  The calculation unit 34c uses the corrected target secondary torque τ_cmd_c value (previous value) calculated by the calculation unit 34a in the previous control processing cycle and the secondary torque measurement unit 34f in the current control processing cycle. The difference in the measured value (current value) of the calculated secondary side torque τ_act is calculated.

  Then, the low-pass filter 34b extracts the offset component by performing a low-pass characteristic filtering process on the output value of the calculation result of the calculation unit 34c, and outputs the offset component to the calculation unit 34a.

  Next, the control input determination unit 34 calculates the value of the corrected target secondary torque τ_cmd_c calculated by the calculation unit 34a (current value) and the measured value of the secondary torque τ_act calculated by the secondary torque measurement unit 34f ( The deviation from the present value) is calculated by the computing unit 34d as the actual secondary torque deviation τ_err_act.

  Then, the control input determination unit 34 sequentially inputs the secondary side torque deviation τ_err_act to the sliding mode control processing unit 34e.

  The sliding mode control processing unit 34e includes a function as a switching hyperplane variable setting unit 34g that determines the inclination of the switching hyperplane σ = 0 (or the convergence time constant Tc on the switching hyperplane σ = 0). Therefore, the target viscosity characteristic coefficient Kdp_cmd is sequentially input to the sliding mode control processing unit 34e in addition to the secondary torque deviation τ_err.

  Then, the switching hyperplane variable setting unit 34g of the sliding mode control processing unit 34e uses the input target viscosity characteristic coefficient Kdp_cmd (current value) to execute the calculation of the right side of the equation (22), thereby A convergence time constant Tc corresponding to the target viscosity characteristic coefficient Kdp_cmd is determined. Further, the switching hyperplane variable setting unit 34g determines the inclination of the switching hyperplane σ = 0 corresponding to the convergence time constant Tc.

  The switching hyperplane variable setting unit 34g further determines the values of the coefficient components s1, s2 of the switching function σ. Specifically, the values of s1 and s2 are set so that the relationship of the above equation (19) (the relationship of Tc = s2 / s1) is satisfied with respect to the convergence time constant Tc corresponding to the switching hyperplane σ = 0. It is determined. In this case, one value of s1 and s2 may be an arbitrary constant value (for example, 1). Therefore, for example, s1 = 1, s2 = Tc, or s1 = 1 / Tc, s2 = 1.

  The sliding mode control processing unit 34e calculates the value of the switching function σ according to the equation (18), and further performs the calculation of the right side of the equation (17) using the calculated value of the switching function σ. The target torque τm_cmd of the power source motor 5 as an input is sequentially calculated.

  In this case, the value of the secondary torque deviation dτ_err necessary for the calculation of Expression (18) is τ_err_act input from the calculation unit 34d, and the value of the secondary torque deviation speed dτ_err is the rate of change with time of τ_err_act. Is a value calculated as Also, the values of the coefficient components s1, s2 are values determined as described above by the switching hyperplane variable setting unit 34g.

  Further, each component of the matrix A and the column vector B necessary for the calculation of the equation (17) is a value determined based on the proviso definition equation of the equation (15). Since these values are constant values in the present embodiment, they are stored and held in advance in a memory (not shown).

  Further, in this embodiment, the value of Ksld necessary for the calculation of Expression (17) is determined according to the calculated value of the switching function σ by Expression (25). The value of δ is a constant value determined in advance and is stored and held in a memory (not shown).

  The control input determination unit 34 sequentially determines the target torque τm_cmd (control input) of the power source motor 5 by the above processing.

  When the offset component (steady error component) included in the secondary torque τ_act is sufficiently small, the calculation units 34a and 34c and the low-pass filter 34b are omitted, and τ_cmd is input to the calculation unit 34d as it is. It may be.

  The control device 30 inputs the target torque τm_cmd sequentially determined by the control input determination unit 34 as described above to the motor control unit 35 and executes processing of the motor control unit 35. The motor control unit 35 determines a command value (target value) of an energization current of an armature winding (not shown) of the power source motor 5 according to the input target torque τm_cmd, and sets the actual energization current as the command value. The energization current of the armature winding is feedback controlled so as to match.

  Thereby, the actual output torque of the power source motor 5 is controlled to the target torque τm_cmd. As a result, the actual secondary torque τ_act is controlled to follow the target secondary torque τ_cmd.

  According to the first embodiment described above, the control input determination unit 34 of the control device 30 uses the secondary side to determine the control input (target torque τm_cmd of the power source motor 5) by the sliding mode control method. The measured value of the torque τ is sequentially acquired by the secondary side torque measuring unit 34f based on the formula (13a). Since this measured value corresponds to the measured value of the torque (elastic force torque τsp) due to the elastic force generated by the elastic force generating mechanism 4, the relative rotation Δθ between the driving pulley 2 and the driven pulley 3 is measured. The value can be obtained with high reliability.

  In this case, the measured value of the secondary side torque τ and the temporal change rate thereof do not include the torque component due to the viscous force generated by the elastic force generating mechanism 4, but the switching hyperplane for sliding mode control. The slope of σ = 0 (or the convergence time constant Tc on the switching hyperplane σ = 0) is variably determined according to the target viscosity characteristic coefficient Kdp_cmd as the control value of the viscosity characteristic coefficient Kdp of the elastic force generation mechanism 4. The

  For this reason, regardless of the control state of the viscosity characteristic coefficient of the elastic force generating mechanism 4, the influence of the viscous force generated by the elastic force generating mechanism 4 is appropriately compensated for, and the vibration of the measured value of the secondary torque τ, etc. It is possible to stably converge the actual secondary torque τ to the secondary target torque τ_cmd with high robustness while suppressing the occurrence of.

  Further, the correlation between the gradient of the switching hyperplane σ = 0 for sliding mode control (or the convergence time constant Tc) and the target viscosity characteristic coefficient Kdp_cmd is determined for each of a plurality of representative values of Kdp_cmd, and so on. Among the plurality of response characteristic data created by using the general-purpose control method, the inclination determination response characteristic data satisfying the selection requirements 1 and 2 is used.

  In this case, the response characteristic data for determining the inclination satisfy the selection requirement 1 so that the rotational angular acceleration and the rotational angular velocity of the drive pulley 2 are the primary acceleration limit values dω1_lim, which are the respective allowable limit values, and This is appropriate data that can satisfy the condition that the primary speed limit value ω1_lim is not exceeded.

  In addition, the response characteristic data trajectory often causes vibration in the region near the origin of the phase plane. However, the response characteristic data for tilt determination satisfies the selection requirement 2 so that the response characteristic data near zero is near zero. This is data including a set of values of the secondary torque deviation τ_err and the secondary torque deviation speed dτ_err that are not very small.

  The line LL1p, LL1n where the value of τ_err is the first set value τ_err_a and the line LL2p, LL2n where the value of dτ_err is the second set value dτ_err_a; Based on the intersection satisfying the intersection requirement condition corresponding to the required range of the convergence time constant Tc among the intersection points with the trajectory of the response characteristic data for slope determination satisfying the selection requirements 1 and 2, by a method such as the least square method The slope of the switching hyperplane σ = 0 (or the convergence time constant Tc) is determined.

  In this case, the magnitude of the value of the secondary torque deviation τ_err (= first set value τ_err_a) at the intersection is smaller than the first allowable limit value τ_err_lim and has a magnitude that is not too close to zero. Have. Similarly, the magnitude of the value of the secondary torque deviation speed dτ_err (= second set value τ_err_a) at the intersection is smaller than the second allowable limit value dτ_err_lim, and is a certain level that does not approach zero too much. Have

  In addition, since the intersection satisfies the intersection requirement condition, it satisfies the design requirement regarding the convergence time constant Tc and the time constant constraint condition.

  Therefore, the intersection used to determine the inclination of the switching hyperplane σ = 0 for each representative value of the target viscosity characteristic coefficient Kdp_cmd is on the trajectory of the response characteristic data that can suitably converge the secondary torque deviation τ_err to zero. As a point, it becomes a thing with high reliability and stability. Therefore, by determining the slope of the switching hyperplane σ = 0 for each representative value of the target viscosity characteristic coefficient Kdp_cmd using the intersection point, the secondary torque deviation τ_err can be suitably converged to zero. The inclination will be determined.

  Each representative value of the target viscosity characteristic coefficient Kdp of the elastic force generating mechanism 4 and the inclination of the switching hyperplane σ = 0 determined corresponding to each value (or the convergence time constant Tc on the switching hyperplane σ = 0) Based on the above, the correlation between the viscosity characteristic coefficient Kdp and the gradient of the switching hyperplane σ = 0 (or the convergence time constant Tc on the switching hyperplane σ = 0) (in the present embodiment, the equation (22) ) Is specified.

  For this reason, the said correlation can be specified suitably according to the actual behavior characteristic of the power transmission device 1.

  In addition, since response characteristic data may be obtained using a general-purpose control method, a large number of response characteristic data can be collected efficiently and easily. Consequently, a suitable gradient of the switching hyperplane σ = 0 for each representative value of the target viscosity characteristic coefficient Kdp_cmd is specified, and further, the gradient of the viscosity characteristic coefficient Kdp and the switching hyperplane σ = 0 (or the switching hyperplane σ = 0). It is possible to efficiently identify a suitable correlation with the above convergence time constant Tc).

  In the process of the control input determination unit 34 of the control device 30, the gradient of the switching hyperplane σ = 0 (and the values of the coefficient components s1, s2) is determined based on the correlation specified as described above. . Then, the control input determination unit 34 uses the switching function σ defined corresponding to the inclination to perform the control input (the target torque τm_cmd of the electric motor 5) by the sliding mode control process (the calculation process of the equation (17)). ).

  For this reason, it is possible to suitably realize the control of the secondary side torque τ with the required convergence characteristic and high robustness regardless of the control state of the viscosity characteristic coefficient of the elastic force generating mechanism 4.

  In particular, in this embodiment, the gradient of the switching hyperplane σ = 0 (or the convergence time constant Tc on the switching hyperplane σ = 0) is set so as to satisfy the time constant constraint condition. The secondary torque τ can be stably controlled with high robustness in a wide operating range.

