JP2010200797A - Motion assisting apparatus - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a motion assisting apparatus which assists actions in an abnormal movable range on the basis of actions in a normal movable range. <P>SOLUTION: The motion assisting apparatus includes speed trajectory predicting means 18 for calculating prediction speed trajectory F(t) of the fore arm on the basis of bending movement of the fore arm within a normal movable range and drive control means 20 for driving a drive motor on the basis of the prediction speed trajectory F(t). The speed trajectory predicting means 18 includes a speed measuring section 30 for measuring sample speeds v<SB>1</SB>, v<SB>2</SB>, and v<SB>3</SB>of the fore arm, and a speed trajectory calculating section 32 for calculating the prediction speed trajectory F(t) as a bell-shaped trajectory on the basis of the sample speeds v<SB>1</SB>, v<SB>2</SB>, and v<SB>3</SB>. The drive control means 20 controls driving of the drive motor on the basis of the prediction speed trajectory F(t) and actuates a second support member assisting the fore arm. By doing so, the fore arm undergoes bending movement within an abnormal movable range in accordance with the prediction speed trajectory F(t) and reaches a target attainment position. <P>COPYRIGHT: (C)2010,JPO&INPIT

Description

  The present invention relates to an operation support apparatus that supports a body part by a support means in a non-normal movable range where normal operation based on the force of the operation subject is impossible, and flexes the body part.

  2. Description of the Related Art Conventionally, various motion support devices that are attached to the body and assist physical motion have been developed. For example, the basic movement of the wearing part of the apparatus is patterned, the intention of movement is determined using the myoelectric signal accompanying the body movement, the apparatus is operated according to the set pattern, and the muscle strength of the user is increased by the apparatus power. A power assist type motion support device has been proposed (see Non-Patent Document 1, for example). Also, a master robot is attached to one body part (for example, the left arm) that can be operated symmetrically such as an arm or a leg, and the slave side is attached to the other body part (for example, the right arm) that operates symmetrically with the body part. There has been proposed a master-slave type motion support device (see, for example, Non-Patent Document 2 and Non-Patent Document 3) in which a robot is mounted and the slave-side robot is operated following the operation of the master-side robot.

KENTA SUZUKI, GOUJI MITO, HIROAKI KAWAMOTO, YASUHISA HASEGAWA, YOSHIYUKI SANKAI, “Intention-based walking support for paraplegia patients with Robot Suit HAL”, Advanced Robotics, Vol.21, No.12, pp.1441-1469, 2007 Mitsunori Uemura, Katsuya Kanaoka, Sadao Kawamura, “Power Assist System for Periodic Motion Using Mechanical Elastic Elements”, Journal of the Robotics Society of Japan, Vol. 26, No. 3, pp. 294-300, 2008 Katsuya Kanaoka, Mitsunori Uemura, “Stabilization of Power Assist Control with Virtual Power Limiter System”, 4th SICE System Integration Division Conference, 2E2-1, 2003

  However, the power assist type motion support device described above is a device that supplements the user's muscle strength by using the myoelectric signal to discriminate the intention of motion, so that the person can move the body with his own will Is intended for support. That is, for example, when the patient can be operated only within a certain range of the movable range of the joint, such as a patient with brachial plexus injury, it is difficult to determine the user's movement intention using the myoelectric signal in the non-movable range. For this reason, it has been impossible to provide motion support beyond the range where self-motion is possible.

  On the other hand, in the case of the above-described master-slave type motion support device, the slave-side robot is operated following the master-side robot. Is possible. However, in the master-slave type motion support device, since the slave side robot performs the same operation as the master side robot, the body part to which the master side robot is attached and the body part to which the slave side robot is attached are independently operated. It is impossible to do (work). That is, it has been difficult to use a master-slave type motion support device for motion support in daily life.

  Accordingly, the present invention has been proposed to solve these problems in view of the above-described problems inherent in the prior art, and is based on the operation in the normal movable range in the movable range of the joint. An object of the present invention is to provide an operation support apparatus that can be operated in a normal movable range.

