JP2010182084A - Parameter estimation device - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a parameter estimation device for estimating the changing amounts of the control parameter of an actuator to be generated, with high accuracy according to the change of operating situations. <P>SOLUTION: A controller 6 is provided with a loop-gain measurement means 23 for measuring the gain of a feedback loop; a gain intersection control means 24 for determining the adjustment value of the gain so that the gains measured by the loop gain measurement means 23 can be fixed; and a parameter-estimating means 28 for estimating the changing amounts of the control parameter of an actuator 5, with respect to a regulation value from the adjustment value of the gain determined by the gain intersection control means 24. <P>COPYRIGHT: (C)2010,JPO&INPIT

Description

  The present invention relates to a parameter estimation device that estimates an amount of change with respect to a specified value of a control parameter of an actuator that moves in a straight direction or a rotational direction by supplying a current to a coil placed in a magnetic field.

  When assembling semiconductor parts and precision parts manufactured by MEMS (Micro Electro Mechanical Systems) technology, precise positioning is required according to the parts, and contact between parts is detected accurately, resulting in excessive contact force. It is necessary to control the positioning so that it does not become.

  As an example of a device that performs such high-precision control, an arm that pressurizes a semiconductor element when mounted on a substrate is attached to an elastic guide constituted by a parallel leaf spring, and a pressure is applied by a voice coil motor. In addition, there is known one in which the pressure applied is controlled by adjusting the current supplied to the coil of the voice coil motor with a current regulator (Patent Document 1).

  Further, there is known one that feeds back a signal from a force sensor attached to a wrist of a robot to a servo control system (Patent Document 2). This is to recognize the part position by monitoring the force during the robot's work and confirm the success or failure of the work, while realizing the force control by feeding back the force information to the control system and improving the work reliability. It is what I did.

  Furthermore, a method is also known in which force control is executed using a disturbance observer from a position sensor and an actuator drive current without using a contact sensor (Patent Document 3).

JP-A-7-86317 Japanese Patent Laid-Open No. 7-24665 JP-A-6-52563

  However, the conventional positioning control technique has the following problems.

  With the technique of the above-mentioned patent document 1, it is possible to apply a precise pressing force when stationary, but it is difficult to control the contact force precisely by suppressing the vibration during the conveyance of the parts or by moving while pressing. Moreover, in order to detect a contact, it is necessary to add another contact sensor.

  In the technique of Patent Document 2, it is difficult to perform work while maintaining a minute contact force due to the influence of the mass of the arm and the frictional force of the speed reducer.

  In the technique of Patent Document 3, in order to detect the contact force with the disturbance observer, it is assumed that the control parameter of the actuator and the control parameter of the disturbance observer coincide with each other. If the workpiece to be attached is different and the mass changes, the control parameter changes from the specified value and the force detection accuracy decreases.

  In order to solve the above-described problems, an object of the present invention is to provide a parameter estimation device capable of estimating a change amount of an actuator control parameter caused by a change in a work situation with high accuracy.

  The present invention relates to a parameter estimation device that estimates a change amount of a control parameter of an actuator that is displaced in a straight traveling direction or a rotational direction by supplying a current to a coil placed in a magnetic field. A displacement sensor to be measured; deviation calculating means for calculating a deviation between the displacement measured by the displacement sensor and a target position of the actuator; and a current supplied to the coil based on the deviation calculated by the deviation calculating means A controller for feedback control of the gain, a gain measuring means for measuring the gain of the feedback loop, and a gain adjustment for determining an adjustment value of the gain so that the gain measured by the gain measuring means becomes constant Means and a gain adjustment value determined by the gain adjustment means And having a parameter estimation means for estimating a serial variation.

  According to the parameter estimation apparatus of the present invention, when the control parameter of the actuator changes depending on the work situation, the gain of the feedback loop (one-round transfer function) changes accordingly. Therefore, it is possible to estimate the amount of change with respect to the specified value of the control parameter from the gain adjustment value for keeping the gain constant. Thereby, the change amount of the control parameter of the actuator caused by the change of the work situation can be estimated with high accuracy.

  In the present invention, it is preferable that the control parameter is a mass of a movable part of the actuator.