  Here, the correspondence relationship between the first embodiment described above and the present invention will be supplemented. The first embodiment is an embodiment related to the first invention, the second invention, the fourth to seventh inventions, and the eleventh invention. In this case, the drive pulley 2, the driven pulley 3, the elastic force generation mechanism 4, and the electric motor 5 correspond to the primary side element, the secondary side element, the elastic force generation mechanism, and the actuator in the present invention, respectively. The rotation of the drive pulley 2 corresponds to the displacement of the primary side element in the present invention, the secondary side torque τ corresponds to the secondary side power in the present invention, and the stiffness characteristic coefficient Ksp corresponds to the stiffness characteristic coefficient in the present invention. The viscosity characteristic coefficient Kdp corresponds to the viscosity characteristic coefficient in the present invention.

  Further, the control input determining unit 34, the secondary side torque measuring unit 34f, and the switching hyperplane variable setting unit 34g of the control device 30 are respectively a control input determining unit, a secondary side power measuring unit, and a switching hyperplane variable setting in the present invention. Corresponds to means.

  The secondary torque deviation τ_err and the secondary torque deviation speed dτ_err correspond to the first variable component and the second variable component in the present invention, respectively. Also, the first allowable limit value τ_err_lim, the second allowable limit value dτ_err_lim, the first set value τ_err_a, and the second set value dτ_err_a in the present embodiment are respectively the first allowable limit value, the second allowable limit value, This corresponds to the first set value and the second set value.

  Furthermore, the lines LL1p and LL1n correspond to the first set value line in the present invention, and the lines LL2p and LL2n correspond to the second set value line in the present invention. Further, the specific time constant Tcx corresponds to the specific time constant in the present invention.

  Moreover, said Formula (22) corresponds to Formula (A) in this invention. Further, the first term on the right side of Equation (17) corresponds to the first control input component represented by Equation (C) in the present invention, and the second term on the right side of Equation (17) is the second control in the present invention. Corresponds to the input component.

[Second Embodiment]
Next, a second embodiment of the present invention will be described with reference to FIG. Note that this embodiment is different from the first embodiment only in a part of the process of the control input determination unit 34 of the control device 30. Therefore, the description of the present embodiment will be focused on matters that are different from those of the first embodiment, and description of the same matters as those of the first embodiment will be omitted.

  In the present embodiment, in the process of the control input determination unit 34, in order to reduce the influence of disturbance, the values of the secondary side torque deviation τ_err and the secondary side torque deviation speed dτ_err are sequentially estimated using an observer. Then, instead of the secondary side torque deviation τ_err calculated using the measured value of the secondary side torque measuring unit 34f as it is and the secondary side torque deviation speed dτ_err obtained as the rate of change over time, estimation by an observer is performed. The sliding mode control process is executed using the values τ_err_hat and dτ_err_hat to determine the control input (target torque τm_cmd).

  Specifically, in the present embodiment, as shown in FIG. 10, the sliding mode control processing unit 34e of the control input determining unit 34 further includes a function as an observer 34h.

  Then, the sliding mode control processing unit 34e has a deviation between the measured value of the secondary side torque τ_act calculated by the secondary side torque measuring unit 34f and the corrected target secondary side torque τ_cmd_c (of the calculating unit 34d). Output) is input as the actual secondary torque deviation τ_err_act. Note that the calculation process of the corrected target secondary torque τ_cmd_c is the same as that in the first embodiment.

  The observer 34h calculates the estimated value τ_err_hat of the secondary torque deviation τ_err (hereinafter referred to as the secondary torque deviation estimated value τ_err_hat) with the disturbance component reduced by the calculation of the following equation (51) and the secondary with the disturbance component reduced. An estimated value dτ_err_hat (hereinafter referred to as a secondary torque deviation speed estimated value dτ_err_hat) of the side torque deviation speed dτ_err is calculated while being sequentially updated at a control processing cycle of the control device 30.

  A and B in the equation (51) are a matrix (2 rows and 2 columns) and a column vector (2 rows and 1 column) defined by the proviso of the equation (15), respectively. As in the first embodiment, it is determined according to the definition of the proviso in Expression (15).

  In this case, the value (previous value) of the target torque Tm_cmd determined in the previous control processing cycle is used as the value of u (n-1) on the right side of the equation (51). Kobs is a predetermined default value and is stored and held in a memory (not shown) of the control device 30.

  Further, dτ_err_act_filt is a value obtained by performing low-pass characteristic filtering processing on the temporal change rate value of the secondary torque deviation τ_err_act input to the sliding mode control processing unit 34e.

  Note that a value obtained by removing the noise component from the value of the temporal change rate of the estimated value of the secondary torque deviation τ_err_act may be used instead of dτ_err_act_filt by an appropriate process other than the low-pass characteristic filtering process.

  The sliding mode control processing unit 34e according to the present embodiment uses the secondary side torque deviation estimated value τ_err_hat and the secondary side torque deviation speed estimated value dτ_err_hat calculated by the observer 34h as described above to obtain the equation (18). Then, the value of the switching function σ is calculated, and the calculation of the right side of the equation (17) is further performed using the calculated value of the switching function σ, so that the target torque τm_cmd of the power source motor 5 as the control input is sequentially obtained. calculate.

  The present embodiment is the same as the first embodiment except for the matters described above.

  Also in this embodiment, the same operational effects as in the first embodiment can be achieved.

  In addition, in the present embodiment, the secondary side torque deviation estimated value τ_err_hat and the secondary side torque deviation speed estimated value dτ_err_hat calculated by the observer 34h are sequentially calculated by the secondary side torque measuring unit 34f. Since the measured value of the torque τ_act is used in place of the actual secondary torque deviation τ_err_act calculated using the measured value and the secondary torque deviation speed dτ_err_act, which is the rate of change over time, the sliding mode control process is performed. The control input (target torque τm_cmd) can be determined by reducing the influence of the disturbance component included in the measured value of the secondary side torque τ_act and its temporal change rate.

  For this reason, the robustness of the control of the secondary torque τ by the control device 30 can be further enhanced.

  Here, the correspondence relationship between the second embodiment described above and the present invention will be supplemented. The second embodiment is an embodiment related to the first invention, the second invention, the fourth to seventh inventions, the eleventh invention, and the twelfth invention. In this case, in the second embodiment, the observer 34h corresponds to the observer in the present invention. Except for this, the correspondence between the second embodiment and the present invention is the same as that of the first embodiment.

[Third Embodiment]
Next, a third embodiment of the present invention will be described with reference to FIGS. Note that this embodiment is different from the first embodiment only in part of the control processing of the power transmission device. For this reason, the description of the present embodiment will be focused on matters that are different from the first embodiment, and the description of the same matters as the first embodiment will be omitted.

  In the first embodiment, the target stiffness characteristic coefficient Ksp_cmd of the elastic force generation mechanism 4 is kept constant. On the other hand, in this embodiment, the target stiffness characteristic coefficient Ksp_cmd is variably set within a predetermined range, and the target stiffness characteristic coefficient Ksp_cmd is sequentially input to the control device 30. .

  Then, the rigidity control unit 36 of the control device 30 has a variable stiffness motor 18 for setting the phase angle φ of the rotating bar 14 in the reference state of the elastic force generating mechanism 4 to an angle value corresponding to the target stiffness characteristic coefficient Ksp_cmd. The target value of the rotation angle of the rotation drive shaft 18a is determined from Ksp_cmd by a predetermined map or arithmetic expression. Then, the rigidity control unit 36 controls the actual rotation angle of the rotation drive shaft 18a (observed value indicated by the output of the angle detector 33) to the target value determined from Ksp_cmd by servo control.

  At this time, the rotation angle of the rotation drive shaft 18a of the stiffness variable motor 18 is controlled to change following the change in the target stiffness characteristic coefficient Ksp_cmd. Thereby, the actual stiffness characteristic coefficient Ksp of the elastic force generating mechanism 4 is variably controlled.

  In the present embodiment, as shown in FIG. 11, the target stiffness characteristic coefficient Ksp_cmd is sequentially input to the sliding mode control processing section 34e of the control input determination section 34 in addition to the target viscosity characteristic coefficient Kdp_cmd. The switching hyperplane variable setting unit 34g of the sliding mode control processing unit 34e is an elastic force generation mechanism that is not reflected in the measured value based on the equations (13a) and (13b) and the state equation of the equation (15). Target viscosity characteristic coefficient Kdp_cmd as a control value (value realized by control) of the viscosity characteristic coefficient Kdp of the elastic force generation mechanism 4 in order to compensate for the influence of the viscosity force 4 (viscosity between both pulleys 2 and 3) And the inclination of the switching hyperplane σ = 0 (or the switching hyperplane σ = 0) according to the target stiffness characteristic coefficient Ksp_cmd as the control value (value realized by the control) of the stiffness characteristic coefficient Ksp of the elastic force generation mechanism 4 The above convergence time constant Tc) is variably determined.

  In this case, the correlation between the inclination of the switching hyperplane σ = 0 (or the convergence time constant Tc on the switching hyperplane σ = 0) and the target viscosity characteristic coefficient Kdp_cmd and the target stiffness characteristic coefficient Ksp_cmd is determined in advance. In accordance with the correlation, the slope of the switching hyperplane σ = 0 (or the convergence time constant Tc on the switching hyperplane σ = 0) is variably determined.

  In the present embodiment, the switching hyperplane variable setting unit 34g changes the slope of the switching hyperplane σ = 0 determined according to Kdp_cmd and Ksp_cmd (or the convergence time constant Tc on the switching hyperplane σ = 0). As in the first embodiment, the values of the coefficient components s1, s2 of the switching function σ are determined. Then, the sliding mode control processing unit 34e uses the values of the coefficient components s1 and s2 to determine the control input τm_cmd according to the equations (17) and (18) as in the first embodiment.

  Here, according to various experiments and examinations of the inventors of the present application, when the stiffness characteristic coefficient Ksp of the elastic force generation mechanism 4 is variably controlled, τ_err and dτ_err when τ_err converges to zero stably Is approximated by the relationship of the following formula (61).

τ_err =-((a2 / sqrt (Ksp_cmd) + a1 ・ Ksp_cmd + a0)
+ (B · Kdp_cmd) / Ksp_cmd) · dτ_err (61)

Therefore, in the present embodiment, the slope on the switching hyperplane σ = 0 (or the switching hyperplane so that the relationship between τ_err and dτ_err on the switching hyperplane σ = 0 becomes the relationship of Expression (61). The convergence time constant Tc) over σ = 0 is determined according to the target viscosity characteristic coefficient Kdp_cmd and the target stiffness characteristic coefficient Ksp_cmd.