In order to solve the above-mentioned problem and to achieve the intended purpose suitably, an operation support apparatus according to claim 1 comprises:
A support unit mounted on a body part that bends about a joint; and a drive unit that operates the support unit, and drives the drive unit in a non-normal movable range in which normal operation based on the force of the operation subject is impossible. Then, an operation support device for bending the body part by the support means,
Based on the bending motion of the body part in the normal movable range in which the normal motion based on the force of the motion subject is possible, the desired target reaching position in the non-normal motion range desired by the operating subject from the initial position in the normal motion range A speed trajectory prediction means for calculating a predicted speed trajectory of the body part up to,
Drive control means for controlling the drive of the drive means based on the predicted speed trajectory calculated by the speed trajectory prediction means,
The velocity trajectory prediction means includes
A speed measuring unit that measures a sample speed composed of a plurality of speeds at any time of the body part in the normal movable range and a speed of the body part at a boundary position between the normal movable range and the abnormal movable range;
A speed trajectory calculating unit that calculates the predicted speed trajectory as a trajectory having one convex peak upward based on the sample speed measured by the speed measuring unit;
The drive control means controls the drive of the drive means based on the predicted speed trajectory calculated by the speed trajectory calculation unit so that the body part bends the abnormally movable range in the predicted speed trajectory. The support means is operated.
According to the first aspect of the present invention, since the body part is bent in the non-normal movable range based on the movement in the normal movable range, the movement support of the body part to be supported can be performed independently. It is possible to use this apparatus without disturbing daily life. In addition, bending motions of the body part within the abnormal movable range are supported according to the predicted speed trajectory calculated based on the motion within the normal movable range. Is continuously performed without interruption, and a bending motion with a little uncomfortable feeling in the abnormal movable range can be reproduced.

In the motion support apparatus according to claim 2, the velocity trajectory calculation unit calculates a bell-shaped trajectory that is substantially symmetrical in the time axis direction, with the acceleration gradually increasing at the start point and the acceleration gradually decreasing at the end point as the predicted velocity trajectory. .
According to the second aspect of the present invention, the bell-shaped trajectory, which is a general speed waveform of a human action, is calculated as the predicted speed trajectory. It can be bent in the orbit and can realize motion support with little uncomfortable feeling.

In the motion support apparatus according to claim 3, the speed trajectory calculation unit calculates a parameter of a predicted speed function that is a quadratic or higher-order function and has one convex peak from the sample speed to calculate the predicted speed function. And a bell-type conversion region for calculating the predicted speed trajectory by converting the predicted speed function determined in the speed function determination region into the bell-shaped trajectory.
According to the invention of claim 3, since the predicted speed function is calculated from the sample speed by the speed function determination region, the predicted speed function reflecting the bending motion by the motion subject in the normal movable range can be determined. The predicted speed trajectory can be brought close to the natural speed waveform of the body part by converting the predicted speed function into a bell-shaped trajectory using the bell-shaped conversion region. Therefore, it is possible to support the movement of the body part without feeling uncomfortable within the abnormal movable range. Further, since the preset predicted speed function is determined from the sample speed, the amount of calculation required for calculating the predicted speed function and the predicted speed trajectory can be suppressed, and the influence of the time lag of the apparatus can be reduced.

In the motion support apparatus according to the fourth aspect, the predicted speed function is a quadratic function.
According to the fourth aspect of the present invention, the prediction speed function can be determined with a small number of parameters by setting the prediction speed function to a quadratic function. Therefore, it is possible to reduce the amount of calculation required for determining the predicted speed function and realize a natural support operation without a time lag.

  According to the present invention, since the motion support is performed based on the motion within the normal movable range of the motion subject, the support can be performed without hindering the activities of daily life. In addition, by calculating the predicted speed trajectory based on the motion within the normal movable range, it is possible to perform smooth motion support with less discomfort in the abnormal movable range.

It is the schematic which shows the whole structure of the operation | movement assistance apparatus which concerns on an Example. It is explanatory drawing which shows the movable range of the forearm part which concerns on an Example. It is a block diagram which shows the speed prediction means and drive control means which concern on an Example. It is an experimental example which shows the bell-shaped effect of a variable damper. It is a conceptual diagram which shows a bell-type speed orbit. It is a schematic explanatory drawing which shows the estimated speed trajectory. It is explanatory drawing which shows the function of a bell type | mold conversion area | region. It is a block diagram which shows a 1st delay element filter. It is explanatory drawing which shows a mode that an input is eased at the time of mode switching. It is an experimental result which shows the behavior of the system at the time of mode switching. It is the experimental result which verified about the accuracy of the prediction orbit.

  Next, the operation support apparatus according to the present invention will be described below with reference to the accompanying drawings by way of preferred embodiments. In addition, in an Example, the case where a support member is mounted | worn with the upper arm part and the forearm part which pinch | interpose the elbow joint in the right arm of an operation | movement main body, and the case where the bending / extending operation | movement of a forearm part is supported is illustrated.

(Overview of overall motion support device)
FIG. 1 is a schematic diagram illustrating an overall configuration of an operation support apparatus 10 according to an embodiment. The motion support apparatus 10 according to the embodiment includes a first support member 12 mounted on the upper arm A of the motion subject, and a pivotally supported pivotally supported by the first support member 12, and a forearm portion (body part) of the motion subject. ) A second support member (support means) 14 mounted on B, a drive motor (drive means) 16 that operates (rotates) the second support member 14, and a second support member 14 (forearm B). Based on the speed calculated by the speed trajectory predicting means 18, the speed trajectory predicting means 18 for calculating the speed at which the first support member 12 is bent (support speed Vt and predicted speed trajectory F (t) described later). And a drive control means 20 for driving the drive motor 16.

