JP4601691B2 - Wearable movement assist device calibration device and calibration program - Google Patents

Description

  The present invention relates to a calibration apparatus for a wearable movement assist apparatus that assists or substitutes for a wearer's movement, and an improvement in a calibration program for the wearable movement assist apparatus.

  For a disabled person who has lost muscular strength or an elderly person who has lost muscular strength, it is often very difficult even for a healthy person to perform a simple operation. For this reason, various power assist devices have been developed today in order to assist or substitute for the operations of these people.

  As these power assist devices, for example, there are wearable motion assist devices (hereinafter simply referred to as “motion assist devices”) worn by users (hereinafter referred to as “wearers”).

  This type of motion assisting device is attached to the wearer's skin and has a myoelectric potential sensor (detection means) for detecting a myoelectric potential signal (biological signal) accompanying the muscle activity of the wearer, and assists the wearer. The thing of the structure provided with the actuator (drive source) for providing motive power is being developed (for example, nonpatent literature 1).

  This motion assisting device is characterized in that an actuator such as a motor is driven based on the detection result of the detecting means, and the actuator is computer-controlled so as to apply assist force (auxiliary power) according to the wearer's intention. For this reason, the motion assisting device can apply an assist force based on the wearer's intention to the wearer, and provides an assist force necessary for the wearer's operation so as to be linked to the wearer's operation. be able to.

  By the way, in the above-described motion assisting device, for example, a control signal having a predetermined correlation with a signal obtained by amplifying the detection signal of the myoelectric potential sensor is supplied to a driver circuit that controls the actuator, whereby the myoelectric potential generated by the wearer is generated. The assist force is generated so as to satisfy the required correlation with the signal.

  That is, the myoelectric potential signal and power accompanying the muscle activity of the wearer have a positive correlation with each other, and the ratio of the magnitudes of each other is a predetermined value. It is necessary to give the assist power according to the response. In other words, if the assisting force by the motion assisting device does not satisfy the required relationship with respect to the myoelectric potential signal, the assisting power applied to the wearer may be excessively large or small, and the convenience for the user may be significantly impaired. .

  The myoelectric potential signal generated by the wearer is a weak electric signal, and the proportional relationship between the myoelectric potential signal and the muscular force generated by the myoelectric potential signal varies depending on the individual, and even in the same person, the electrical resistance of the skin depends on the physical condition of the day. Since the value is not constant, the myoelectric potential signal and the muscle force generated by the myoelectric potential signal are often not constant. For this reason, the motion assisting device includes a so-called calibration device in order to correct the control amount to the actuator by multiplying the control signal by a predetermined coefficient (parameter). Specifically, a calibration device is provided that performs correction using a coefficient (parameter) by associating the myoelectric signal and the assist force with a required relationship when the motion assisting tool is worn by the wearer. This calibration device acquires a myoelectric potential signal when a predetermined amount of load is applied to the wearer, and corrects it so that it can be derived by changing the coefficient (parameter) from the correspondence between these loads and the myoelectric potential signal. It is supposed to be configured.

  In this calibration device, when the load on the wearer is changed in stages and muscle strength is generated in the wearer so as to antagonize the load at each stage, the correspondence between the load at each stage and the myoelectric potential signal Based on this, the myoelectric potential signal and the auxiliary power can be associated with the required relationship.

  Here, as a method of changing the load on the wearer step by step, a weight having a different weight is prepared in advance and replaced with another weight each time a surface myoelectric potential is detected, or a coil spring is used. It is possible to consider a method in which the amount of extension of the coil spring is changed stepwise.

In the motion assisting device provided with the calibration device to which these methods are applied, the myoelectric potential signal and the assisting power can be reliably associated with the required relationship as necessary. The situation of becoming too small can be prevented.
Takao Nakai, Suwoong Lee, Hiroaki Kawamoto and Yoshiyuki Sankai, "Development of Power Assistive Leg for Walking Aid using EMG and Linux," Second Asian Symposium on Industrial Automation and Robotics, BITECH, Bangkok, Thailand, May 17-18, 2001

  By the way, in the above-mentioned power assist device, as described above, the myoelectric potential sensor for detecting the myoelectric potential signal is directly pasted on the wearer's skin, and the surface myoelectric potential is detected through the skin. For this reason, for example, even if the wearer is the same, if the attachment position of the myoelectric potential sensor is shifted or the physical condition changes, the electrical resistance value is different or changed, and the detection sensitivity of the myoelectric potential signal is changed. Therefore, it is necessary to perform the above-described correction (calibration) every time it is mounted. For this reason, in the above-described calibration method, each time the motion assisting tool is worn on the wearer, the weights with different weights are replaced many times, or a coil spring is attached to change the extension amount of the coil spring stepwise. Such a complicated work is forced on the wearer.

  As described above, in the conventional calibration method, the work required for one wearer for calibration becomes extremely complicated, and it takes a considerable time until the calibration is completed, and the strength is weak. This places an extra burden on the wearer, and for these reasons, there has been a problem that it is a major limitation in the practical application and popularization of motion assist devices.

  Therefore, in view of the above circumstances, the present invention provides a calibration apparatus and calibration program for a wearable motion assisting device that can reduce the complexity of the work required by the wearer for the calibration of the wearable motion assisting device. The purpose is to provide.