  In other words, as the correlation between the convergence time constant Tc on the switching hyperplane σ = 0 and Kdp_cmd and Ksp_cmd is expressed by the following equation (62), the slope of the switching hyperplane σ = 0 ( Alternatively, the convergence time constant Tc) on the switching hyperplane σ = 0 is determined according to the target viscosity characteristic coefficient Kdp_cmd and the target stiffness characteristic coefficient Ksp_cmd.

Tc = (a2 / sqrt (Ksp_cmd) + a1 ・ Ksp_cmd + a0)
+ (B · Kdp_cmd) / Ksp_cmd (62)

(A2 / sqrt (Ksp_cmd) + a1 · Ksp_cmd + a0) in the equations (61) and (62) is a switching hyperplane when the viscosity characteristic coefficient Kdp of the elastic force generation mechanism 4 is zero (when Kdp_cmd = 0). It means the value of the convergence time constant Tc on σ = 0. Therefore, when the viscosity characteristic coefficient Kdp of the elastic force generation mechanism 4 is zero, the convergence time constant Tc on the switching hyperplane σ = 0 changes according to the target stiffness characteristic coefficient Ksp_cmd.

  Supplementally, in the first embodiment, since Ksp_cmd is a constant value, (a2 / sqrt (Ksp_cmd) + a1 · Ksp_cmd + a0) is a constant value, and this value is the value of the variable a in the equation (22). It will be equivalent.

  Further, the variable b in the equations (61) and (62) is the target viscosity characteristic coefficient Kdp_cmd, as in the variable b of the equation (22) in the first embodiment. It means a correction coefficient for correcting to an estimated value of the coefficient Kdp.

  Also in this embodiment, the relationship between the viscous force torque τdp and the temporal change rate dτsp of the elastic force torque τsp is expressed by the equation (23).

  Accordingly, as in the case of the first embodiment, the right side of the equation (62) indicates that the actual viscosity characteristic coefficient Kdp of the elastic force generation mechanism 4 is zero (the viscosity force between the pulleys 2 and 3 is zero). In some cases, it has a meaning as a deviation between the elastic force torque τsp as the actual secondary torque τ and the secondary target torque τ_cmd. When the actual viscosity characteristic coefficient Kdp of the elastic force generating mechanism 4 is not zero, the actual secondary torque τ and the secondary target torque obtained by combining the elastic force torque τsp and the viscous force torque τdp. It has a meaning as a deviation from τ_cmd.

  Hereinafter, in the present embodiment, the correlation between the gradient of the switching hyperplane σ = 0 (or the convergence time constant Tc on the switching hyperplane σ = 0) and the target viscosity characteristic coefficient Kdp_cmd and the target stiffness characteristic coefficient Ksp_cmd. The pre-preparation process for determining the will be described.

  In the preliminary preparation processing in the present embodiment, first, a plurality of representative values of the target viscosity characteristic coefficient Kdp_cmd and a plurality of representative values of the target stiffness characteristic coefficient Ksp_cmd are determined in advance.

  A set of one representative value of the target viscosity characteristic coefficient Kdp_cmd and one representative value of the target stiffness characteristic coefficient Ksp_cmd is selected, and an elastic force generating mechanism is selected according to the selected representative value of the target viscosity characteristic coefficient Kdp_cmd. 4 is controlled, and the rotation angle of the rotational drive shaft 18a of the variable stiffness motor 18 (= rotation angle of the spring worm 17) is controlled according to the selected target stiffness characteristic coefficient Ksp_cmd. The

  In this state, the same processing as in steps 1 to 3 described in the first embodiment is executed, whereby one representative value of the selected target viscosity characteristic coefficient Kdp_cmd and one of the target stiffness characteristic coefficient Ksp_cmd are set. An appropriate slope of the switching hyperplane σ = 0 corresponding to a set of two representative values (or an appropriate convergence time constant Tc on the switching hyperplane σ = 0) is determined.

  In this case, in the process requiring the value of Kdp (for example, the process of calculating the values of Mu and Edp in the equation (44)), the representative viscosity characteristic coefficient Kdp_cmd is selected as the Kdp value. A value is used. Similarly, in a process that requires a value of Ksp (for example, a process of determining the first allowable limit value τ_err_lim by the above equation (29), etc.), the representative value during selection of the target stiffness characteristic coefficient Ksp_cmd is used as the value of Ksp. Used.

  Then, by repeating the same process as the process of steps 1 to 3, the switching hyperplane σ = for each set of the representative value of the target viscosity characteristic coefficient Kdp_cmd and the representative value of the target stiffness characteristic coefficient Ksp_cmd, respectively. An appropriate slope of 0 (or an appropriate convergence time constant Tc on the switching hyperplane σ = 0) is determined.

  Next, a set of each representative value of the target viscosity characteristic coefficient Kdp_cmd and each representative value of the target stiffness characteristic coefficient Ksp_cmd, and a slope of the switching hyperplane σ = 0 determined corresponding to each set of the representative values Using the convergence time constant Tc, the values of the variables a2, a1, a0, b in the correlation of the equation (62) are determined by, for example, the least square method. As a result, the correlation between the target viscosity characteristic coefficient Kdp_cmd, the target rigidity characteristic coefficient Ksp_cmd, and the convergence time constant Tc on the switching hyperplane σ = 0, that is, the target viscosity characteristic coefficient Kdp_cmd, and the target rigidity characteristic coefficient Ksp_cmd And the inclination of the switching hyperplane σ = 0 is finally determined.

  The above is the correlation between the gradient of the switching hyperplane σ = 0 (or the convergence time constant Tc on the switching hyperplane σ = 0) and the target viscosity characteristic coefficient Kdp_cmd and the target stiffness characteristic coefficient Ksp_cmd in the present embodiment. This is a pre-preparation process for determining.

  The processing of the control device 30 in the present embodiment is the same as that in the first embodiment except for the matters described above.

  Therefore, the actual secondary torque τ_act is controlled to follow the target secondary torque τ_cmd while the viscosity characteristic coefficient Kdp and the rigidity characteristic coefficient Ksp of the elastic force generating mechanism 4 are appropriately changed.

  According to the third embodiment described above, the control input determination unit 34 of the control device 30 obtains the measured value of the secondary torque τ based on the formula (13a) as in the first embodiment, and The gradient of the switching hyperplane σ = 0 for sliding mode control (or the convergence time constant Tc on the switching hyperplane σ = 0) is used as the target viscosity characteristic coefficient Kdp_cmd as the control value of the viscosity characteristic coefficient Kdp of the elastic force generating mechanism 4. And a target stiffness characteristic coefficient Ksp_cmd as a control value of the stiffness characteristic coefficient Ksp of the elastic force generation mechanism 4 is variably determined.

  Therefore, regardless of the control state of the viscosity characteristic coefficient and the rigidity characteristic coefficient of the elastic force generation mechanism 4, the influence of the viscous force generated by the elastic force generation mechanism 4 is appropriately compensated for, and the secondary side torque τ is measured. It is possible to stably converge the actual secondary torque τ to the secondary target torque τ_cmd with high robustness while suppressing the occurrence of vibration of the value.

  Further, in the pre-preparation process, the inclination (or the switching hyperplane) of the switching hyperplane σ = 0 corresponding to each set of representative values of the target viscosity characteristic coefficient Kdp and the target stiffness characteristic coefficient Ksp of the elastic force generation mechanism 4. The convergence time constant Tc) on σ = 0 is determined using a plurality of response characteristic data in the same manner as in the first embodiment.

  Further, a set of representative values of the target viscosity characteristic coefficient Kdp and the target rigidity characteristic coefficient Ksp of the elastic force generation mechanism 4 and the inclination of the switching hyperplane σ = 0 determined corresponding to each set (or the switching hyperplane σ Based on the convergence time constant Tc) = 0, the viscosity characteristic coefficient Kdp and the stiffness characteristic coefficient Ksp and the gradient of the switching hyperplane σ = 0 (or the convergence time constant Tc on the switching hyperplane σ = 0) (In this embodiment, the correlation represented by the equation (62)) is specified.

  For this reason, the said correlation can be specified suitably according to the actual behavior characteristic of the power transmission device 1.

  Further, similarly to the first embodiment, the response characteristic data can be efficiently and easily collected using a general-purpose control method. As a result, a suitable inclination of the switching hyperplane σ = 0 for each representative value set of the target viscosity characteristic coefficient Kdp_cmd and the target rigidity characteristic coefficient Ksp_cmd is specified, and the viscosity characteristic coefficient Kdp, the rigidity characteristic coefficient, It is possible to efficiently identify a suitable correlation with the slope of the plane σ = 0 (or the convergence time constant Tc on the switching hyperplane σ = 0).

  In the process of the control input determination unit 34 of the control device 30, the gradient of the switching hyperplane σ = 0 (and the values of the coefficient components s1, s2) is determined based on the correlation specified as described above. . Then, the control input determination unit 34 uses the switching function σ defined corresponding to the inclination to perform the control input (the target torque τm_cmd of the electric motor 5) by the sliding mode control process (the calculation process of the equation (17)). ).

  For this reason, the control of the secondary torque τ is suitably realized with the required convergence characteristics and high robustness regardless of the control state of the viscosity characteristic coefficient Kdp and the rigidity characteristic coefficient Ksp of the elastic force generation mechanism 4. can do.

  In particular, since the gradient of the switching hyperplane σ = 0 (or the convergence time constant Tc on the switching hyperplane σ = 0) is set so as to satisfy the time constant constraint condition, the power transmission device 1 has a wide operating range. The secondary torque τ can be stably controlled with high robustness.

  Here, the correspondence relationship between the third embodiment described above and the present invention will be supplemented. The third embodiment is an embodiment related to the first invention, the third invention, the fourth invention, and the eighth to eleventh inventions. In this case, the formula (62) corresponds to the formula (B) in the present invention. Except for this, the correspondence between the third embodiment and the present invention is the same as in the first embodiment.

  As in the case of the first embodiment, in the third embodiment, when the offset component (steady error component) included in the secondary torque τ_act is sufficiently small, the calculation shown in FIG. The units 34a and 34c and the low-pass filter 34b may be omitted, and τ_cmd may be input to the calculation unit 34d as it is.

  In the third embodiment, similarly to the second embodiment, the sliding mode control processing unit 34e may include an observer 34h.