  Here, the movable range of the forearm B which is the main subject of movement refers to the range from the initial position of the forearm B to the movable limit position where the forearm B is bent to the upper arm A to the maximum extent. As shown in FIG. 2, the movable range of the forearm portion B in the embodiment is a normal movable range in which the operation subject can bend by its own force (input to the pressure sensor 22 described later by the operation of the forearm portion B). And an abnormally movable range in which the bending operation by itself is impossible (input to the pressure sensor 22 is impossible due to the operation of the forearm portion B).

  In the normal movable range, the motion support device 10 assists the bending motion based on the intention of the motion subject input from the forearm B. On the other hand, in the abnormally movable range in which almost no input is made by the motion subject, the motion support device 10 operates by fully automatic control until the forearm B reaches a target reaching position (described later) within the abnormally movable range. Support is to be provided. That is, the motion support device 10 operates in a manual mode that assists the motion of the forearm B in the normal movable range, and automatically mode that automatically supports the motion of the forearm B in the abnormal movable range. It is set to work with.

  In the embodiment, the second support member 14 is bent at a predetermined angle (about 90 degrees) with respect to the first support member 12 at the initial position (see FIG. 2). In addition, the target reaching position refers to an arbitrary reaching position desired by the operating subject within the abnormal movable range, and the operating subject flexes the forearm portion B from the initial position toward the target reaching position. Furthermore, the boundary between the normal movable range and the abnormal movable range is defined as a boundary position, and the angle formed by the second support member 14 that arrives at the boundary position with respect to the initial position is referred to as a switching angle. That is, the operation support device 10 switches from the manual mode to the automatic mode when the second support member 14 reaches the switching angle.

(About support members and drive motors)
The first support member 12 is attached to the upper arm portion A so as to extend along the longitudinal direction of the upper arm portion A. The first support member 12 is provided with a first fastening belt 38 configured to be elastically stretchable, and the first support member 12 is wound around the upper arm portion A by winding the first fastening belt 38 around the upper arm portion A. It is supposed to be fixed to. The second support member 14 is attached to the forearm B so as to extend along the longitudinal direction of the forearm B. The second support member 14 includes a second tightening belt 40 configured to be elastically stretchable in the same manner as the first tightening belt 38, and the second tightening belt 40 is wound around the forearm portion B to be second. The support member 14 is fixed to the forearm B. On the wrist side of the second support member 14, a ring-shaped mounting portion 24 that is mounted so as to surround the wrist is provided. The ring-shaped mounting portion 24 is provided with a plurality of thin pressure sensors 22, and these pressure sensors 22 function as force sensors that detect the force from the forearm B. Then, when the forearm portion B operated by the operating subject presses the pressure sensor 22, the pressure sensor 22 detects the force (operating force) of the forearm portion B and sends an input signal based on the force. It is like that.

  The second support member 14 is pivotally supported with respect to the first support member 12 through a shaft portion 26 at a position corresponding to the elbow joint C as a main subject of operation. The drive motor 16 in which a rotation shaft (not shown) is connected to the shaft portion 26 is provided at a connection portion between the first and second support members 12 and 14. When the drive motor 16 is driven under the control of the drive control means 20, the second support member 14 rotates with respect to the first support member 12 around the shaft portion 26, and the forearm portion B Is supposed to work. The speed (angular speed) of the second support member 14 can be changed by increasing or decreasing the rotational speed of the drive motor 16.

(Velocity trajectory prediction means)
As shown in FIG. 3, the speed trajectory predicting means 18 includes a variable impedance control unit 28 that calculates the support speed Vt in the manual mode, and a sample of the forearm B (second support member 14) in the normal movable range. A velocity measuring unit 30 for measuring velocities v ₁ , v ₂ , v ₃ (described later), and a predicted velocity trajectory F (t) based on the sample velocities v ₁ , v ₂ , v ₃ measured by the velocity measuring unit 30 And a velocity trajectory calculation unit 32 for calculation. As shown in FIG. 3, the variable impedance control unit 28 is electrically connected to the pressure sensor 22, and the support is based on an input signal from the pressure sensor 22 (operation force from the forearm B). The speed Vt is calculated. The support speed Vt is a speed calculated in real time based on an input signal from the pressure sensor 22, and in the manual mode, the second support member 14 operates at the support speed Vt. Therefore, the forearm portion B performs a bending operation at the support speed Vt calculated based on the input of the operation subject in the normal movable range.