In order to solve the above problems, the present invention has the following means.
(1) The present invention of claim 1 is a detection means for detecting a potential as a biological signal generated in accordance with the operation of the wearer;
An operation assisting wearing device having a drive source for providing auxiliary power to the wearer;
A calibration device for a wearable motion assisting device having control means for controlling the drive source by performing arithmetic processing so as to generate auxiliary power according to a detection signal detected by the detecting means,
The detection means detects a potential as a biological signal when generating a muscle force against a driving force as a load applied stepwise from the driving source, and the wearer generates in response to the detection signal. Estimating means for estimating a virtual torque to be
Calibration means for updating a parameter for correcting the auxiliary power by the arithmetic processing based on the relationship between the virtual torque estimated by the estimation means and the detection signal ;
It is provided with.
(2) In the present invention of claim 2, the calibration means includes:
When the wearer wearing the motion assisting wearing device is in a stationary state, the detection signal detected by the detection unit and the virtual torque due to the muscle force according to the muscle strength against the input torque as the load input from the drive source First storage means for storing the correspondence as initial setting data;
Comparing the input torque as the load input from the drive source and the estimated torque due to the muscular force corresponding to the detection signal detected by the detection means in the process in which the wearer performs a predetermined operation to counter the load A second storage means for storing the correspondence relationship between the differences as reset data;
The virtual torque corresponding to the detection signal and the input torque coincide with each other based on the correspondence relationship between the detection signal stored in the first storage means and the virtual torque each time the movement assisting wearing device is mounted on the wearer. First parameter setting means for setting parameters as follows:
Second parameter setting means for setting a parameter such that the estimated torque corresponding to the detection signal and the input torque match based on the difference between the input torque stored in the second storage means and the estimated torque;
It is provided with.
(3) The present invention of claim 3 has phase specifying means for specifying a phase constituting a partial operation included in the predetermined operation performed by the wearer,
The second parameter setting means sets the parameter so that the estimated torque corresponding to the detection signal matches the input torque based on the correspondence relationship between the estimated torque and the input torque in the phase specified by the phase specifying means. It is characterized by doing.
(4) In the present invention of claim 4, the correspondence stored in the first storage means is such that the auxiliary power has a positive correlation with the detection signal,
The correspondence relationship stored in the second storage means is a relationship between a change in the detection signal and a change in the auxiliary power in the predetermined operation.
(5) The present invention of claim 5 is an operation having a drive source for generating auxiliary power in accordance with a detection signal obtained from a detection means for detecting a potential as a biological signal generated in accordance with the operation of the wearer. A calibration program for a wearable motion assisting device that causes a computer to execute calibration for associating the biological signal and the auxiliary power with a required relationship each time an auxiliary wearing tool is worn on the wearer.
In the computer,
When the wearer wearing the motion assisting wearing device is in a stationary state, the detection signal detected by the detection unit and the virtual torque due to the muscle force according to the muscle strength against the input torque as the load input from the drive source A first procedure for storing the correspondence as initial setting data;
The detection means detects a potential as a biological signal when generating a muscle force against a driving force as a load applied stepwise from the driving source, and the wearer generates in response to the detection signal. A second procedure for estimating the virtual torque to be
An input torque as the load inputted from the drive source, before comparing the Ki推 constant torque, and a third step of storing the correspondence between both by the difference as resetting data,
Each time the motion assisting wearing device is worn by the wearer, the virtual torque corresponding to the detection signal and the input torque are matched based on the correspondence relationship between the detection signal stored in the first procedure and the virtual torque. A fourth procedure for setting parameters in
A fifth procedure for setting parameters so that the estimated torque corresponding to the detection signal and the input torque match based on the difference between the input torque stored in the third procedure and the estimated torque;
Is executed.

  According to the present invention, the detection means detects the potential as a biological signal when generating muscle force against the driving force as a load applied stepwise from the driving source, and wears corresponding to this detection signal Estimates the virtual torque generated by the user, and updates parameters for correcting auxiliary power through arithmetic processing based on the relationship between the virtual torque and the detection signal, greatly reducing the labor and time required for calibration This makes it possible to further promote the practical application and spread of the wearable movement assist device.

  In addition, it does not impose an extra burden to calibrate a wearer with weakened muscle strength, and the wearer simply performs a simple operation when the operation assisting device is worn on the wearer. The calibration is automatically performed, the correction value according to the wearer's condition is set, and the driving force based on the wearer's myoelectric potential signal can be accurately applied in conjunction with the wearer's operation. It becomes possible.

  Therefore, when performing calibration, an assist force according to the intention of the wearer is applied from the drive source, and the assist force does not become excessive or small, and the operation of the wearer is stably assisted. The reliability of the wearable movement assist device can be further increased.

  In particular, the wearer can calibrate with peace of mind even in a situation where it is difficult to use the attached motion assisting wearer as expected, such as when the wearer is a beginner. Therefore, even for physically disabled people who cannot move freely, it is possible to perform calibration without avoiding the wearer's physical disadvantages without performing special operations. Thus, calibration can be performed so as to compensate for the physical weakness of the wearer.

  Hereinafter, an embodiment of the present invention will be described with reference to the drawings.

  FIG. 1 is a block diagram showing a control system of a wearable movement assist apparatus to which an embodiment of a calibration apparatus according to the present invention is applied. As shown in FIG. 1, the control system of the movement assist device 10 detects a driving source 140 that applies assist force to the wearer 12 and a joint angle (physical phenomenon) corresponding to the movement of the wearer 12. A physical phenomenon detection means 142 that detects the muscle potential (biological signal) corresponding to the muscle force generated by the wearer 12.

  The data storage unit 146 includes a calibration database 148 for correcting the parameter of the command signal (control signal) according to the detection sensitivity of the myoelectric potential (biological signal) with respect to the muscular force generated by the wearer 12, and the command signal database 150. And are stored. As will be described later, the calibration database 148 stores the first power (muscle strength) and biological signal (myoelectric potential signal) generated by the wearer 12 wearing the motion assisting wearing tool 18 (see FIGS. 2 and 3). A first storage area (first storage means) that stores the correspondence relationship between the first and second correspondences between the power (muscle strength) and the biological signal (myoelectric potential signal) generated in the course of the wearer 12 performing a predetermined basic motion And a second storage area (second storage means) in which the relationship is stored in advance.

  The joint angles (θknee, θhip) detected by the physical phenomenon detection unit 142 and the myoelectric potential signals (EMGknee, EMGhip) detected by the biological signal detection unit 144 are input to the calibration database 148 and the command signal database 150.

  The control device 100 includes a phase specifying unit 152, a difference deriving unit 154, a parameter correcting unit 156, a control unit 160, a calibration control unit 162, and a load generating unit 164. Then, each time the motion assisting wearing device 18 is worn on the wearer 12, the calibration control means 162 performs the first correspondence based on the biological signal generated in the basic motion by the wearer 12 and the second correspondence relationship. The auxiliary power is corrected according to the biological signal so as to satisfy the relationship.

  That is, the calibration control means 162 executes the calibration control process when the wearer 12 wears the motion assisting wearing tool 18 and the power switch is turned on, and the load generating means 164 causes the power amplifying means 158. In contrast, the driving force from the driving source 140 is applied to the wearer 12 stepwise as a load (input torque), and the wearer 12 is caused to generate muscular strength so as to antagonize this driving force.