  According to this aspect, even when the mass of the movable part changes due to the work situation due to replacement of the attachment jig to the actuator, the control parameter corresponding to this can be estimated with high accuracy.

  In the present invention, the actuator driving force calculated from the current supplied to the coil, the acceleration calculated from the displacement detected by the displacement sensor, and the change in the movable part mass estimated by the parameter estimation means It is preferable to provide an external force estimating means for estimating an external force acting on the actuator based on the amount.

  According to this aspect, the external force is calculated by the difference between the driving force of the actuator and the value obtained by multiplying the mass of the movable part by the acceleration. Even when the mass of the movable part changes, this can be calculated with high accuracy. The estimated value can be used, and the external force acting on the actuator can be estimated with high accuracy.

The external view of the voice coil motor attached between the arm and hand of the robot. It is a figure which shows the structure of a voice coil motor, (a) is a top view, (b) is sectional drawing. The block diagram which shows the control system of a voice coil motor. The block diagram which shows control of a loop gain adjustment means. The block diagram which shows control of an external force estimation means.

  As an embodiment of the present invention, FIG. 1 includes a controller 6 (see FIG. 3) in addition to a voice coil motor 5 as an actuator attached between an arm 2 and a hand 3 of a robot 1.

  As shown in FIG. 2, the voice coil motor 5 includes a cylindrical casing 11 having an open top surface, a magnet 13 fixed to the inner side surface of the casing 11 together with the side wall of the annular groove-shaped yoke 12, The coil 15 wound around the bobbin 14 disposed so as to face the magnet 13, the cylindrical frame 16 having an open bottom surface disposed inside the yoke 12, and the upper and lower ends of the frame 16 together with the bobbin 14 are encased. An elastic guide member 17 made up of a plurality of arc-shaped flat springs supported in the body 11 and a displacement sensor 18 installed on the bottom surface of the housing 11 inside the gantry 16 are provided.

  As the displacement sensor 18, a proximity sensor, a strain gauge, a linear encoder, or the like is used.

  According to the voice coil motor 5 configured as described above, the gantry 16 is driven in the vertical direction together with the coil 15 by supplying a current to the coil 15 placed in the magnetic field by the magnet 13. The displacement (position change) due to the vertical movement is measured by the displacement sensor 18.

  As shown in FIG. 3, the controller 6 receives as input a comparator 21 that calculates a deviation between the displacement measured by the displacement sensor 18 and the target position of the voice coil motor 5, and the deviation calculated by the comparator 21. A phase compensator 22 that stabilizes the characteristics of the feedback loop, a loop gain control unit 26 including a loop gain measurement unit 23, a gain intersection control unit 24, and a variable gain adjuster 25, and a constant that generates a current proportional to the input voltage. And a current driver 27. The loop gain measuring means 23 corresponds to the gain measuring means of the present invention, and the gain intersection control means 24 corresponds to the gain adjusting means of the present invention.

  Further, the controller 6 uses the adjustment value of the variable gain adjuster 25 to estimate the amount of change with respect to the specified value of the control parameter of the voice coil motor 5, and the estimated value of the parameter estimate 28 and the position sensor 18 are used for measurement. And an external force estimating means 29 for estimating the external force from the determined displacement.

  Next, each component and function of the loop gain adjusting means 26 will be described with reference to FIG.

  The loop gain measuring means 23 is a means for measuring the loop gain of the feedback loop. Specifically, the loop gain measuring means 23 is a sine wave having a frequency at which the loop gain becomes 1 by the transmitter 31, that is, a frequency set at the gain intersection. Is added to the feedback loop via the adder 32.

  The gain intersection control means 24 is a comparator that compares the amplitude ratios of the signals before and after the adder 32. Here, the sine wave signal added to the feedback loop and the signal a that has returned through the loop a The comparator 33 compares the ratio of the amplitude of the sum b with the amplitude of the signal a that has returned through the loop.

  The variable gain adjuster 25 is an adjuster capable of adjusting the loop gain so that the magnification of the amplitude ratio compared by the gain intersection control means 42 becomes 1, and specifically, is configured by a variable gain amplifier. Is done.