  In this case, the sliding mode control processing unit 34e calculates the secondary side torque deviation estimated value τ_err_hat and the secondary side torque deviation speed estimated value dτ_err_hat, which are sequentially calculated by the observer 34h according to the equation (51), to the calculating unit 34d. Is used instead of dτ_err_act calculated as the rate of change over time, and the calculation of the above formulas (17) and (18) is performed, so that the target of the power source motor 5 as a control input is obtained. Torque τm_cmd is calculated sequentially. The other processes may be the same as those in the third embodiment. By doing so, another embodiment of the twelfth aspect of the present invention is constructed.

[Fourth Embodiment]
Next, a fourth embodiment of the present invention will be described with reference to FIGS. Note that this embodiment is different from the first embodiment only in part of the processing of the sliding mode control processing unit. Therefore, the description of the present embodiment will be focused on matters that are different from those of the first embodiment, and description of the same matters as those of the first embodiment will be omitted.

  In the first embodiment, as described above, the calculation process for determining the control input by the sliding mode control method assumes that the viscosity characteristic coefficient Kdp of the elastic force generation mechanism 4 is zero, that is, In addition, it is constructed based on a model that expresses the behavior of the power transmission device 1 when it is assumed that no viscous force is generated between the pulleys 2 and 3.

  In order to compensate for the influence of the viscous force generated by the elastic force generating mechanism 4, the gradient of the switching hyperplane σ = 0 (or the convergence time constant Tc on the switching hyperplane σ = 0) is set as the target viscosity characteristic coefficient Kdp_cmd. It was determined variably according to.

  On the other hand, in the present embodiment, the arithmetic processing for determining the control input by the sliding mode control technique expresses the behavior of the power transmission device 1 assuming that the viscosity characteristic coefficient Kdp of the elastic force generation mechanism 4 is not zero. Built based on the model to be. The slope of the switching hyperplane σ = 0 (or the convergence time constant Tc on the switching hyperplane σ = 0) is determined so as not to depend on the target viscosity characteristic coefficient Kdp_cmd, and is used to determine the control input. By changing the value of one parameter (a parameter other than the parameter related to the switching function σ) according to the target viscosity characteristic coefficient Kdp_cmd, the influence of the viscous force generated by the elastic force generating mechanism 4 is compensated.

  Specifically, when the state equation of the equation (10) is rearranged using the equations (13a) and (13b), as a model expressing the behavior regarding the secondary side torque τ and the secondary side torque change rate dτ. The following equation (71) is obtained.

  In the sliding mode control in the present embodiment, a control process for determining a control input is constructed based on the state equation (model) of the equation (71).

  In this case, the expression (71) is an expression in which the matrix A of the expression (15) is replaced with the matrix A defined by the proviso of the expression (71). Therefore, the target torque τm_cmd of the power source motor 5 as a control input by the sliding mode control in the present embodiment is determined by an arithmetic expression having the same format as the expression (17).

  However, in the present embodiment, the coefficient components s1 and s2 of the switching function σ represented by the equation (18) are the slope of the switching hyperplane σ = 0 (or the convergence time constant Tc on the switching hyperplane σ = 0). ) Is determined in advance in a manner different from that of the first embodiment.

  Specifically, in this embodiment, the preliminary preparation process is executed in the same manner as in the first embodiment, so that the target viscosity characteristic coefficient Kdp_cmd of the elastic force generation mechanism 4 and the convergence time constant Tc are set. The values of the variables a and b in the equation (22) that define the correlation are determined.

  Here, in the first embodiment, Tc in Expression (22), that is, Tc as a function value of Kdp_cmd is used as it is as a convergence time constant on the switching hyperplane σ = 0. On the other hand, in this embodiment, Tc in the equation (22) is not set as the convergence time constant on the switching hyperplane σ = 0, and the proper convergence as a suitable convergence time constant for converging τ_err to zero. Considered as a time constant. In the correlation between the value of the variable a in the equation (22), that is, the appropriate time constant Tc defined by the equation (22) and Kdp_cmd, the target viscosity characteristic coefficient Kdp_cmd is assumed to be zero. The value of the proper convergence time constant Tc (= a) was determined as the convergence time constant on the switching hyperplane σ = 0.

  Therefore, in the present embodiment, the inclination of the switching hyperplane σ = 0 is such that the convergence time constant on the switching hyperplane σ = 0 (the convergence time constant on the switching hyperplane) is the value of the variable a determined in advance (this implementation) It is determined to be a constant value). Hereinafter, the convergence time constant on the switching hyperplane σ = 0 in this embodiment will be referred to as Tc0. In this case, the convergence time constant Tc0 is given by the following equation (72).

Tc0 = a (72)

Therefore, the convergence time constant Tc0 on the switching hyperplane σ = 0 in the present embodiment is a constant value that does not depend on the target viscosity characteristic coefficient Kdp_cmd.

  Then, the values of the coefficient components s1, s2 of the switching function σ in the equation (18) are determined in advance so that s2 / s1 matches the convergence time constant Tc0 (= -a) on the switching hyperplane σ = 0. Has been. For example, s1 = 1, s2 = a, or s1 = 1 / a, s2 = 1.

  On the other hand, in the present embodiment, the matrix A used for calculating the target torque τm_cmd of the power source motor 5 as the control input by the equation (17) is a matrix defined by the proviso of the equation (71). The 2 × 2 component of the matrix A is a value that depends on the viscosity characteristic coefficient Kdp.

  Here, as described above in the first embodiment, the value (= b · Kdp_cmd) obtained by multiplying the value of the variable b in the equation (22) by the target viscosity characteristic coefficient Kdp_cmd is an actual value of the elastic force generation mechanism 4. This corresponds to an estimated value of the viscosity characteristic coefficient Kdp.

  Therefore, in this embodiment, b · Kdp_cmd is used as the value of Ksp for calculating the 2 × 2 component of the matrix A. That is, the values of Cin and Cout in the 2 × 2 component of the matrix A are calculated by the following equations (72a) and (72b), respectively. Note that the values of the inertias Iin and Iout are predetermined values that are determined in advance, as in the first embodiment.

Cin = b · Kdp_cmd / Iin (72a)
Cout = b · Kdp_cmd / Iout (72b)

Based on the above, the control input determination unit 34 of the control device 30 in the present embodiment will be described with reference to FIG. As shown in FIG. 12, the control input determination unit 34 in the present embodiment is similar to the first embodiment in that the calculation units 34a, 34c, 34d, the low-pass filter 34b, the secondary side torque measurement unit 34f, and the sliding mode control. A processing unit 34e is provided. However, in the present embodiment, the sliding mode control processing unit 34e does not include the switching hyperplane variable setting unit, and the convergence time constant a on the switching hyperplane σ = 0 corresponding to the inclination of the switching hyperplane σ = 0. Is determined in advance as described above. The values of the coefficient components s1 and s2 of the switching function σ are also determined in advance corresponding to the convergence time constant a, and the values are stored and held in advance in a memory (not shown).

  Then, the measured value of the actual secondary torque deviation τ_err_act sequentially calculated by the calculation unit 34d is sequentially input to the sliding mode control processing unit 34e, and the target viscosity characteristic coefficient Kdp_cmd is sequentially input.

  Then, the sliding mode control processing unit 34e is calculated as the predetermined values (constant values) of the coefficient components s1 and s2, the secondary torque deviation τ_err_act input from the calculation unit 34d, and the temporal change rate thereof. From the secondary torque deviation speed dτ_er_act, the value of the switching function σ is calculated by the calculation of the equation (18).

  Further, the value of the switching function σ, the values of coefficient components s1 and s2 stored in advance in a memory (not shown), the matrix A and the column vector B defined by the proviso of the equation (71), τ_err_act and dτ_er_act Is used to calculate the target torque τm_cmd of the power source motor 5 as a control input.

  In this case, the 2 × 2 component of the matrix A is calculated using b · Kdp_cmd as the value of Ksp as described above.

  Note that the values of Iin and Iout necessary for calculating the 2 × 2 component of the matrix A are predetermined values and are stored and held in advance in a memory (not shown).

  Further, each component other than the 2-row and 2-column components of the matrix A, each component of the column vector B, and δ in Expression (17) are constant values in the present embodiment, as in the first embodiment. Not stored in memory. Further, the value of Ksld in the equation (17) is calculated by the equation (25) according to the value of the switching function σ as in the first embodiment.

  The present embodiment is the same as the first embodiment except for the matters described above.

  According to the fourth embodiment described above, the control input determination unit 34 of the control device 30 uses the secondary side to determine the control input (target torque τm_cmd of the power source motor 5) by the sliding mode control method. Similarly to the first embodiment, the measured value of the torque τ is sequentially acquired by the secondary side torque measuring unit 34f based on the formula (13a).

  On the other hand, in the fourth embodiment, unlike the first embodiment, the gradient of the switching hyperplane σ = 0 for sliding mode control (or the convergence time constant Tc0 on the switching hyperplane σ = 0) is the target viscosity characteristic coefficient Kdp_cmd. It is decided not to depend on. Specifically, the inclination of the switching hyperplane σ = 0 is determined to be an inclination corresponding to a convergence time constant Tc0 having a predetermined value (= a).

  In this case, the convergence time constant Tc0 corresponding to the inclination of the switching hyperplane σ = 0 is determined by the preparatory processing performed in the same manner as in the first embodiment, and the appropriate convergence time constant Tc and the viscosity characteristic coefficient Kdp specified in advance. Is the value of the appropriate convergence time constant Tc when the value of the viscosity characteristic coefficient Kdp is zero.

  In the fourth embodiment, the gradient of the switching hyperplane σ = 0 for sliding mode control (or the convergence time constant Tc0 on the switching hyperplane σ = 0) does not depend on the target viscosity characteristic coefficient Kdp_cmd as described above. On the other hand, the matrix A of the arithmetic expression (17) for calculating the control input (τm_cmd) includes a component that depends on the viscosity characteristic coefficient Kdp of the elastic force generation mechanism 4.

  As a value of the viscosity characteristic coefficient Kdp that defines the value of the component, a value (= b · Kdp_cmd) obtained by correcting the target viscosity characteristic coefficient Kdp_cmd with the correction coefficient b determined in advance based on the above correlation. used.

  For this reason, also in the fourth embodiment, as a result, similarly to the first embodiment, it depends on the viscous force generated by the elastic force generating mechanism 4 regardless of the control state of the viscosity characteristic coefficient of the elastic force generating mechanism 4. The actual secondary torque τ is stable and highly stable with high robustness while suppressing the occurrence of vibration of the measured value of the secondary torque τ by appropriately compensating the influence. Can be made to converge.