  Thereafter, the wearer 12 to which the driving force from the driving source 140 is applied performs a predetermined calibration operation (for example, task A: an operation of getting up from a seated state) to generate muscle force from the skeletal muscle. As a result, the physical phenomenon detection unit 142 detects the joint angle along with the calibration operation, and the biological signal detection unit 144 detects the myoelectric potential signal.

  Then, the phase identification unit 152 identifies the phase of the calibration operation pattern of the wearer 12 by comparing the joint angle detected by the physical phenomenon detection unit 142 with the joint angle stored in the calibration database 148.

  Further, the difference deriving unit 154 starts the calibration control process, and the load (input torque) of the driving source 140 applied by the load generating unit 164 and the myoelectric potential signal (actually measured value) detected by the biological signal detecting unit 144. ) And the muscular strength (estimated torque) corresponding to each other, and the difference between the two is obtained to obtain the second correspondence.

  The parameter correction unit 156 also uses the first correspondence relationship based on the difference between the load (input torque) and the muscle strength (estimated torque) calculated by the difference deriving unit 154 in the phase specified by the phase specifying unit 152. The parameter K is corrected so as to satisfy The reference parameter is not corrected when there is no difference between the input torque from the drive source 140 applied by the load generating unit 164 and the muscle force corresponding to the myoelectric potential signal (actually measured value) detected by the biological signal detecting unit 144. . However, if there is a difference between the input torque from the drive source 140 applied by the load generating means 164 and the muscle force corresponding to the myoelectric potential signal (actually measured value) detected by the biological signal detecting means 144, the two match. The parameter K is corrected as follows. At this time, the correction parameter K ′ is set so that the input torque and the estimated torque are equal (second parameter setting means).

  Then, the calibration control unit 162 sets the parameter corrected by the parameter correction unit 156 as the parameter of the wearer 12, and performs calibration for the next phase.

  As described above, since the driving source 140 is controlled so as to generate the assist force according to the biological signal detected by the biological signal detection unit 144 using the parameters set by the calibration, the state of the wearer 12 of the day Regardless of the (skin resistance value) or the displacement of the attachment position of the biological signal detection means 144, the muscle force and the assist force can be controlled to maintain a predetermined ratio of 1: 1, for example.

  Further, the control means 160 is supplied with the joint angle (θknee, θhip) detected by the physical phenomenon detection means 142 and the myoelectric potential signal (EMGknee, EMGhip) detected by the biological signal detection means 144, and the joint angle and The assist force from the driving source 140 for each phase according to the myoelectric potential signal is calculated using the correction parameter K ′ set by the calibration control means 162, and the command signal obtained from the calculation result is the power amplifying means. 158.

  Here, a specific configuration example of the wearable movement assist device 10 will be described in detail.

  FIG. 2 is a perspective view of a state in which one embodiment of the wearable movement assisting device is worn as seen from the front side. FIG. 3 is a perspective view of a state in which one embodiment of the wearable motion assisting device is worn as seen from the rear side.

  As shown in FIGS. 2 and 3, the motion assisting device 10 is self-supporting, for example, a person with lower limb movement dysfunction who cannot walk due to a decrease in skeletal muscle strength, or a patient who performs rehabilitation of walking movement. It is a device that assists the walking movement of people who have difficulty walking, and it detects the biological signal (surface myoelectric potential) that is generated when the muscle force is generated by the signal from the brain, and from the actuator based on this detection signal Operates to apply driving force.

  When the wearer 12 wearing the motion assisting device 10 performs a walking motion with his / her own intention, a driving torque corresponding to the biological signal generated at that time is applied from the motion assisting device 10 as an assisting force. It is possible to walk with half the strength of the muscle strength required. Therefore, the wearer 12 can walk while supporting the overall weight by the resultant force of his / her muscle strength and the driving torque from the actuator (in this embodiment, an electric driving motor is used).

  At that time, the motion assisting device 10 controls the assisting force (motor torque) applied according to the movement of the center of gravity accompanying the walking motion to reflect the intention of the wearer 12 as described later. Therefore, the actuator of the motion assisting device 10 is controlled so as not to apply a load that is against the intention of the wearer 12 and is controlled so as not to hinder the operation of the wearer 12.

  In addition to the walking motion, the motion assisting device 10 can assist, for example, an operation when the wearer 12 stands up from a state of sitting on a chair, or an operation when sitting on the chair from a standing state. Furthermore, power assist can be performed even when the wearer 12 goes up or down the stairs. In particular, when the muscular strength is weak, it is difficult to move up the stairs or move up from the chair, but the wearer 12 wearing the movement assisting device 10 is given drive torque according to his / her intention. Therefore, it becomes possible to operate without worrying about the decrease in muscle strength.

  Here, an example of the configuration of the motion assisting device 10 will be described. As shown in FIGS. 2 and 3, the motion assisting device 10 is provided with an actuator (corresponding to the drive source 140) in the motion assisting wearing tool 18 to be worn by the wearer 12. Actuators include a right thigh drive motor 20 located at the right hip joint of the wearer 12, a left thigh drive motor 22 located at the left hip joint of the wearer 12, and a right knee drive motor located at the right knee joint of the wearer 12. 24 and a left knee drive motor 26 located at the left knee joint of the wearer 12. These drive motors 20, 22, 24, and 26 are drive sources composed of servo motors whose drive torque is controlled by a control signal from a control device, and a reduction mechanism (not shown) that reduces motor rotation at a predetermined reduction ratio. And a sufficient driving force can be applied although it is small.

  In addition, batteries 32 and 34 that function as a power source for driving the drive motors 20, 22, 24, and 26 are attached to the waist belt 30 attached to the waist of the wearer 12. The batteries 32 and 34 are rechargeable batteries, and are distributed on the left and right so as not to hinder the walking motion of the wearer 12.

  In addition, the control bag 36 to be worn on the back of the wearer 12 houses devices such as a control device, a motor driver, a measuring device, and a power circuit, which will be described later. The lower portion of the control back 36 is supported by the waist belt 30 and attached so that the weight of the control back 36 does not become a burden on the wearer 12.