The variable impedance control unit 28 calculates the assist speed Vt based on the torque-based impedance control law I as shown in Equation 1 (see FIG. 3). This impedance control law I employs a model in which a virtual spring is installed in the direction in which the second support member 14 (forearm B) is bent from the initial position. Therefore, in the normal movable range, the reaction force by the virtual spring is always presented to the forearm portion B via the second support member 14, and the forearm portion B is supported without feeling delayed by the operating subject. ing.
Here, θ _d is an angle with respect to the initial position of the second support member 14 (see FIG. 1), M _I and D _I are inertia and viscosity in mechanical impedance, and τ _h and τ _k are human operating force and virtual spring. The reaction force presentation by, k _amp is the force amplification factor. In addition, an angle limiter is provided, and safety is improved by adding a speed suppression coefficient (weighting coefficient) W so that it does not operate excessively (ROM) near the movable limit (initial position and movable limit position). . This weighting factor W is
And is given by the following equation:
This weight coefficient W also contributes in the vicinity of the movable limit position in the automatic mode.

Here, it is said that the mechanical impedance is preferably low in inertia and high in viscosity in a quick operation such as a bending operation of the elbow joint C. Moreover, it is known that the control ratio of viscosity with respect to operability is high. Therefore, the the variable impedance control unit 28 employs the impedance control law I where only viscosity D _I variable. Conventionally, many damper variable methods have been proposed. In this embodiment, a variable method depending on the operating force (a method for increasing or decreasing the viscosity with respect to the input amount) is employed. Specifically, as shown in Equation 4, a technique for decreasing the viscosity according to the operation force is employed.

Here, FIG. 4 illustrates an output waveform (trajectory of the support speed Vt) with respect to the input of the impedance control law I using the damper variable method of Formula 4. The graph of FIG. 4 shows, from the top, the input to the impedance control law I, the viscosity corresponding to the input, the angle with respect to the initial position of the second support member 14, and the velocity trajectory (trajectory of the assist velocity Vt) as the output. The input τ _h is a positive input (amplitude 1 [V], frequency π [rad / s]) for one cycle of the sine wave, M _I = 0.001 [kg · deg ² ], τ _k = 0 [V] It was. Further, A = 0.5, D _s = 0.2 [N · s / deg]. As is apparent from FIG. 4, the output velocity waveform has a bell shape that is convex upward and substantially symmetrical in the time axis direction. That is, the impedance control law I employing the damper variable method of Equation 4 has the effect of making the velocity waveform input from the forearm B into a bell shape.

Further, the virtual spring reaction force τ _k of Equation ₁ is used as rise delay compensation in the bending operation. By continuously presenting a small spring reaction force by the virtual spring reaction force τ _k , the operating subject can obtain a sense of being supported without being delayed from his / her own operation. The virtual spring reaction force τ _k is changed according to the following equation according to the angle θ _d .
θ _max and θ _mid are the angle of the forearm B and the maximum power output angle at the movable limit position, τ _ksi is the initial value of the spring reaction force (initial value at the initial position), and τ _ks is the spring reaction force. Is the maximum value.

(About speed measurement unit and speed trajectory calculation unit)
The speed trajectory predicting means 18 calculates a predicted speed trajectory F (t) from the bending motion of the forearm B in the normal movable range when switching from the manual mode to the automatic mode. This predicted speed trajectory F (t) is a virtual one from the initial position of the forearm B that is predicted based on the bending motion of the forearm B (second support member 14) in the normal movable range to the target arrival position. A velocity trajectory. That is, in the automatic mode (unnormal movable range), the second support member 14 operates according to the calculated predicted speed trajectory F (t), so that the forearm B is moved to the predicted speed trajectory F (t). It is made to bend along. During the automatic mode, all inputs from the forearm B to the pressure sensor 22 are set to be cut, and the operation support is not affected by the input.

  By the way, the reaching movement that reaches for a certain target generally is a basic action that is performed without particular consciousness on a daily basis, and at first glance seems to be an extremely simple movement. However, despite the infinite number of trajectories to reach the target, it has been conventionally known that the trajectory of the hand position consistently exhibits characteristic properties, and the human brain It is clear that some kind of trajectory planning is in progress. Specifically, the hand draws a substantially linear trajectory in the external world coordinate system, the hand speed in the external world coordinate system shows a substantially symmetrical bell shape, and the acceleration is zero at the start and end points of the movement. It is to become.