  Next, a method for adjusting the loop gain of the feedback loop by the loop gain adjusting means 26 will be described.

  First, the frequency at which the loop gain is 1 is added to the feedback loop via the adder 32 by the transmitter 31.

  Then, the gain intersection control means 24 adjusts the gain of the amplitude ratio of the signals before and after the adder 32 by variable gain adjustment so that the amplitude ratio of the signal returned through the feedback loop is 1. The adjustment value of the device 25 is used.

  That is, when the amplitude level Ea at the point a before the adder 32 and the amplitude level Eb at the point b after the adder 32 are Ea <Eb, the gain intersection control means 24 adjusts the gain of the variable gain adjuster 25. When Ea> Eb, the variable gain adjuster 25 is lowered so that the loop gain of the feedback loop is adjusted to 1. Then, the gain adjustment value of the variable gain adjuster 25 when the loop gain becomes 1 is held.

  Here, the frequency at which the loop gain is 1 is added to the feedback loop by the transmitter 31 at the frequency at which the loop gain is 1 (the frequency at which the gain crosses 0 dB, also referred to as gain intersection frequency). This is because the gain intersection can always be maintained at 1 by adjusting the loop gain so that Ea / Eb = 1. The parameter estimation means 28 estimates the change in the transfer characteristic of the feedback loop from the adjustment value of the variable gain adjuster 25 for always maintaining the gain intersection point at 1.

  Next, a method for estimating the amount of change with respect to the specified value of the control parameter of the voice coil motor 5 by the parameter estimation means 28 will be described.

  The parameter estimation means 28 is a control parameter of the voice coil motor 5 (A) a voltage / current conversion coefficient of the voice coil motor 5, (B) a torque constant, (C) a mass of the movable part (a mass of the gantry 16 and an attachment thereto) The fixed parameter and one variation parameter are set according to the situation in which the voice coil motor 5 is used.

  As for the setting of the fixed parameter and the fluctuation parameter, one fluctuation parameter may be selected in accordance with a selection pattern determined in advance from the usage state of the voice coil motor 5, and one fluctuation parameter is arbitrarily set by the user. You may comprise.

  Next, the parameter estimation means 28 estimates the amount of change with respect to the specified value of one variation parameter set from the adjustment value of the variable gain adjuster 25.

  For example, when the movable part mass is set as a fluctuation parameter, if the jig attached to the gantry 16 is replaced and the movable part mass becomes twice the specified value, the loop gain of the feedback loop is It is 1/2 of the inverse number, and Ea / Eb = 1/2. On the other hand, in order to make the loop gain constant, the variable gain adjuster 25 is adjusted to double the magnification. The parameter estimation means 28 performs a reverse calculation from the adjustment value of the variable gain adjuster 25 and estimates that the mass of the movable part has changed to twice the specified value.

  In this way, the amount of change with respect to the specified value of the control parameter can be estimated from the adjustment value of the variable gain adjuster 25 for keeping the loop gain of the feedback loop constant. Can be estimated with high accuracy.

  Next, a method for estimating the external force acting on the voice coil motor 5 by the external force estimating means 29 will be described with reference to FIG.

  The external force estimating means 29 estimates the external force using an equivalent model (lower side in FIG. 5) corresponding to the control model (upper side in FIG. 5) of the voice coil motor 5.

  That is, the control block of the control model of the voice coil motor 5 includes the voltage / current conversion coefficient Ki (41) of the constant current driver 27, the torque constant Kt (42) for operating the gantry 16, and the movable part mass M. It is composed of an inverse number (43) and two integrators (44, 45). At this time, a driving force is generated in the gantry 16 with the voltage / current conversion coefficient Ki (41) and the torque constant Kt (42) as control parameters for the torque command Rt that is an input signal. At the same time, if an external force F is applied to the gantry 16, the gantry 16 is displaced with respect to these resultant forces using the movable part mass M (43) as a control parameter (43 to 45).