  The motion assisting device 10 includes surface myoelectric potential sensors 38a and 38b that detect surface myoelectric potential (EMGhip) accompanying the movement of the right thigh of the wearer 12, and surface myoelectric potential ( EMGhip), myoelectric sensors 42a and 42b for detecting the surface myoelectric potential (EMGknee) associated with the movement of the right knee, and the surface myoelectric potential (EMGknee) associated with the movement of the left knee EMG sensors 44a and 44b are provided.

  These myoelectric potential sensors 38a, 38b, 40a, 40b, 42a, 42b, 44a, 44b are detection means for measuring the surface myoelectric potential when the skeletal muscles generate muscle force, and are weak potentials generated in the skeletal muscles. Has an electrode (not shown). In the present embodiment, each myoelectric potential sensor 38a, 38b, 40a, 40b, 42a, 42b, 44a, 44b is attached so as to be attached to the skin surface of the wearer 12 with an adhesive seal covering the periphery of the electrode. .

  In the human body, acetylcholine, a synaptic transmitter, is released on the surface of muscles that form skeletal muscles by commands from the brain. As a result, the ionic permeability of muscle fiber membranes changes and action potential (EMG: Electro MyoGram Myoelectricity) appear. The action potential causes contraction of muscle fibers and generates muscle force. Therefore, by detecting the myoelectric potential of the skeletal muscle, it becomes possible to estimate the muscle force generated during the walking motion, and the assist force necessary for the walking motion can be obtained from the virtual torque based on the estimated muscle strength. It becomes possible.

  Muscles expand and contract when proteins called actin and myosin are supplied by blood, but muscles are generated when they contract. For this reason, in a joint in which two bones are connected so as to be rotatable, a flexor that generates a force in the direction of bending the joint and an extensor that generates a force in the direction of extending the joint are mounted between the two bones. It is built.

  The human body has several muscles to move the legs from the waist down, the iliopsoas muscle that raises the thigh forward, the gluteal muscle that lowers the thigh, the quadriceps muscle to extend the knee, and the knee. There are biceps femoris.

  The myoelectric potential sensors 38a and 40a are affixed to the front side of the base of the thigh of the wearer 12, and measure the myoelectric potential corresponding to the muscle strength when the leg is pushed forward by detecting the surface myoelectric potential of the iliopsoas muscle. .

  The myoelectric potential sensors 38b and 40b are affixed to the buttocks of the wearer 12, and measure the myoelectric potential according to the muscle force when, for example, the kicking force or the stairs rises by detecting the surface myoelectric potential of the gluteus To do.

  The myoelectric potential sensors 42a and 44a are attached to the front side of the wearer 12 above the knee, detect the surface myoelectric potential of the quadriceps femoris muscle, and measure the myoelectric potential according to the muscle force that moves downward from the knee.

  The myoelectric potential sensors 42b and 44b are attached to the rear side of the wearer 12 above the knee, detect the surface myoelectric potential of the biceps femoris muscle, and measure the myoelectric potential according to the muscle force to return the knee back.

  Therefore, in the motion assisting device 10, the four drive motors 20, 22, 24, 26 are based on the surface myoelectric potentials detected by these myoelectric potential sensors 38a, 38b, 40a, 40b, 42a, 42b, 44a, 44b. The driving current to be supplied to the vehicle is obtained, and the driving motors 20, 22, 24, and 26 are driven by this driving current, so that an assist force is applied to assist the walking motion of the wearer 12.

  In addition, in order to smoothly move the center of gravity by walking, it is necessary to detect the load applied to the back of the leg. Therefore, reaction force sensors 50a, 50b, 52a, 52b (shown by broken lines in FIGS. 2 and 3) are provided on the backs of the left and right legs of the wearer 12.

  The reaction force sensor 50a detects a reaction force against the load on the front side of the right leg, and the reaction force sensor 50b detects a reaction force on the load on the rear side of the right leg. The reaction force sensor 52a detects a reaction force against the load on the left leg front side, and the reaction force sensor 52b detects a reaction force against the load on the left leg rear side. Each of the reaction force sensors 50a, 50b, 52a, and 52b includes, for example, a piezoelectric element that outputs a voltage corresponding to an applied load, and changes in load accompanying weight shift and grounding of the wearer's 12 leg and the ground. The presence or absence of each can be detected.

  Here, the configuration of the motion assisting wearing device 18 will be described with reference to FIGS. 4 and 5 together. FIG. 4 is a left side view of the motion assisting wearing tool 18. FIG. 5 is a rear view of the motion assisting wearing tool 18.

  As shown in FIGS. 4 and 5, the movement assisting wearing device 18 includes a waist belt 30 that is worn on the waist of the wearer 12, and a right leg auxiliary portion 54 that is provided below the right side of the waist belt 30. And a left leg auxiliary portion 55 provided below the left side portion of the waist belt 30.

  The right leg auxiliary part 54 and the left leg auxiliary part 55 are arranged symmetrically, and extend downward from the first frame 56 so as to support the waist belt 30 and the first frame 56. A second frame 58 formed along the outer thigh of the wearer 12, a third frame 60 extending downward from the second frame 58 and formed along the outer shin of the wearer 12, and the wearer And a fourth frame 62 on which the soles of the twelve legs (shoe soles when wearing shoes) are placed.

  A first joint 64 having a bearing structure is interposed between the lower end of the first frame 56 and the upper end of the second frame 58, and the first frame 56 and the second frame 58 are rotatably connected. is doing. The first joint 64 is provided at a height position coinciding with the hip joint, the first frame 56 is coupled to the support side of the first joint 64, and the second frame 58 is disposed on the rotation side of the first joint 64. Are combined.

  Further, a second joint 66 having a bearing structure is interposed between the lower end of the second frame 58 and the upper end of the third frame 60, and the second frame 58 and the third frame 60 can be rotated. It is linked to. The second joint 66 is provided at a height position that coincides with the knee joint, the second frame 58 is coupled to the support side of the second joint 66, and the third frame 60 is the rotation side of the second joint 66. Is bound to.

  Therefore, the second frame 58 and the third frame 60 are attached so that a pendulum motion can be performed with respect to the first frame 56 fixed to the waist belt 30 with the first joint 64 and the second joint 66 as pivot points. ing. That is, the second frame 58 and the third frame 60 are configured to perform the same operation as the leg of the wearer 12.