In this way, the brain generates a trajectory satisfying a certain property despite the infinite number of trajectories from the start point to the end point. Therefore, the brain sets some evaluation function and takes the optimum value. It is thought that the orbit was selected. A typical evaluation function is a square integral of jerk (time differentiation of acceleration, that is, three-time differentiation of hand position). Flash and Hogan et al. Have the optimal trajectory that minimizes the following evaluation function, and the velocity waveform is bell-shaped as shown in Fig. 5 (the acceleration at the start and end points is approximately 0 and approximately symmetrical in the time axis direction). (T. Flash, N. Hogan, “The coordination of arm movement: An experimentally confirmed mathematical model”, Journal of Neuroscience, 5, pp.1688-703, 1985.).

The trajectory that minimizes this evaluation function can be obtained analytically, and the relationship between the hand position and time is expressed by the following equation.
Here, s is a normalized time obtained by dividing the physical time t by the total movement time T, and (x ₀ , y ₀ ) is the coordinates of the start point (x ₁ , y ₁ ) is the coordinates of the end point.

  Based on the above knowledge, the speed trajectory predicting means 18 of the embodiment uses a bell-shaped trajectory as the predicted speed trajectory F (t) from the bending motion of the second support member 14 (forearm B) in the normal movable range. Is set so as to calculate a trajectory having one convex peak (that is, a trajectory that gradually increases at the start point and gradually decreases at the end point and is substantially symmetrical in the time axis direction). A specific method for calculating the predicted speed trajectory F (t) will be described below. As shown in Fig. 6, the operation after reaching the normal movable range from the normal movable range is as follows: (1) If it is currently accelerating, it will decelerate after further acceleration. (2) If it is currently decelerating For example, it can be divided into three categories: continue deceleration, and (3) continue constant speed if it is currently at constant speed.

Therefore, the speed measuring unit 30 includes a plurality of speeds v ₂ and v ₃ at an arbitrary time in the second support member 14 (forearm B) in the normal movable range, and the second support member 14 ( The sample velocities v ₁ , v ₂ , v ₃ consisting of the velocity v ₁ of the forearm B) are measured. Specifically, the rotation speed of the shaft portion 26 that pivotally supports the first support member 12 and the second support member 14 is measured by an encoder, and the speed (angular speed) of the second support member 14 is obtained. The velocity measuring unit 30 measures _three sample velocities v ₁ , v ₂ , and v ₃ at a sampling interval T within a prediction range (see FIG. 2) set to a normal movable range including the boundary position.

The speed trajectory calculation unit 32 calculates the predicted speed trajectory F (t) from the sample speeds v ₁ , v ₂ , v ₃ . As shown in FIG. 3, the speed trajectory calculation unit 32 includes a speed function determination area 34 and a bell-type conversion area 36, and a predicted speed trajectory F (t) is calculated through both areas 34 and 36. The In the speed function determination area 34, a predicted speed function E (t) that approximates the bell-shaped trajectory is set. This predicted speed function E (t) is a function of quadratic or higher and has one convex peak, and in the embodiment, a quadratic function is set as shown by the following equation.
Then, the speed function determination area 34 calculates parameters a, b, and c that satisfy Equation 10 from the _three sample speeds v ₁ , v ₂ , and v ₃ to determine the predicted speed function E (t). When the predicted motion is (3) above, it can be predicted that the forearm B continues to move at a constant speed, but it cannot be predicted at which timing the forearm B begins to decelerate. Therefore, in the embodiment, the constant velocity movement is continued for a certain time and then decelerated, which is defined by the following equation.

  Here, the predicted speed function E (t) approximates a bell-shaped speed waveform, which is preferably a bell shape, to a quadratic function. Therefore, if this is assumed to be a predicted speed trajectory F (t), the forearm in the abnormal movable range is used. The bending motion of the part B is a strange motion. Therefore, in the bell-type conversion area 36, the predicted speed function E (t) (quadratic function) determined in the speed function determination area 34 is converted into a bell-shaped trajectory, and the converted trajectory is converted into the predicted speed trajectory F. It is calculated as (t). Specifically, the bell-type conversion region 36 converts the predicted speed function E (t) into a bell shape by using the bell-type effect by the variable damping of the impedance control law I described above. That is, as shown in FIG. 7, the bell-type transformation region 36 solves the impedance control law I (see Equation 1) in reverse using the predicted speed function E (t), so that the predicted speed function E ( t) is once converted into a force to generate a virtual operation force (predicted operation force (force in FIG. 7)). By assuming this predicted operating force as an input and introducing it into the impedance control law I, the bell-type conversion region 36 converts the predicted speed function E (t) into a bell-type due to the bell-type effect of the variable damper. Then, the predicted speed function E (t) bell-shaped in the bell-shaped conversion area 36 becomes the predicted speed trajectory F (t).