  In the preparatory process, the appropriate convergence time constant Tc corresponding to each representative value of the target viscosity characteristic coefficient Kdp_cmd of the elastic force generation mechanism 4 is set in the same manner as in the first embodiment in a plurality of response characteristics. Determined using data.

  Further, based on each representative value of the target viscosity characteristic coefficient Kdp_cmd of the elastic force generation mechanism 4 and the appropriate time constant Tc determined corresponding to each representative value, between the viscosity characteristic coefficient Kdp and the appropriate convergence time constant Tc. A correlation (in this embodiment, a correlation represented by the formula (22)) is specified.

  For this reason, the said correlation can be specified suitably according to the actual behavior characteristic of the power transmission device 1.

  Further, similarly to the first embodiment, the response characteristic data can be efficiently and easily collected using a general-purpose control method. As a result, a suitable proper convergence time constant Tc for each representative value of the target viscosity characteristic coefficient Kdp_cmd is specified, and a suitable correlation between the viscosity characteristic coefficient Kdp and the appropriate convergence time constant Tc is specified. It can be done efficiently.

  Then, the control input determination unit 34 of the control device 30 includes a switching function σ corresponding to the switching hyperplane σ = 0 having the slope (or the convergence time constant Tc0) defined by the correlation specified as described above, The control input (target torque τm_cmd of the electric motor 5) is performed by the sliding mode control process (the calculation process (17) based on the state equation of the expression (71)) using the correction coefficient b defined by the correlation. ).

  For this reason, it is possible to suitably realize the control of the secondary side torque τ with the required convergence characteristic and high robustness regardless of the control state of the viscosity characteristic coefficient Kdp of the elastic force generating mechanism 4.

  In particular, since the proper convergence time constant Tc of the correlation in the equation (22) is set so as to satisfy the time constant constraint condition, the secondary torque τ can be controlled in a wide operating region of the power transmission device 1. It can be performed stably with high robustness.

  Here, the correspondence relationship between the fourth embodiment described above and the present invention will be supplemented. The fourth embodiment is an embodiment related to the thirteenth to eighteenth inventions and the twenty-third invention. In this case, the formula (22) corresponds to the formula (E) in the present invention. Further, the first term on the right side of the equation (17) corresponds to the first control input component represented by the equation (D) in the present invention, and the second term on the right side of the equation (17) is the second control in the present invention. Corresponds to the input component. Further, the convergence time constant Tc0 represented by the equation (72) corresponds to the convergence time constant on the switching hyperplane in the present invention. Except for this, the correspondence between the fourth embodiment and the present invention is the same as that of the first embodiment.

  As in the case of the first embodiment, in the fourth embodiment, when the offset component (steady error component) included in the secondary torque τ_act is sufficiently small, the calculation shown in FIG. The units 34a and 34c and the low-pass filter 34b may be omitted, and τ_cmd may be input to the calculation unit 34d as it is.

  In the fourth embodiment, similarly to the second embodiment, the sliding mode control processing unit 34e may include an observer 34h.

  In this case, the observer 34h uses the matrix A defined by the proviso of the formula (71) as the matrix A of the formula (51), and uses the secondary side torque deviation estimated value τ_err_hat and the secondary side torque. A deviation speed estimated value dτ_err_hat is calculated. In this case, Ksp_cmd and b · Kdp_cmd are used as the values of Ksp and Kdp, respectively.

  Then, the sliding mode control processing unit 34e calculates the secondary side torque deviation estimated value τ_err_hat and the secondary side torque deviation speed estimated value dτ_err_hat, which are sequentially calculated by the observer 34h, the τ_err_act calculated by the calculating unit 34d, and the time The target torque τm_cmd of the power source motor 5 as the control input is sequentially calculated by performing the calculations of the above formulas (17) and (18) instead of dτ_err_act calculated as the dynamic change rate. The other processes may be the same as those in the fourth embodiment. By doing so, one embodiment of the twenty-fourth aspect of the invention is constructed.

[Fifth Embodiment]
Next, a fifth embodiment of the present invention will be described with reference to FIGS. This embodiment is different from the fourth embodiment only in a part of the processing of the power transmission device. Therefore, the description of this embodiment will be focused on matters that are different from those of the fourth embodiment, and description of the same matters as those of the fourth embodiment will be omitted.

  In the fourth embodiment, the target stiffness characteristic coefficient Ksp_cmd of the elastic force generation mechanism 4 is kept constant. On the other hand, in this embodiment, the target stiffness characteristic coefficient Ksp_cmd is variably set within a predetermined range, and the target stiffness characteristic coefficient Ksp_cmd is sequentially input to the control device 30. .

  Then, similarly to the third embodiment, the stiffness control unit 36 of the control device 30 determines the target value of the rotation angle of the rotation drive shaft 18a of the stiffness variable motor 18 according to the target stiffness characteristic coefficient Ksp_cmd, The actual rotation angle of the rotary drive shaft 18a (observed value indicated by the output of the angle detector 33) is controlled to the determined target value. As a result, the actual stiffness characteristic coefficient Ksp of the elastic force generation mechanism 4 is controlled to the variably set target stiffness characteristic coefficient Ksp_cmd.

  In the present embodiment, as shown in FIG. 13, in addition to the target viscosity characteristic coefficient Kdp_cmd, the target stiffness characteristic coefficient Ksp_cmd is sequentially input to the sliding mode control processing unit 34e of the control input determination unit 34.

  The sliding mode control processing unit 34e in the present embodiment includes a switching hyperplane variable setting unit 34g that variably sets the gradient of the switching hyperplane σ = 0 (or the convergence time constant on the switching hyperplane σ = 0). Yes. The switching hyperplane variable setting unit 34g variably determines the gradient of the switching hyperplane σ = 0 (or the convergence time constant on the switching hyperplane σ = 0) according to the input target stiffness characteristic coefficient Ksp_cmd.

  More specifically, in the present embodiment, the pre-preparation process is executed in the same manner as in the third embodiment, so that the target viscosity characteristic coefficient Kdp_cmd and the target rigidity characteristic coefficient Ksp_cmd of the elastic force generation mechanism 4 are converged. The values of the variables a2, a1, a0, b in the equation (62) that define the correlation with the time constant Tc are determined.

  In the present embodiment, the switching hyperplane variable setting unit 34g sets the value of (a2 / sqrt (Ksp_cmd) + a1 · Ksp_cmd + a0) in Expression (62), that is, Tc, Kdp_cmd, and Ksp_cmd defined by Expression (62). The value of the convergence time constant Tc (= a2 / sqrt (Ksp_cmd) + a1 · Ksp_cmd + a0) when the target viscosity characteristic coefficient Kdp_cmd is zero in the correlation between and the switching hyperplane σ = 0 The convergence time constant is variably determined according to Ksp_cmd.

  Therefore, in this embodiment, the inclination of the switching hyperplane σ = 0 is such that the convergence time constant on the switching hyperplane σ = 0 is (a2 / sqrt (Ksp_cmd) + a1 · Ksp_cmd + a0). It is variably determined according to Ksp_cmd. Hereinafter, the convergence time constant on the switching hyperplane σ = 0 in this embodiment will be referred to as Tc0. In this case, the convergence time constant Tc0 is given by the following equation (81).

Tc0 = a2 / sqrt (Ksp_cmd) + a1 / Ksp_cmd + a0 (81)

For this reason, the convergence time constant Tc0 on the switching hyperplane σ = 0 determined by the switching hyperplane variable setting unit 34g in the present embodiment is a value that does not depend on the target viscosity characteristic coefficient Kdp_cmd and is equal to the target stiffness characteristic coefficient Ksp_cmd. It becomes a corresponding value.

  Then, the switching hyperplane variable setting unit 34g is configured so that s2 / s1 matches the convergence time constant Tc0 on the switching hyperplane σ = 0 (the convergence time constant Tc0 defined by the expression (81)). 18) The values of the coefficient components s1, s2 of the switching function σ are determined, for example, s1 = 1, s2 = Tc0, or s1 = 1 / Tc0, s2 = 1.

  The processing of the control device 30 in the present embodiment is the same as that in the fourth embodiment except for the matters described above.

  Therefore, the actual secondary torque τ_act is controlled to follow the target secondary torque τ_cmd while the viscosity characteristic coefficient Kdp and the rigidity characteristic coefficient Ksp of the elastic force generating mechanism 4 are appropriately changed.

  According to the fifth embodiment described above, the control input determination unit 34 of the control device 30 calculates the measured value of the secondary side torque τ based on the equation (13a) as in the first embodiment and the fourth embodiment. To acquire sequentially.

  In the fifth embodiment, the gradient of the switching hyperplane σ = 0 for sliding mode control (or the convergence time constant Tc0 on the switching hyperplane σ = 0) is the target viscosity characteristic coefficient as in the fourth embodiment. It is determined not to depend on Kdp_cmd. However, in the fifth embodiment, the inclination is determined to be an inclination corresponding to the convergence time constant Tc0 represented by the equation (81) according to the target stiffness characteristic coefficient Ksp.

  In this case, the convergence time constant Tc0 corresponding to the gradient of the switching hyperplane σ = 0 is determined by the preliminarily prepared appropriate convergence time constant Tc and the viscosity characteristic coefficient Kdp by the preparatory processing performed in the same manner as in the fourth embodiment. In the correlation with the stiffness characteristic coefficient Ksp, the value of the proper convergence time constant Tc when the value of the viscosity characteristic coefficient Kdp is zero is used.

  In the fifth embodiment, the gradient of the switching hyperplane σ = 0 (or the convergence time constant Tc0 on the switching hyperplane σ = 0) does not depend on the target viscosity characteristic coefficient Kdp_cmd but depends on the target rigidity characteristic coefficient Ksp_cmd. On the other hand, the matrix A of the arithmetic expression (17) for calculating the control input (τm_cmd) includes a component that depends on the viscosity characteristic coefficient Kdp of the elastic force generation mechanism 4.

  As a value of the viscosity characteristic coefficient Kdp that defines the value of the component, a value (= b · Kdp_cmd) obtained by correcting the target viscosity characteristic coefficient Kdp_cmd with the correction coefficient b determined in advance based on the above correlation. used.