  A motor bracket 68 is provided on the support side of the first joint 64 and the second joint 66. The motor bracket 68 has a motor support portion 68a protruding in the outer horizontal direction, and the drive motors 20, 22, 24, and 26 are attached to the motor support portion 68a in a vertical state. Therefore, the drive motors 20, 22, 24, and 26 are provided so as not to protrude greatly to the side and to make it difficult to contact surrounding obstacles or the like during a walking motion.

  In addition, the first joint 64 and the second joint 66 transmit the drive torque to the second frame 58 and the third frame 60 on which the rotation shafts of the drive motors 20, 22, 24, and 26 are driven via gears. Is configured to do.

  Furthermore, the drive motors 20, 22, 24, and 26 have angle sensors (corresponding to the physical phenomenon detection means 142) 70, 72, 74, and 76 that detect joint angles. The angle sensors 70, 72, 74, and 76 include, for example, a rotary encoder that counts the number of pulses proportional to the joint angles of the first joint 64 and the second joint 66, and corresponds to the number of pulses corresponding to the joint angle. An electric signal is output as a sensor output.

  The angle sensors 70 and 72 detect a rotation angle between the first frame 56 and the second frame 58 corresponding to the joint angle (θhip) of the hip joint of the wearer 12. Further, the angle sensors 74 and 76 detect a rotation angle between the lower end of the second frame 58 and the third frame 60 corresponding to the joint angle (θknee) of the knee joint of the wearer 12.

  It should be noted that the first joint 64 and the second joint 66 are configured to be rotated only within an angular range in which the wearer's 12 hip joint and knee joint can be rotated, so that the wearer's 12 hip joint and knee joint can move unreasonably. A stopper mechanism (not shown) is built in so as not to give the noise.

  A first fastening belt 78 that is fastened to the thigh of the wearer 12 is attached to the second frame 58. A second fastening belt 80 that is fastened under the knee of the wearer 12 is attached to the third frame 60. Accordingly, the drive torque generated by the drive motors 20, 22, 24, 26 is transmitted to the second frame 58 and the third frame 60 via gears, and further via the first fastening belt 78 and the second fastening belt 80. Then, it is transmitted as an assist force to the leg of the wearer 12.

  The fourth frame 62 is rotatably connected to the lower end of the third frame 60 via a shaft 82. Furthermore, the lower end of the fourth frame 62 is provided with a heel receiving portion 84 on which the heel portion of the shoe sole of the wearer 12 is placed. The second frame 58 and the third frame 60 can be adjusted in length in the axial direction by a screw mechanism, and can be adjusted to any length according to the length of the leg of the wearer 12. ing.

  Each of the frames 56, 58, 60, 64 is made of metal, and can support the weights of the batteries 32, 34, the control back 36, and the motion assisting attachment 18 provided on the waist belt 30. That is, the motion assisting device 10 is configured such that the weight of the motion assisting wearing tool 18 or the like does not act on the wearer 12, and is attached so as not to apply an extra load to the wearer 12 whose muscle strength has decreased.

  FIG. 6 is a block diagram of each device constituting the motion assisting device 10. As shown in FIG. 6, the batteries 32 and 34 supply power to the power supply circuit 86. The power supply circuit 86 converts the power into a predetermined voltage and supplies a constant voltage to the input / output interface 88. Further, the charging capacity of the batteries 32 and 34 is monitored by the battery charging warning unit 90. When the battery charging warning unit 90 decreases to a preset remaining amount, a warning is issued to the wearer 12 for battery replacement or Notify charging.

  The first to fourth motor drivers 92 to 95 that drive the drive motors 20, 22, 24, and 26 amplify the drive voltage corresponding to the control signal from the control device 100 via the input / output interface 88 and drive each drive. Output to motors 20, 22, 24, and 26.

The surface myoelectric potential detection signals output from the myoelectric potential sensors 38a, 38b, 40a, 40b, 42a, 42b, 44a, 44b are first to eighth differential amplifiers (corresponding to the power amplification means 158) 101 to 101. The signal is amplified by 108, converted into a digital signal by an A / D converter (not shown), and input to the control device 100 via the input / output interface 88. Note that the myoelectric potential generated in the muscle is weak. Therefore, for example, a first to eighth differential amplifiers 101 to 108, to amplify the myoelectric potential of 30μV about computer can determine 3V, it is necessary to 10 ⁵ times become 100dB amplification factor of.

  Further, the angle detection signals output from the angle sensors 70, 72, 74, and 76 are input to the first to fourth angle detection units 111 to 114, respectively. The first to fourth angle detectors 111 to 114 convert the number of pulses detected by the rotary encoder into an angle data value corresponding to the angle, and the detected angle data is transmitted via the input / output interface 88 to the control device. 100 is input.

  The reaction force detection signals output from the reaction force sensors 50a, 50b, 52a, and 52b are input to the first to fourth reaction force detection units 121 to 124, respectively. The first to fourth reaction force detectors 121 to 124 convert the voltage detected by the piezoelectric element into a digital value corresponding to the force, and the detected reaction force data is transmitted to the control device via the input / output interface 88. 100 is input.

  A memory (corresponding to the data storage means 146) 130 is a storage means for storing each data, and control data for each phase set for each operation pattern (task) such as a standing motion, a walking motion, and a seating motion is stored in advance. A stored database storage area 130A, a control program storage area 130B in which a control program for controlling each motor is stored, and the like are provided.

In this embodiment, a calibration database 148 and a command signal database 150 are stored in the database storage area 130A. Further, as shown in FIG. 8, the calibration database 148 includes the muscular strength (power) e _A1 (t)... And the biometric signal E _A1 (t)... Generated by the wearer 12 wearing the motion assisting wearing device 18. 1 correspondence relationship and reference parameter K _A1 ... _Are stored. Further, the first correspondence relationship is that the muscular strength e _A1 (t)... Is proportional to the biological signal E _A1 (t).

In addition, the calibration database 148 includes the second correspondence relationship between the muscle strength (power) e ′ _A1 (t)... And the biological signal E _A1 (t). Parameter K ′ _A1 ... Is stored. The second correspondence relationship is a relationship between a change in the biological signal E _A1 (t)... And a change in the muscle strength e _A1 (t).

  The control data output from the control device 100 is output to the data output unit 132 or the communication unit 134 via the input / output interface 88, and displayed on a monitor (not shown) or a data monitoring computer, for example. It can also be transferred to data communication (not shown).