  Here, in the motion support device 10 according to the embodiment, when the movement exceeding the switching angle occurs, the operation supporting device 10 switches to the automatic mode, and the predicted operation force becomes 0 (when the target reaching position is reached). Is set to switch to manual mode. However, in the abnormally movable range, the main subject of movement is a range in which the force cannot be exerted originally, and thus the second support member 14 is in a state of supporting all the load of the forearm portion B. That is, the forearm portion B in the abnormal movable range is in a state in which the pressure sensor 22 is constantly given input in the extension direction (direction in which the arm is extended). Therefore, in this state, when the automatic mode is switched to the manual mode, a sudden input such as a step input is given to the pressure sensor 22 and it is dangerous. Therefore, in the operation support device 10 of the embodiment, the input cancellation using the first-order lag element is performed to avoid sudden input to the pressure sensor 22 (see FIGS. 8 and 9). . In other words, by setting the applied input at the moment of switching from the automatic mode to the manual mode by delay output using the first-order delay element, the operation force is set to be relaxed for the delay time until there is no residual input.

  The drive control means 20 drives the drive motor 16 based on the support speed Vt and the predicted speed trajectory F (t) calculated by the speed trajectory prediction means 18. That is, in the manual mode, the speed trajectory predicting means 18 calculates the support speed Vt in real time based on the input from the forearm B, and the drive control means 20 drives the drive motor 16 based on the support speed Vt. Control. Then, the drive motor 16 rotates the second support member 14 at the support speed Vt so that the forearm B is bent at the support speed Vt. On the other hand, in the automatic mode, the drive control means 20 controls the drive motor 16 based on the predicted speed trajectory F (t) calculated by the speed trajectory calculation unit 32. Then, the drive motor 16 is configured to rotate the second support member 14 in accordance with the predicted speed trajectory F (t) so that the forearm portion B bends in the predicted speed trajectory F (t). .

(Operation of Example)
Next, the operation of the operation support apparatus 10 according to the embodiment will be described. The moving subject sets the first support member 12 and the second support member 14 to the upper arm portion A and the forearm portion B, respectively, and extends the forearm portion B to the initial position. Next, the moving subject determines an arbitrary target reaching position within the abnormal movable range, and bends the forearm B toward the target reaching position.

  An operating force is input to the pressure sensor 22 through the forearm portion B by the bending operation of the forearm portion B by the moving subject. Then, the pressure sensor 22 sends the input signal to the variable impedance control unit 28, and the control unit 28 calculates the support speed Vt. That is, an input signal is input to the impedance control law I, and the assist speed Vt at the time is calculated. The support speed Vt calculated by the variable impedance control unit 28 is sent to the drive control means 20, and the drive control means 20 controls the drive of the drive motor 16 based on the support speed Vt. Then, the drive motor 16 rotates the second support member 14 so that the forearm portion B bends at the support speed Vt. As described above, in the manual mode, the variable impedance control unit 28 calculates the support speed Vt in real time based on the input from the forearm B, and the drive motor 16 operates at the support speed Vt calculated at each time. 2 The support member 14 is rotated.

As described above, the motion support device 10 according to the embodiment calculates the support speed Vt according to the input from the operation subject in the manual mode, and assists the bending motion of the forearm B. . In addition, since the viscosity D _I (see Equation 4) that decreases according to the operation force (input) of the operation subject is adopted, the input of the operation subject is corrected to a bell-shaped velocity trajectory even in the manual mode, and the forearm portion B ( The second support member 14) can be bent by natural movement. In addition, since the virtual spring reaction force is used as the rise delay compensation, the moving subject can obtain a sense of being supported without delay with respect to his operation.

Next, when the forearm B (second support member 14) reaches the predicted range, the velocity measuring unit 30 measures three sample velocities v ₁ , v ₂ , and v ₃ including the velocity v ₁ at the boundary position. To do. The measured sample velocities v ₁ , v ₂ , v ₃ are sent to the velocity function determination area 34, which predicts using the sample velocities v ₁ , v ₂ , v ₃ using Equation 10 Parameters a, b, and c of the speed function E (t) (secondary function) are calculated. When the parameters a, b, and c are calculated and the predicted speed function E (t) is determined, the predicted speed function E (t) is sent to the bell-type conversion area 36. Note that when the sample velocities v ₁ , v ₂ , and v ₃ are substantially equal (when the second support member 14 is moving at a constant speed), the speed function determination region 34 is the predicted speed function shown in Equations 11 and 12. E (t) is determined.

  In the bell-shaped conversion region 36, as shown in FIG. 7, the impedance control law I is calculated back using the calculated predicted speed function E (t), and the prediction corresponding to the predicted speed function E (t) is performed. Calculate the operating force. Then, by introducing the predicted operation force once calculated into the impedance control law I, the predicted speed function E (t) is bell-shaped. That is, the bell-type conversion region 36 uses the bell-type effect due to the variable damping of the impedance control law I to convert the predicted speed function E (t), which is a quadratic function, into a bell-type trajectory that is a natural human action. Then, the trajectory is calculated as the predicted speed trajectory F (t).