  Therefore, in the fifth embodiment, as in the third embodiment, as a result, the elastic force generation mechanism 4 does not depend on the control state of the viscosity characteristic coefficient and the rigidity characteristic coefficient of the elastic force generation mechanism 4. The actual secondary torque τ can be stably increased with high robustness while properly compensating for the influence of the generated viscous force and suppressing the vibration of the measured value of the secondary torque τ. It can be made to converge to the secondary target torque τ_cmd.

  In the pre-preparation process, the appropriate convergence time constant Tc corresponding to each set of representative values of the target viscosity characteristic coefficient Kdp_cmd and the target stiffness characteristic coefficient Ksp_cmd of the elastic force generation mechanism 4 is the same as in the first embodiment. In this way, it is determined using a plurality of response characteristic data.

  Furthermore, the viscosity characteristic coefficient Kdp and the rigidity are determined based on the set of representative values of the target viscosity characteristic coefficient Kdp_cmd and the target rigidity characteristic coefficient Ksp_cmd of the elastic force generation mechanism 4 and the appropriate time constant Tc determined corresponding to each set. A correlation between the characteristic coefficient Ksp and the appropriate convergence time constant Tc (in this embodiment, a correlation represented by the equation (62)) is specified.

  For this reason, the said correlation can be specified suitably according to the actual behavior characteristic of the power transmission device 1.

  Further, similarly to the first embodiment, the response characteristic data can be efficiently and easily collected using a general-purpose control method. As a result, a suitable proper convergence time constant Tc for each set of representative values of the target viscosity characteristic coefficient Kdp_cmd and the target stiffness characteristic coefficient Ksp_cmd is specified, and further, the viscosity characteristic coefficient Kdp, the stiffness characteristic coefficient Ksp, and the proper convergence time constant It is possible to efficiently identify a suitable correlation with Tc.

  Then, the control input determination unit 34 of the control device 30 includes a switching function σ corresponding to the switching hyperplane σ = 0 having the slope (or the convergence time constant Tc0) defined by the correlation specified as described above, Using the correction coefficient b defined by the correlation, as in the fourth embodiment, control input is performed by the sliding mode control process (the calculation process of (17) based on the state equation of the expression (71)). (Target torque τm_cmd of the electric motor 5) is determined.

  For this reason, the control of the secondary torque τ is suitably realized with the required convergence characteristics and high robustness regardless of the control state of the viscosity characteristic coefficient Kdp and the rigidity characteristic coefficient Ksp of the elastic force generation mechanism 4. can do.

  In particular, since the proper convergence time constant Tc of the correlation in the equation (62) is set so as to satisfy the time constant constraint condition, the secondary side torque τ can be controlled in a wide operating region of the power transmission device 1. It can be performed stably with high robustness.

  Here, the correspondence relationship between the fifth embodiment described above and the present invention will be supplemented. The fifth embodiment is an embodiment related to the thirteenth invention, the fourteenth invention, and the nineteenth to twenty-third invention. In this case, the formula (62) corresponds to the formula (F) in the present invention. The switching hyperplane variable setting unit 34g corresponds to the switching hyperplane variable setting means in the present invention. Further, the convergence time constant Tc0 represented by the equation (81) corresponds to the convergence time constant on the switching hyperplane in the present invention. Except for this, the correspondence between the fifth embodiment and the present invention is the same as that of the fourth embodiment.

  As in the case of the first embodiment, also in the fifth embodiment, when the offset component (steady error component) included in the secondary torque τ_act is sufficiently small, the calculation shown in FIG. The units 34a and 34c and the low-pass filter 34b may be omitted, and τ_cmd may be input to the calculation unit 34d as it is.

  In the fifth embodiment, similarly to the second embodiment, the sliding mode control processing unit 34e may include an observer 34h.

  In this case, the observer 34h uses the matrix A defined by the proviso of the equation (71) as the matrix A of the equation (51) as in the case of supplementary explanation regarding the fourth embodiment. A secondary torque deviation estimated value τ_err_hat and a secondary torque deviation speed estimated value dτ_err_hat are calculated.

  Then, the sliding mode control processing unit 34e calculates the secondary side torque deviation estimated value τ_err_hat and the secondary side torque deviation speed estimated value dτ_err_hat, which are sequentially calculated by the observer 34h, the τ_err_act calculated by the calculating unit 34d, and the time The target torque τm_cmd of the power source motor 5 as the control input is sequentially calculated by performing the calculations of the above formulas (17) and (18) instead of dτ_err_act calculated as the dynamic change rate. The other processes may be the same as those in the fourth embodiment. By doing so, another embodiment of the twenty-fourth aspect of the present invention is constructed.

[Modification]
Next, some modifications related to the above embodiments will be described.

  In each of the embodiments described above, the electric motor 5 (power source motor 5) is used as the actuator that generates the driving force. However, an electric actuator other than the electric motor 5 or an actuator of another form such as a hydraulic actuator is used. Also good.

  In the first embodiment, the correlation between the viscosity characteristic coefficient Kdp and the gradient of the switching hyperplane σ = 0 (or the convergence time constant Tc on the switching hyperplane σ = 0) is expressed by the above equation (22). Instead of being specified by the arithmetic expression represented by the following equation, it is specified by map data, and when the operation of the power transmission device 1 is controlled, based on the map data, the gradient of the switching hyperplane σ = 0 (or the switching hyperplane σ The convergence time constant Tc) = 0 may be determined according to the target viscosity characteristic coefficient Kdp_cmd.

  Similarly, in the third embodiment, the correlation between the viscosity characteristic coefficient Kdp, the rigidity characteristic coefficient Ksp, and the slope of the switching hyperplane σ = 0 (or the convergence time constant Tc on the switching hyperplane σ = 0) is obtained. , Instead of specifying by the arithmetic expression represented by the equation (62), it is specified by map data, and when the operation of the power transmission device 1 is controlled, the inclination of the switching hyperplane σ = 0 based on the map data (Or the convergence time constant Tc on the switching hyperplane σ = 0) may be determined according to the target viscosity characteristic coefficient Kdp_cmd and the target stiffness characteristic coefficient Ksp_cmd.

  In the fifth embodiment, during the operation control of the power transmission device 1, the gradient of the switching hyperplane σ = 0 (or the convergence time constant Tc on the switching hyperplane σ = 0) is determined according to the target stiffness characteristic coefficient Ksp_cmd. Instead of determining based on the equation (81), the gradient of the switching hyperplane σ = 0 (or the convergence time constant Tc on the switching hyperplane σ = 0) is calculated based on the map data corresponding to the equation (81). You may make it determine according to the target rigidity characteristic coefficient Ksp_cmd.

  In the first, second, and fourth embodiments in which the stiffness characteristic coefficient Ksp of the elastic force generating mechanism 4 is kept constant, the stiffness variable motor 18 in the elastic force generating mechanism 4 is omitted, and the spring One end of the worm 17 opposite to the cylinder 19 (one end on the spring seat member 20a side) may be fixed to a member that rotatably supports the driving pulley 2, the driven pulley 3, and the like.

  In each of the above embodiments, the elastic force generation mechanism 4 having the configuration shown in FIGS. 2 to 4 is adopted. However, the elastic force generation mechanism 4 includes the drive pulley 2 (primary element) and the driven pulley 3 (two It is possible to generate elastic force and viscous force with the secondary element) and to control the viscosity characteristic coefficient variably, or to control both the viscosity characteristic coefficient and the rigidity characteristic coefficient. Any other structure may be used as long as it exists.

  For example, the elastic force generation mechanism may be configured using a conductive polymer actuator.

  The power transmission device 1 in each of the embodiments transmits a rotational driving force. However, in the power transmission device of the present invention, the secondary side element moves in a translational manner relative to the primary side element. In this way, they may be connected by an elastic force generation mechanism so as to transmit a translational force between both elements.

  DESCRIPTION OF SYMBOLS 1 ... Power transmission device, 2 ... Drive pulley (primary side element), 3 ... Driven pulley (secondary side element), 4 ... Elastic force generation mechanism, 5 ... Electric motor (actuator), 30 ... Control device, 34 ... Control Input determining unit (control input determining unit), 34f... Secondary torque measuring unit (secondary power measuring unit), 34g... Switching hyperplane variable setting unit (switching hyperplane variable setting unit), 34h.

Claims (24)

A primary side element that is displaced by a driving force of an actuator; and a secondary side element that is connected to the primary side element via an elastic force generating mechanism so as to be relatively displaceable with respect to the primary side element. And a power transmission device configured to transmit power between the two elements by the elastic force generated by the elastic force generation mechanism between the two elements in accordance with the relative displacement between the two elements. A control device for controlling secondary power, which is power given to the secondary element by the power transmission, to a target value,
The elastic force generation mechanism generates a viscous force according to a relative speed between the primary side element and the secondary side element, and a ratio of a change in the viscous force to a change in the relative speed between the two elements. The viscosity characteristic coefficient that represents is variably controllable,
The relative displacement amount between the primary side element and the secondary side element is measured, the measured value of the relative displacement amount, and the ratio of the change in the elastic force generated by the elastic force generation mechanism to the change in the relative displacement amount A secondary-side power measuring means for obtaining a measured value of the secondary-side power applied to the secondary-side element by the generated elastic force from the value of the stiffness characteristic coefficient representing
By control processing of sliding mode control using a switching function configured with a deviation between the measured value of the secondary power and the target value as a first variable component, and a temporal change rate of the deviation as a second variable component, Control input determining means for sequentially determining a control input for controlling the driving force of the actuator so that the first variable component converges to zero on a switching hyperplane defined by a switching function;
A switching hyperplane variable setting for setting the inclination of the switching hyperplane in a phase plane having the first variable component and the second variable component as two coordinate axis components according to the control value of the viscosity characteristic coefficient Means and
The control input determining means is configured to determine the control input by using the switching function corresponding to the switching hyperplane having the set inclination. . In the control device of the power transmission device according to claim 1,
The elastic force generation mechanism is configured such that the stiffness characteristic coefficient is a constant value,
The switching hyperplane variable setting means includes a convergence time constant on the switching hyperplane defined by an inclination of the switching hyperplane as a time constant of convergence of the first variable component to zero on the switching hyperplane, The inclination of the switching hyperplane is determined according to the control value of the viscosity characteristic coefficient so that the correlation between the viscosity characteristic coefficient is expressed by the following equation (A). A control device for a power transmission device.