  In addition, the control device 100 identifies the phase of the movement pattern of the wearer 12 by comparing the joint angles detected by the angle sensors 70, 72, 74, and 76 with the joint angle of the reference parameter, and according to this phase Autonomous control means (corresponding to the control means 160) 100A for generating a command signal for generating the motive power.

  Furthermore, the control device 100 generates a load that applies a predetermined driving force from the drive motors (drive sources) 20, 22, 24, and 26 as an external load when the motion assisting wearing device 18 is worn by the wearer 12. The biosignal generated against the means 100D and the applied driving force is detected by myoelectric potential sensors (detection means) 38a, 38b, 40a, 40b, 42a, 42b, 44a, 44b, and based on this detection signal. A correction value setting means (corresponding to the parameter correction means 156) that generates a parameter (for example, a proportional gain in proportional control) performed by the autonomous control means 100A and sets this parameter as a correction value unique to the wearer. By appropriately controlling the operation of 100E and the correction value setting means 100E, etc., the corrected parameter is set as a parameter specific to the wearer 12 And a constant calibrating control means 100F.

  As calibration of the present embodiment, for example, initial setting calibration that is performed when the purchased motion auxiliary wearing tool 18 is first worn, and after the initial setting calibration is performed, the motion auxiliary wearing tool 18 is worn. There is reset calibration to be performed each time.

  In the initial setting calibration, the correction value setting process is performed by maintaining the state where the wearer 12 is stationary in a predetermined posture as described later.

  In the reset calibration, as described later, the wearer 12 determines a reference action that is determined in advance, for example, a stationary state calibration in which correction value update processing is performed when the wearer 12 generates muscle strength in a stationary state. There is one-motion calibration in which correction value update processing is performed by once bending and stretching the knee from a bent state while the wearer 12 is standing.

  Here, when performing the calibration, the load applied to the initial wearer 12 is set small, and the drive motors 20, 22, 24, and 26 are set so that the load gradually increases as the calibration operation progresses. A biological signal generated against the driving force is detected while being controlled. Further, in the wearing type movement assist device 10 of the present embodiment, calibration in a stationary state is performed as initial setting calibration, and calibration by one motion (one operation) is performed as reset calibration that is performed every time the wearing is performed. Either of them can be selected.

  Here, the calibration operation performed by the wearer 12 will be described with reference to FIGS.

  FIG. 7 is a diagram showing an example of each task and phase stored in each database. As shown in FIG. 7, the tasks for classifying the movement of the wearer 12 include, for example, a task A having rising movement data for shifting from the seated state to the away state, and a walking operation in which the worn wearer 12 walks. The memory 130 stores a task B having data, a task C having seating operation data for transitioning from a standing state to a seating state, and a task D having stairs climbing operation data that goes up the stairs from the standing state.

  Each task is set with a plurality of phase data that further define the minimum unit motion. For example, the task B for walking motion has a phase B1 having motion data in a state where both left and right legs are aligned, Phase B2 having motion data when the right leg is put forward, Phase B3 having motion data in a state where the left leg is put forward and aligned with the right leg, and the left leg is put in front of the right leg The phase B4 having the operation data is stored.

FIG. 8 is a diagram schematically showing the calibration database 148. As shown in FIG. 8, the calibration database 148 includes a surface myoelectric potential e _A1 (t) detected for each phase obtained by dividing each of the tasks A, B,. Reference parameters K _A1 ... Corresponding to the potential are stored.

  In the present embodiment, the wearer 12 wearing the motion assisting wearing tool 18 performs a predetermined calibration operation determined in advance. Here, for example, it is assumed that the wearer 12 as shown in FIG. 9 performs the rising operation (phases A1 to A4) from the seated state as the reference operation, and then performs the seating operation (phases A4 to A1) again.

  Here, the principle of calibration of the biological signal detection means 144 that detects the myoelectric potential corresponding to the muscle force generated by the wearer 12 will be described in more detail.

It has been found that when the wearer 12 makes a static movement, the relationship between the surface myoelectric potential and the muscle force produced by the wearer 12 is almost linear. From this, a method has been developed for estimating the torque generated by the wearer 12 from the surface myoelectric potential measured by the following formulas (1) and (2). The estimated torque is referred to as “virtual torque”.
τ _hip = K ₁ e ₁ -K ₂ e ₂ (1)
τ _knee = K ₄ e ₄ -K ₃ e ₃ (2)
In equations (1) and (2), τ _hip is the virtual torque of the _hip joint, τ _knee is the virtual torque of the knee joint, e _{1 to} e ₄ are surface myoelectric potentials generated by the muscles, and K _{1 to} K ₄ are It is a parameter. The hip joint and knee joint of the wearer 12 operate by the balance of contraction between the flexor and extensor. As shown in FIGS. 10A and 10B and FIGS. 11A and 11B, e ₁ is the surface myoelectric potential of the rectus femoris, e ₂ is the surface myoelectric potential of the gluteal muscle, and e ₃ is the medial vastus , E ₄ is the surface myoelectric potential of the biceps femoris.

  For the calculation of the virtual torque, a value that has passed through a digital filter in consideration of noise and the like is used. In the present embodiment, a value obtained by passing through a low-pass filter is acquired as a surface myoelectric potential value.

In the calibration of the control system for detecting the surface myoelectric potential, each parameter K of the following equation (3) obtained from the virtual torque τ of each muscle is obtained from the equations (1) and (2).
τ = Ke (3)
That is, in the calibration of the present embodiment, the value of the parameter K in Expression (3) is calculated so that the value of the surface myoelectric potential becomes 1 when the target muscle exerts a force of 1 Nm, and this value is updated. To do.

  As described above, in this embodiment, the myoelectric potential sensors 38a, 38b, 40a, 40b, 42a, 42b, 44a, 44b are detected in both cases of the initial setting calibration and the resetting calibration, and the detection results are obtained. Is used to correct the value of the parameter K.

  Next, the above-described reset calibration will be described. As an operation of this calibration, for example, one motion that moves the knee from a bent state to a stretched state while the wearer 12 is sitting can be applied, and the burden on the wearer 12 can be reduced, and in a short time. It becomes possible to finish the calibration.