  The predicted speed trajectory F (t) calculated by the speed trajectory calculation unit 32 is sent to the drive control means 20, and the drive control means 20 drives the drive motor 16 based on the predicted speed trajectory F (t). Control. Then, the drive motor 16 causes the second support member 14 to rotate in accordance with the predicted speed trajectory F (t) within the abnormally movable range. That is, the forearm portion B bends along the predicted speed trajectory F (t) with the assistance of the second support member 14 along the abnormal speed range. When the predicted speed trajectory F (t) becomes 0, the drive motor 16 stops the second support member 14 and the forearm B stops at a predetermined position (angle). At this time, the forearm B is stopped according to the predicted speed trajectory F (t) calculated based on the operation of the second support member 14 (the forearm B) in the normal movable range. Will stop near the target position determined by the operating subject.

  As described above, according to the motion support device 10 according to the embodiment, since the support based on the bending motion in the normal movable range is performed, the left and right forearm portions are similar to the master-slave method described in the conventional example. Do not perform the same bending motion, and the motion support device 10 can be used for motion support in daily life. Further, since the forearm B is bent in the abnormal movable range according to the predicted speed trajectory F (t) calculated based on the movement of the forearm B in the normal movable range, the bending operation is abnormal from the normal movable range. It is possible to move to the movable range without interruption and provide operation support with less sense of incongruity.

  Since the predicted speed trajectory F (t) calculated by the speed trajectory predicting means 18 is approximated to a bell-shaped trajectory based on a natural human motion, the bending motion of the forearm portion B supported in the abnormal movable range. Becomes a natural movement, and a higher rehabilitation effect can be obtained by the action subject being aware of the natural movement. In addition, by calculating the bell-shaped predicted speed trajectory F (t), it is possible to predict the target arrival position desired by the operating subject with high accuracy. Furthermore, in the motion support apparatus 10 according to the embodiment, the predicted speed function E (t) is a quadratic function with few parameters, so that the amount of calculation required for calculating the predicted speed trajectory F (t) can be reduced, and the normal movable range. Operation support with little discomfort can be realized at the time of transition from the normal to the non-normal movable range.

  When the second support member 14 is stopped, the operation support device is switched from the automatic mode to the manual mode, and input from the forearm portion B to the pressure sensor 22 becomes possible. However, since the forearm B, which is the main subject of the operation, hardly receives any force in the abnormally movable range, the load on the forearm B is supported by the second support member 14, and a large input is input to the pressure sensor 22. It has been added. However, in the motion support device 10 according to the embodiment, the applied input at the moment of switching from the automatic mode to the manual mode is relaxed by performing input cancellation using the first-order lag element (see FIGS. 8 and 9). ). Therefore, the second support member 14 does not suddenly extend when the mode is switched, and the forearm portion B can be slowly extended to be safe.

(Verification experiment 1)
Next, a verification experiment was performed in order to confirm the system behavior of the operation support apparatus 10 according to the example. FIG. 10 shows the behavior of the system when the mode is switched. From the upper part of the graph, the input to the impedance control law I, the operating force of the forearm B by the subject, the predicted operating force by trajectory estimation, automatic control determination, angle and angular velocity. The impedance parameters were M = 0.0002, D ₀ = 0.5, A = 0.5, τ _ks = 0.05, τ _ksi = 0.01, and the switching angle was 80 [deg]. In order to make the switching state easy to understand, once the automatic mode is entered, the automatic mode state is set to be maintained for 3 seconds regardless of the predicted operation force. From FIG. 10, it can be seen that when switching from the manual mode to the automatic mode, there is no discontinuity, and the velocity waveform forms a substantially bell shape. During the automatic mode, no matter how much force a person applies, it is not reflected in the input, and even if the input continues to be given at the moment of switching to the manual mode, no input is suddenly added to the system, and it is slowly System input is being generated.

(Verification experiment 2)
FIG. 11 shows the results of an experiment that verified the accuracy of the predicted trajectory. The content of the experiment is that the target reaching position is set to 60, 70, 80, 70, 90 [deg], and the arm is moved so as to reach the target reaching position. The switching angle is 60 [deg], and the sensor input in the bending direction is cut off at an angle larger than that. From the upper part of the graph, there are input to the impedance control law I, predicted operation force by trajectory estimation, automatic control determination, angle, and angular velocity. The impedance parameter is the same as in verification experiment 1. The target arrival position is reached with an error of about ± 5 [deg], and it can be said that the target arrival position can be predicted well. Both the system input and the velocity waveform are continuous and smooth, and are almost symmetrical in the time axis direction, achieving a very natural reaching movement.