Tc = a + b · Kdp / Ksp (A)
However,
Tc: convergence time constant on the switching hyperplane a, b: predetermined constant value Ksp: value of the stiffness characteristic coefficient Kdp: value of the viscosity characteristic coefficient In the control device of the power transmission device according to claim 1,
The elastic force generation mechanism is configured to be able to variably control the viscosity characteristic coefficient and the rigidity characteristic coefficient,
The switching hyperplane variable setting means includes a convergence time constant on the switching hyperplane defined by an inclination of the switching hyperplane as a time constant of convergence of the first variable component to zero on the switching hyperplane, According to the control value of the viscosity characteristic coefficient and the control value of the rigidity characteristic coefficient so that the correlation between the viscosity characteristic coefficient and the rigidity characteristic coefficient is represented by the following equation (B): A control device for a power transmission device configured to determine an inclination of the switching hyperplane.

Tc = (a2 / sqrt (Ksp)) + a1 * Ksp + a0 + b * Kdp / Ksp (B)
However,
Tc: convergence time constant on the switching hyperplane a2, a1, a0, b: constant values determined in advance Ksp: value of the stiffness characteristic coefficient Kdp: value of the viscosity characteristic coefficient In the control apparatus of the power transmission device according to claim 2 or 3,
The control input determining means includes a first control input component calculated by the following equation (C) as a control input component having a function of converging the first variable component to zero on the switching hyperplane, and the switching function A value obtained by combining a second control input component determined according to the value of the switching function as a control input component having a function of converging the value to zero is determined as the control input. A control device for a power transmission device.

In the control device of the power transmission device according to claim 1,
The elastic force generation mechanism is configured such that the stiffness characteristic coefficient is a constant value,
The switching hyperplane variable setting means is a correlation between the inclination of the switching hyperplane and the viscosity characteristic coefficient, and includes a plurality of representative values of the viscosity characteristic coefficient, and representative values of the viscosity characteristic coefficient. The inclination of the switching hyperplane is determined according to the control value of the viscosity characteristic coefficient in accordance with a correlation specified in advance based on the inclination of the switching hyperplane determined in advance corresponding to And
The inclination of the switching hyperplane corresponding to each representative value of the viscosity characteristic coefficient is such that the first variable component is changed from an arbitrary value to zero while the viscosity characteristic coefficient of the elastic force generation mechanism is controlled to the representative value. Specific response characteristics satisfying a predetermined requirement among a plurality of response characteristic data acquired in advance as data indicating a locus of transition of a set of the value of the first variable component and the value of the second variable component at the time of convergence Based on data,
A line indicating, on the phase plane, a first set value set according to the first allowable limit value between a first allowable limit value which is an allowable limit value of the magnitude of the value of the first variable component and zero. , A first set value line, a second set value set according to the second allowable limit value between zero and a second allowable limit value which is an allowable limit value of the value of the second variable component Is defined as a second set value line, the specific response characteristic data satisfying the predetermined requirement is the first variable component on the locus indicated by the specific response characteristic data. And the value of the second variable component fall within the first allowable limit value and the second allowable limit value, respectively, and the locus is the first set value line or the second set value. Response characteristic data that satisfies the requirement to cross the value line It is in,
The inclination of the switching hyperplane corresponding to each representative value of the viscosity characteristic coefficient is the locus indicated by each of the specific response characteristic data and the first set value line or the second set value line in the phase plane. The constraint that the time constant of convergence of the first variable component to zero on the line connecting the intersection and the origin of the phase plane is a value within a predetermined required range among A control device for a power transmission device, wherein the control device is determined so as to match or approximate an inclination of a line connecting a satisfying intersection and an origin of the phase plane. In the control apparatus of the power transmission device according to claim 5,
The first permissible limit value is a permissible limit value of the displacement speed of the primary side element so that a displacement speed and a displacement acceleration of the primary side element due to the driving force of the actuator do not exceed predetermined permissible limit values, respectively. And an allowable limit value of displacement acceleration of the primary side element and a value of the stiffness characteristic coefficient. In the control apparatus of the power transmission device according to claim 5 or 6,
The second permissible limit value depends on the value of the stiffness characteristic coefficient so as to prevent the occurrence of vibration of the power transmission system according to the natural vibration of the power transmission system from the actuator to the secondary element. A control device for a power transmission device, which is a set value. In the control device of the power transmission device according to claim 1,
The elastic force generation mechanism is configured to be able to variably control the viscosity characteristic coefficient and the rigidity characteristic coefficient,
The switching hyperplane variable setting means is a correlation among the inclination of the switching hyperplane, the viscosity characteristic coefficient, and the stiffness characteristic coefficient, wherein a plurality of representative values of the viscosity characteristic coefficient, and the stiffness In accordance with a predetermined correlation based on a plurality of representative values of the characteristic coefficient and the inclination of the switching hyperplane determined in advance corresponding to each set of representative values of the viscosity characteristic coefficient and the rigidity characteristic coefficient The inclination of the switching hyperplane is determined according to the control value of the viscosity characteristic coefficient and the control value of the rigidity characteristic coefficient,
The inclination of the switching hyperplane corresponding to each set of representative values of the viscosity characteristic coefficient and the rigidity characteristic coefficient is in a state where the viscosity characteristic coefficient and the rigidity characteristic coefficient of the elastic force generation mechanism are controlled to the representative values, respectively. , A plurality of response characteristics acquired in advance as data indicating a locus of transition of a set of the value of the first variable component and the value of the second variable component when the first variable component is converged from an arbitrary value to zero Determined based on specific response characteristics data that meets certain requirements of the data,
A line indicating, on the phase plane, a first set value set according to the first allowable limit value between a first allowable limit value which is an allowable limit value of the magnitude of the value of the first variable component and zero. , A first set value line, a second set value set according to the second allowable limit value between zero and a second allowable limit value which is an allowable limit value of the value of the second variable component Is defined as a second set value line, the specific response characteristic data satisfying the predetermined requirement is the first variable component on the locus indicated by the specific response characteristic data. And the value of the second variable component fall within the first allowable limit value and the second allowable limit value, respectively, and the locus is the first set value line or the second set value. Response characteristic data that satisfies the requirement to cross the value line It is in,
The inclination of the switching hyperplane corresponding to each set of representative values of the viscosity characteristic coefficient and the rigidity characteristic coefficient is the trajectory indicated by each of the specific response characteristic data and the first setting on the phase plane. Among the intersections with the value line or the second set value line, the time constant for convergence of the first variable component to zero on the line connecting the intersection and the origin of the phase plane is within a predetermined required range. A control device for a power transmission device, wherein the control device is determined so as to coincide with or approximate to an inclination of a line connecting an intersection satisfying a constraint condition that is a value of and an origin of the phase plane. The control device for a power transmission device according to claim 8,
The first permissible limit value is a value set corresponding to each representative value of the stiffness characteristic coefficient, and the first permissible limit value corresponding to each representative value of the stiffness characteristic coefficient is the elastic force generation In a state in which the stiffness characteristic coefficient of the mechanism is controlled to the representative value, the primary side element is controlled such that the displacement speed and the displacement acceleration of the primary side element due to the driving force of the actuator do not exceed predetermined allowable limit values. A control device for a power transmission device, characterized in that the value is set in accordance with a permissible limit value of the displacement speed, a permissible limit value of the displacement acceleration of the primary side element, and the representative value of the stiffness characteristic coefficient. In the control device of the power transmission device according to claim 8 or 9,
The second permissible limit value is a value set corresponding to each representative value of the stiffness characteristic coefficient, and the second permissible limit value corresponding to each representative value of the stiffness characteristic coefficient is the elastic force generation In a state where the stiffness characteristic coefficient of the mechanism is controlled to the representative value, the rigidity of the power transmission system is prevented from occurring in accordance with the natural vibration of the power transmission system from the actuator to the secondary element. A control device for a power transmission device, which is a value determined according to the representative value of the characteristic coefficient. In the control apparatus of the power transmission device according to any one of claims 5 to 10,
The value of the first variable component is changed stepwise from zero to the first permissible limit value, and the convergence of the value of the first variable component to zero is caused by the displacement acceleration of the primary side element by the driving force of the actuator. When the time constant realized when assuming that the displacement acceleration is set to the allowable limit value is defined as a specific time constant, the constraint condition is that the time constant of convergence of the first variable component to zero is A control device for a power transmission device, characterized in that the condition is limited to a value in a range not less than the specific time constant. In the control apparatus of the power transmission device according to any one of claims 1 to 11,
A first variable component measurement value that is a value of the first variable component calculated from the measurement value of the secondary power and the target value of the secondary power, and a temporal change of the first variable component measurement value The estimated value of the first variable component and the estimated value of the second variable component obtained by reducing the influence of disturbance from the second variable component measured value that is the value of the second variable component calculated as a rate are sequentially It also has an observer to calculate,
The control input determining means includes an estimated value of the first variable component and an estimated value of the second variable component calculated by the observer instead of the first variable component measured value and the second variable component measured value. A control device for a power transmission device, wherein the control input is sequentially generated by using the value of the switching function while calculating the value of the switching function using. A primary side element that is displaced by a driving force of an actuator; and a secondary side element that is connected to the primary side element via an elastic force generating mechanism so as to be relatively displaceable with respect to the primary side element. And a power transmission device configured to transmit power between the two elements by the elastic force generated by the elastic force generation mechanism between the two elements in accordance with the relative displacement between the two elements. A control device for controlling secondary power, which is power given to the secondary element by the power transmission, to a target value,
The elastic force generation mechanism generates a viscous force according to a relative speed between the primary side element and the secondary side element, and a ratio of a change in the viscous force to a change in the relative speed between the two elements. The viscosity characteristic coefficient that represents is variably controllable,
The relative displacement amount between the primary side element and the secondary side element is measured, the measured value of the relative displacement amount, and the ratio of the change in the elastic force generated by the elastic force generation mechanism to the change in the relative displacement amount A secondary-side power measuring means for obtaining a measured value of the secondary-side power applied to the secondary-side element by the generated elastic force from the value of the stiffness characteristic coefficient representing
By control processing of sliding mode control using a switching function configured with a deviation between the measured value of the secondary power and the target value as a first variable component, and a temporal change rate of the deviation as a second variable component, Control input determining means for sequentially determining a control input for controlling the driving force of the actuator so that the first variable component converges to zero on a switching hyperplane defined by a switching function;
The control input determination means is a phase plane having a correction value obtained by correcting the control value of the viscosity characteristic coefficient with a predetermined correction coefficient, and the first variable component and the second variable component as two coordinate axis components. The control input is determined using the switching function determined so that the inclination of the switching hyperplane becomes an inclination independent of a change in the control value of the viscosity characteristic coefficient. Control device for power transmission device. The control device for a power transmission device according to claim 13,
The control input determining means includes a first control input component calculated by the following equation (D) as a control input component having a function of converging the first variable component to zero on the switching hyperplane, and the switching function A value obtained by combining a second control input component determined according to the value of the switching function as a control input component having a function of converging the value to zero is determined as the control input. A control device for a power transmission device.