FIG. 12 is a schematic view showing a state of the flexor of the knee joint of the wearer 12 to which the motion assisting wearing device 18 is attached. As shown in FIG. 12, input torque τ _m is applied as a load to the knee joint using the drive motors 20, 22, 24, and 26 with respect to the wearer 12 in a state in which the motion assisting wearing device 18 is worn. The wearer 12 applies a muscle force that antagonizes the input torque τ _m to keep the knee joint stationary. At this time, it can be said that the input torque τ _m applied by the drive motors 20, 22, 24, 26 is the same as the muscular torque τ _h generated by the wearer 12.

From this, the following formula (4) is established.
τ _m (t) = τ _h (t) (4)
From the above formula (3), the muscle strength generated by the wearer 12 is
τ _h = Ke (5)
Equation (4) can be expressed as
τ _m = Ke (6)
Can be rewritten.

Next, an initial setting calibration procedure performed in the stationary state will be described. The initial setting calibration in the stationary state is performed by the following procedure.
(Procedure 1) The surface myoelectric potential e when the wearer 12 is exerting a muscular force that opposes the driving force (torque τ _m ) of the driving motors 20, 22, 24, and 26 is measured.
(Step 2) by using the least square method from the torque tau _m input and when the measured surface myoelectric potential Request parameter K of Formula (6) holds.

An arithmetic expression for obtaining the parameter K using the least square method is as shown in the following expression (7).
K = Στ _m (t) e (t) / Σe ² (t) (7)
From the above, for example, as shown in FIG. 12, the surface myoelectric potential value is 1 when the wearer 12 outputs a force of 1 Nm while sitting still with the knee joint bent at an angle of approximately 90 degrees. The parameter K can be obtained as follows. In this stationary state, the driving force (torque τ _m ) of the drive motors 20, 22, 24, and 26 is applied to the wearer 12 in stages as a load (input torque), whereas the wearer 12 By generating muscular strength so as to antagonize, it will remain stationary.

Next, a reset calibration procedure for performing a predetermined reference operation will be described. The reset calibration in this one motion (one operation) is performed by the following procedure.
(Procedure 1) The wearer 12 rotates the knee joint so that the knee angle changes from 90 degrees to 180 degrees, and then returns the knee joint so that the knee angle changes from 180 degrees to 90 degrees.
(Procedure 2) A driving force (torque τ _m ) of the driving motors 20, 22, 24, 26 corresponding to the knee joint angles detected by the angle sensors 74, 76 is applied.
(Procedure 3) The surface myoelectric potential e is measured when the wearer 12 performs knee expansion and contraction.
(Procedure 4) By using the least square method from the measured surface myoelectric potential and the torque τ _m inputted at that time, a parameter K that satisfies Equation (6) is obtained.

  Here, the principle when the initial setting calibration is performed will be described with reference to FIGS.

For example, the parameter K is obtained by applying a torque of 8 Nm, 16 Nm, 24 Nm, and 32 Nm from the drive motors 20, 22, 24, and 26 to the wearer 12 as the input torque τ _m . In this case, the virtual torque is calculated from the surface myoelectric potential when calibration is performed using the obtained parameter K, and the comparison result with the input torque applied at that time is shown in FIGS. FIG. 13 is a graph showing the input torque (a) and virtual torque (b) for the extensor of the right hip joint. FIG. 14 is a graph showing the input torque (a) and virtual torque (b) for the flexor muscle of the right hip joint.

  The parameter K obtained by the above method from the input torque graph (a), the virtual torque graph (b) shown in FIG. 13, the input torque graph (b) and the virtual torque graph (b) shown in FIG. It can be seen that the virtual torque calculated by using the input torque and the input torque applied at that time substantially coincide with each other.

  In FIG. 13, the input torque from the drive motors 20, 22, 24, and 26 is controlled such that the torque value increases stepwise with time as can be seen from the graph (a) of FIG. The Accordingly, the drive motors 20, 22, 24, and 26 are initially driven to have a small torque value, and the input torque is applied in a pulse manner at a predetermined time interval, and the torque value is stepwise. It is controlled to become larger.

  This prevents the wearer 12 from being given excessive torque when the motion assisting wearing device 18 is worn, and resists the input torque by gradually increasing the value of the input torque. The burden of muscles that generate muscle strength is reduced, and muscle fatigue during calibration can be reduced.

  Further, as can be seen from FIGS. 13 and 14, similar results could be obtained with the left and right hip joints and the left and right knee joints. When the assist force is generated by the virtual torque using the parameter K obtained as described above, the drive motors 20, 22, 24, and 26 are driven similarly to the muscular force 1Nm generated by the wearer 12. Since the force 1Nm can be applied to the wearer 12 as an assist force, the wearer 12 can operate with a muscular force that is half of the force required for the predetermined operation.

  Furthermore, in the present embodiment, when the wearer 12 wearing the movement assisting wearing device 18 performs the calibration, it is necessary to generate a muscular force that opposes the input torque, so that the burden on the wearer 12 does not increase. Thus, the driving force of the drive motors 20, 22, 24, and 26 is controlled so as to perform calibration while suppressing the input torque.

  In other words, in this embodiment, the wearer 12 performs a predetermined operation (see, for example, FIG. 9 or FIG. 12) to perform calibration of the surface myoelectric potential. It is possible to calibrate the surface myoelectric potential without incurring the burden.

  For example, when a predetermined motion is performed twice, if the muscle strength generated by each joint in the two motions is the same, the virtual torque obtained at that time must also be the same. Therefore, by storing the virtual torque pattern of the reference operation as reference data in the memory 130 in advance, the parameter correction process at the time of calibration can be performed efficiently.

Then, using the parameter K obtained when the wearer 12 performs the calibration operation, the virtual torque when the wearer 12 performs the predetermined predetermined operation is set to τ _i (t), and the same operation as that is newly performed. Assuming that the surface myoelectric potential when e ′ (t) is e ′ (t), the relationship of the following equation (8) is established.
τ _i (t) = Ke ′ (t) (8)
When the surface myoelectric potential is calibrated, as shown in FIG. 15, the surface myoelectric potential when the wearer 12 performs the same operation as the reference predetermined operation (the graph (a) indicated by the solid line in FIG. 15). And the parameter K ′ is determined so that the virtual torque (the graph (b) indicated by the broken line in FIG. 15) is the same as the input torque (first parameter setting means).