The operation support apparatus according to the present invention is not limited to the above embodiment, and the following modifications are possible.
(1) In the embodiment, the motion support device is configured to support the forearm portion of the upper limb. However, for example, the rotation of the upper arm portion may be supported around the shoulder joint of the motion subject. It is also possible to support the movement of the lower limbs (for example, the lower leg) by the movement support device.
(2) In the embodiment, the motion support apparatus is configured to support the forearm portion of the motion subject over the entire movable range (normal movable range and abnormal movable range). However, for example, when the normal movable range can be normally bent by its own force, and the self-moving operation cannot be performed only in the abnormal movable range, it is only necessary to provide support in the automatic mode only in the abnormal movable range. Assistance in the normal movable range (manual mode) is unnecessary.
(3) In the embodiment, the operation support by the impedance control is performed in the normal movable range (manual mode), but the operation support in the normal movable range may be performed by another known control method.
(4) In the embodiment, the prediction speed function is set to a quadratic function. However, a cubic function, a quadratic function, or the like may be adopted as long as the function has one convex peak. For example, when a predicted speed orbit function of a cubic function is adopted, four sample speeds for determining the predicted speed function are required. Further, the velocity trajectory predicting means may directly calculate a bell-shaped velocity trajectory that optimizes the evaluation function shown in Equation 7. In this case, the bell type conversion area is not necessary.
(5) In the embodiment, the bell-type transformation region calculates the predicted operating force by calculating back the impedance control law, and then introduces the predicted operating force into the impedance control law again, so that the predicted speed function (secondary function) Was made bell-shaped. However, for example, the predicted speed function may be bell-shaped using a first-order lag element as shown in FIGS.
(6) Although the variable impedance control unit is provided in the velocity trajectory prediction unit in the embodiment, the variable impedance control unit may be configured separately.

14 Second support member (support means), 16 Drive motor (drive means)
18 speed trajectory prediction means, 20 drive control means, 30 speed measurement section 32 speed trajectory calculation section, 34 speed function determination area, 36 bell type conversion area B forearm, F (t) predicted speed function, C elbow joint (joint)
v ₁ , v ₂ , v ₃ sample rate, E (t) predicted velocity function, a, b, c parameters

Claims (4)

A support means (14) mounted on a body part (B) that bends around a joint (C) as an axis, and a drive means (16) that operates the support means (14). An operation support apparatus that drives the drive means (16) in a non-normal movable range in which normal operation is impossible, and causes the body part (B) to bend by the support means (14),
Based on the bending motion of the body part (B) in the normal movable range in which the normal motion based on the force of the motion subject is possible, any desired in the non-normal motion range desired by the operating subject from the initial position in the normal motion range A speed trajectory prediction means (18) for calculating a predicted speed trajectory (F (t)) of the body part (B) to the target reaching position;
Drive control means (20) for controlling the drive of the drive means (16) based on the predicted speed trajectory (F (t)) calculated by the speed trajectory prediction means (18),
The velocity trajectory predicting means (18)
From a plurality of speeds (v ₂ , v ₃ ) at any time of the body part (B) in the normal movable range and the speed (v ₁ ) of the body part at the boundary position between the normal movable range and the abnormal movable range A velocity measuring unit (30) for measuring the sample velocity (v ₁ , v ₂ , v ₃ ),
Based on the sample velocities (v ₁ , v ₂ , v ₃ ) measured by the velocity measuring unit (30), the predicted velocity trajectory (F (t)) is calculated as a trajectory having one upward convex peak. A speed trajectory calculation unit (32),
The drive control means (20) controls the drive of the drive means (16) based on the predicted speed trajectory (F (t)) calculated by the speed trajectory calculation section (32), and sets the abnormal movable range. The motion support device, wherein the drive means (16) activates the support means (14) so that the body part (B) bends in the predicted speed trajectory (F (t)).   The velocity trajectory calculation unit (32) calculates a bell-shaped trajectory that is substantially symmetrical in the time axis direction as the predicted velocity trajectory (F (t)), with acceleration gradually increasing at the start point and gradually decreasing at the end point. The operation support apparatus according to claim 1. The velocity trajectory calculation unit (32) sets parameters (a, b, c) of a predicted velocity function (E (t)) that is a quadratic or higher-order function and has one convex peak to the sample velocity ( v ₁ , v ₂ , v ₃ ) and calculating the predicted speed function (E (t)) to determine the predicted speed function (34), and the predicted speed function determined in the speed function determined area (34) The motion support apparatus according to claim 2, further comprising: a bell-type conversion region (36) for converting (E (t)) into the bell-type trajectory to calculate the predicted speed trajectory (F (t)).   The motion support apparatus according to claim 3, wherein the predicted speed function (E (t)) is a quadratic function.