In the control device of the power transmission device according to claim 13 or 14,
The elastic force generation mechanism is configured such that the stiffness characteristic coefficient is a constant value,
Correlation between an appropriate time constant for converging the first variable component to zero and the viscosity characteristic coefficient, and a plurality of representative values of the viscosity characteristic coefficient and representative values of the viscosity characteristic coefficient The switching hyperplane to the value of the appropriate time constant when the value of the viscosity characteristic coefficient is zero in the correlation specified in advance based on the value of the appropriate time constant determined in advance corresponding to The inclination of the switching hyperplane is determined so that the convergence time constant on the switching hyperplane, which is the time constant of convergence of the first variable component to zero above, matches, and in the correlation, any The difference between the value of the appropriate time constant corresponding to the value of the viscosity characteristic coefficient and the value of the appropriate time constant when the value of the viscosity characteristic coefficient is zero is the value of the arbitrary viscosity characteristic coefficient. Of the correction value corrected by the correction coefficient Numerical and said correction coefficient are determined so that,
The value of the appropriate time constant corresponding to each representative value of the viscosity characteristic coefficient is set such that the first variable component is changed from an arbitrary value to zero while the viscosity characteristic coefficient of the elastic force generation mechanism is controlled to the representative value. Specific response characteristics satisfying a predetermined requirement among a plurality of response characteristic data acquired in advance as data indicating a locus of transition of a set of the value of the first variable component and the value of the second variable component at the time of convergence Based on data,
A line indicating, on the phase plane, a first set value set according to the first allowable limit value between a first allowable limit value which is an allowable limit value of the magnitude of the value of the first variable component and zero. , A first set value line, a second set value set according to the second allowable limit value between zero and a second allowable limit value which is an allowable limit value of the value of the second variable component Is defined as a second set value line, the specific response characteristic data satisfying the predetermined requirement is the first variable component on the locus indicated by the specific response characteristic data. And the value of the second variable component fall within the first allowable limit value and the second allowable limit value, respectively, and the locus is the first set value line or the second set value. Response characteristic data that satisfies the requirement to cross the value line It is in,
The value of the appropriate time constant corresponding to each representative value of the viscosity characteristic coefficient is the trajectory indicated by each of the specific response characteristic data and the first set value line or the second set value line on the phase plane. The constraint that the time constant of convergence of the first variable component to zero on the line connecting the intersection and the origin of the phase plane is a value within a predetermined required range among Control of the power transmission device, which is set so as to coincide with or approximate a time constant of convergence of the first variable component to zero on a line connecting a satisfying intersection and the origin of the phase plane apparatus. The control device for a power transmission device according to claim 15,
The control device for a power transmission device, wherein the correlation is a relationship represented by the following equation (E):

Tc = a + b · Kdp / Ksp (E)
However,
Tc: value of the appropriate time constant Ksp: value of the stiffness characteristic coefficient Kdp: value of the viscous characteristic coefficient a: predetermined value determined in advance b: predetermined value determined in advance as the correction coefficient The control device for a power transmission device according to claim 15 or 16,
The first permissible limit value is a permissible limit value of the displacement speed of the primary side element so that a displacement speed and a displacement acceleration of the primary side element due to the driving force of the actuator do not exceed predetermined permissible limit values, respectively. And a value determined in accordance with an allowable limit value of displacement acceleration of the primary side element and the value of the stiffness characteristic coefficient. In the control apparatus of the power transmission device according to any one of claims 15 to 17,
The second permissible limit value depends on the value of the stiffness characteristic coefficient so as to prevent the occurrence of vibration of the power transmission system according to the natural vibration of the power transmission system from the actuator to the secondary element. A control device for a power transmission device, wherein the control value is a determined value. In the control device of the power transmission device according to claim 13 or 14,
The elastic force generation mechanism is configured to be able to variably control the viscosity characteristic coefficient and the rigidity characteristic coefficient,
Switching hyperplane variable setting means for setting the inclination of the switching hyperplane to change according to the control value of the stiffness characteristic coefficient,
The switching hyperplane variable setting means is a correlation between an appropriate time constant for converging the first variable component to zero, the viscosity characteristic coefficient, and the stiffness characteristic coefficient, wherein the viscosity characteristic coefficient A plurality of representative values, a plurality of representative values of the stiffness characteristic coefficient, and a value of the appropriate time constant determined in advance corresponding to each set of representative values of the viscosity characteristic coefficient and the stiffness characteristic coefficient. Based on the correlation specified in advance, the value of the viscosity characteristic coefficient is set to zero, and the value of the appropriate time constant when the value of the rigidity characteristic coefficient matches the control value of the rigidity characteristic coefficient, Depending on the control value of the stiffness characteristic coefficient, the switching hyperplane has a convergence time constant on the switching hyperplane that is a time constant for convergence of the first variable component to zero on the switching hyperplane. Configured to determine the tilt Cage,
Further, in the correlation, the difference between the value of the appropriate time constant corresponding to the value of the arbitrary viscosity characteristic coefficient and the value of the appropriate time constant when the value of the viscosity characteristic coefficient is zero is the arbitrary time constant. The correction coefficient is determined in advance to be a function value of a correction value obtained by correcting the value of the viscosity characteristic coefficient with the correction coefficient,
The value of the appropriate time constant corresponding to each set of representative values of the viscosity characteristic coefficient and the rigidity characteristic coefficient is the state in which the viscosity characteristic coefficient and the rigidity characteristic coefficient of the elastic force generation mechanism are controlled to the representative values, respectively. , A plurality of response characteristics acquired in advance as data indicating a locus of transition of a set of the value of the first variable component and the value of the second variable component when the first variable component is converged from an arbitrary value to zero Determined based on specific response characteristics data that meets certain requirements of the data,
A line indicating, on the phase plane, a first set value set according to the first allowable limit value between a first allowable limit value which is an allowable limit value of the magnitude of the value of the first variable component and zero. , A first set value line, a second set value set according to the second allowable limit value between zero and a second allowable limit value which is an allowable limit value of the value of the second variable component Is defined as a second set value line, the specific response characteristic data satisfying the predetermined requirement is the first variable component on the locus indicated by the specific response characteristic data. And the value of the second variable component fall within the first allowable limit value and the second allowable limit value, respectively, and the locus is the first set value line or the second set value. Response characteristic data that satisfies the requirement to cross the value line It is in,
The value of the appropriate time constant corresponding to each set of representative values of the viscosity characteristic coefficient and the stiffness characteristic coefficient is the trajectory indicated by each of the specific response characteristic data and the first setting on the phase plane. Among the intersections with the value line or the second set value line, the time constant for convergence of the first variable component to zero on the line connecting the intersection and the origin of the phase plane is within a predetermined required range. Is determined so as to match or approximate the time constant of convergence of the first variable component to zero on a line connecting the intersection satisfying the constraint that the value is the value of and the origin of the phase plane. A control device for a power transmission device. The control device for a power transmission device according to claim 19,
The control device for a power transmission device, wherein the correlation is a relationship represented by the following equation (F).

Tc = (a2 / sqrt (Ksp)) + a1 · Ksp + a0 + b · Kdp / Ksp (F)
However,
Tc: value of the appropriate time constant Ksp: value of the stiffness characteristic coefficient Kdp: value of the viscosity characteristic coefficient a2, a1, a0: predetermined value determined in advance b: predetermined value determined in advance as the correction coefficient The control device for a power transmission device according to claim 19 or 20,
The first permissible limit value is a value set corresponding to each representative value of the stiffness characteristic coefficient, and the first permissible limit value corresponding to each representative value of the stiffness characteristic coefficient is the elastic force generation In a state in which the stiffness characteristic coefficient of the mechanism is controlled to the representative value, the primary side element is controlled such that the displacement speed and the displacement acceleration of the primary side element due to the driving force of the actuator do not exceed predetermined allowable limit values. A control device for a power transmission device, characterized in that the value is determined in accordance with a permissible limit value of the displacement speed, a permissible limit value of the displacement acceleration of the primary side element, and the representative value of the stiffness characteristic coefficient. In the control apparatus of the power transmission device according to any one of claims 19 to 21,
The second permissible limit value is a value set corresponding to each representative value of the stiffness characteristic coefficient, and the second permissible limit value corresponding to each representative value of the stiffness characteristic coefficient is the elastic force generation In a state where the stiffness characteristic coefficient of the mechanism is controlled to the representative value, the rigidity of the power transmission system is prevented from occurring in accordance with the natural vibration of the power transmission system from the actuator to the secondary element. A control device for a power transmission device, which is a value determined according to the representative value of the characteristic coefficient. In the control apparatus of the power transmission device according to any one of claims 15 to 22,
The value of the first variable component is changed stepwise from zero to the first permissible limit value, and the convergence of the value of the first variable component to zero is caused by the displacement acceleration of the primary side element by the driving force of the actuator. When the time constant realized when assuming that the displacement acceleration is set to the allowable limit value is defined as a specific time constant, the constraint condition is that the time constant of convergence of the first variable component to zero is A control device for a power transmission device, characterized in that the condition is limited to a value in a range not less than the specific time constant. In the control apparatus of the power transmission device according to any one of claims 13 to 23,
A first variable component measurement value that is a value of the first variable component calculated from the measurement value of the secondary power and the target value of the secondary power, and a temporal change of the first variable component measurement value The estimated value of the first variable component and the estimated value of the second variable component obtained by reducing the influence of disturbance from the second variable component measured value that is the value of the second variable component calculated as a rate are sequentially It also has an observer to calculate,
The control input determining means includes an estimated value of the first variable component and an estimated value of the second variable component calculated by the observer instead of the first variable component measured value and the second variable component measured value. A control device for a power transmission device, wherein the control input is sequentially generated by using the value of the switching function while calculating the value of the switching function using.

Images