  On the other hand, the equivalent model of the external force estimating means 29 includes an estimated voltage / current conversion coefficient Ki ′ (46) of the constant current driver 27, an estimated torque constant Kt ′ (47) for operating the gantry 16, and two differentiators. (48, 49). One of the two differentiators is multiplied by the estimated movable part mass M ′ (49). At this time, the actual displacement of the gantry 16 is measured by the sensor 18, and the estimated resultant force of the external force F and the driving force of the gantry 16 is calculated from the displacement using the estimated movable part mass M ′ (49) as a control parameter (48). 49). On the other hand, the estimated driving force generated in the gantry 16 is calculated for the torque command Rt, which is an input signal, using the estimated voltage / current conversion coefficient Ki ′ (46) and the estimated torque constant Kt ′ (47) as control parameters. Then, the estimated external force F ′ is calculated by subtracting the estimated driving force from the estimated resultant force.

  Here, the value estimated by the parameter estimation means 28 is applied to any one of the estimated voltage / current conversion coefficient Ki ′ (46), the estimated torque constant Kt ′ (47), and the estimated movable part mass M ′ (49). Is done.

  That is, among these control parameters, those that are set as fixed parameters use the parameter specified values of the control model as estimated values as they are, and only the control parameters set as the fluctuation parameters are estimated by the parameter estimation means 28. Values are used.

  For example, when only the movable part mass is set as the variable parameter, Ki ′ = Ki and Kt ′ = Kt in the equivalent model, and the estimated external force F is estimated by the movable part mass M ′ estimated by the parameter estimation unit. 'Is calculated.

  As a result, even when various jigs are attached to the gantry 16 and the mass of the movable part changes according to the work situation, it is necessary to newly set the estimated mass M ′ of the equivalent model as a disturbance observer. Rather, the external force F actually acting on the gantry 16 can be estimated with high accuracy using the estimated value of the parameter estimating means 28.

  In the present embodiment, the voice coil motor 5 has been described as an example of an actuator. However, the object to which the present invention is applied is not limited to this, and a straight direction or a straight direction by supplying a current to the coil. Any actuator that is displaced in the rotational direction can be applied to various linear actuators.

  In this embodiment, the actuator that moves in the straight direction has been described as an example. However, the present invention is not limited to this, and the present invention may be applied to an actuator in which the gantry rotates and changes according to torque.

DESCRIPTION OF SYMBOLS 1 ... Robot, 2 ... Arm, 3 ... Hand, 5 ... Voice coil motor (actuator), 6 ... Controller, 11 ... Housing, 12 ... Yoke, 13 ... Magnet, 14 ... Bobbin, 15 ... Coil, 16 ... Mount, DESCRIPTION OF SYMBOLS 17 ... Elastic guide member, 18 ... Displacement sensor, 21 ... Comparator (deviation calculation means), 22 ... Phase compensator, 23 ... Loop gain measurement means (gain measurement means), 24 ... Gain intersection control means (gain adjustment means) , 25 ... variable gain adjuster, 26 ... loop gain adjuster, 27 ... constant current driver, 28 ... parameter estimator, 29 ... external force estimator, 31 ... transmitter.

Claims (3)

A parameter estimation device that estimates a change amount with respect to a specified value of a control parameter of an actuator that is displaced in a straight direction or a rotation direction by supplying a current to a coil placed in a magnetic field,
A displacement sensor for measuring the displacement of the actuator;
Deviation calculating means for calculating a deviation between the displacement measured by the displacement sensor and the target position of the actuator;
A controller that feedback-controls the current supplied to the coil based on the deviation calculated by the deviation calculating means;
The controller includes a gain measuring unit that measures the gain of the feedback loop, a gain adjusting unit that determines an adjustment value of the gain so that the gain measured by the gain measuring unit is constant, and the gain adjusting unit And a parameter estimation means for estimating the amount of change from the adjusted gain value. The parameter estimation device according to claim 1,
The parameter estimation apparatus, wherein the control parameter is a mass of a movable part of the actuator. The parameter estimation device according to claim 2, wherein
Based on the driving force of the actuator calculated from the current supplied to the coil, the acceleration calculated from the displacement detected by the displacement sensor, and the amount of change in the movable part mass estimated by the parameter estimation means, A parameter estimation device comprising external force estimation means for estimating an external force acting on an actuator.

Images