The equation (9) for obtaining the parameter K ′ using the least square method is as follows, similarly to the equation (7) described above.
K ′ = Στ _m (t) e ′ (t) / Σe ′ ² (t) (9)
Since the virtual torque τ _i is obtained by calibration using the wearable motion assisting device 10, the parameter K ′ obtained from the above is the same as that of the calibration using the wearable motion assisting device 10. It can be said. Therefore, it is possible to give the wearer 12 an assist force of 1 Nm with respect to the muscle strength 1 Nm produced by the wearer 12 by the calibration method in which the wearer 12 performs a predetermined operation.

  Next, FIG. 17, FIG. 18, and FIG. 19 show experimental results when the reference operation using the calibration of the present embodiment is a bending / extending operation as shown in FIG. 16, for example.

  The graph (a) shown in FIG. 16 shows the joint angle change of the hip joint accompanying the bending / extending operation, and the graph (b) shown in FIG. 16 shows the joint angle change of the knee joint accompanying the bending / extending operation.

  In FIG. 17, graph (a) shows the virtual torque of the hip joint bending operation accompanying the flexion and extension operation, graph (b) shows the virtual torque of the hip joint extension operation accompanying the flexion and extension operation, and graph (c) The virtual torque of the bending operation of the knee joint accompanying the bending / extending operation is shown, and the graph (d) shows the virtual torque of the extending operation of the knee joint accompanying the bending / extending operation.

  Next, in the case where the reference operation is the bending / extending operation as described above, when calibration is performed by the bending / extending operation, correction results of the right hip flexor and extensor as shown in FIGS. 18 and 19 were obtained. In FIG. 18, graph (a) shows the surface myoelectric potential of the hip joint extension operation accompanying the flexion and extension operation, graph (b) shows the reference virtual torque of the hip joint extension operation accompanying the flexion and extension operation, and graph (c). Indicates the estimated torque of the hip joint extension motion accompanying the flexion and extension motion. In FIG. 19, graph (a) shows the surface myoelectric potential of the bending motion of the hip joint accompanying the bending / extending operation, graph (b) shows the reference virtual torque of the bending motion of the hip joint accompanying the bending / extending motion, c) shows the estimated torque of the bending motion of the hip joint accompanying the bending and stretching motion.

  Therefore, the virtual torque based on the graphs (a) to (c) shown in FIGS. 18 and 19 and the estimated torque based on the parameter K ′ obtained by calibration have the same amplitude waveform, and the estimated torque accompanying the bending / extending operation It can be seen that is approximately the same magnitude as the virtual torque obtained from the surface myoelectric potential.

  As described above, in this embodiment, the wearer 12 can perform the calibration of the surface myoelectric potential by performing a predetermined operation, so that it is not necessary to place a large burden on the wearer 12 and the virtual torque is reduced. It is possible to instantaneously calculate the parameter K ′ for obtaining (in other words, the calibrated surface myoelectric potential).

  The torque given to the wearer 12 as a load in the above description can be set so as to give a load according to the physical strength of each person. For example, by setting a lower limit value and an upper limit value of the load in advance. It is also possible to adjust so that the burden on the wearer 12 is not excessive during calibration.

  Here, a procedure of main control processing executed by the control device 100 will be described with reference to a flowchart shown in FIG. As shown in FIG. 20, in step S <b> 11 (hereinafter, “step” is omitted), the control device 100 operates the power switch (not shown) to be turned on by mounting the motion assisting wearing tool 18 on the wearer 12. In step S12, it is checked whether the power-on operation is the first time. In S12, if it is the first time, the process proceeds to S13 to shift to the initial setting mode, and the initial setting calibration process described above is executed in S14.

  That is, in S14, a biosignal for the driving force as a load applied from the driving motors 20, 22, 24, and 26 is output from each myoelectric potential sensor 38a, 38b, 40a, 40b, 42a, 42b, 44a, and 44b. Detection is performed using a surface myoelectric potential detection signal, and a correction value is obtained based on the detection signal. In S15, the voltage applied to the motor is increased by one rank to increase the load. Then, it progresses to S16 and it is confirmed whether load has reached the preset upper limit. In S16, when the load is not the preset upper limit value, the process returns to S14, and the processes of S14 to S16 are repeated.

  In S16, when the load reaches the preset upper limit value, the process proceeds to S17, and the parameter K ′ obtained by the calibration is set.

In the next S17, the correction value (parameter K ′) corresponding to the muscle strength of the wearer 12 obtained by the calibration of the wearer 12 wearing the motion assisting wearing tool 18 in a stationary state as shown in FIG. Is set (first parameter setting means). That is, in S15, as described above, the parameter is set so that the value of the surface myoelectric potential becomes 1 when the wearer 12 outputs a force of 1 Nm while sitting still with the knee joint bent at an angle of approximately 90 degrees. Find K. In this initial calibration, the driving force (torque τ _m ) of the drive motors 20, 22, 24, and 26 is applied to the wearer 12 in stages as a load (input torque), whereas the wearer 12 Generate muscle strength to antagonize the driving force.

  In this way, each myoelectric potential sensor detects a biological signal generated against the driving force applied from the driving source, generates a parameter for arithmetic processing based on this detection signal, and this parameter is unique to the wearer. Is set in the database 148 as a correction value.

  Thus, each time the motion assisting wearing device 18 is worn on the wearer 12, based on the correspondence (second correspondence) between the power and the biological signal generated in the process in which the wearer 12 performs a predetermined basic motion, It becomes possible to correct the auxiliary power according to the biological signal so as to satisfy the correspondence (first correspondence) between the power generated by the wearer 12 and the biological signal.

  After that, it progresses to S18 and transfers to the control mode which performs a normal assist force control process. In S19, the normal control mode is continued until the power switch is turned off.

  In S12, when the power-on operation is performed for the second time or later, the process proceeds to S20 to shift to the above-described reset mode. In S21, the wearer 12 executes the correction value setting calibration in one motion (one operation), and the wearer 12 obtained by performing the calibration operation as shown in FIG. A correction value (parameter K ′) is set according to the muscle strength (second parameter setting means). Thereafter, the processes of S17 to S19 are executed.

  In this embodiment, the one-motion calibration is performed after the second time. However, the present invention is not limited to this, and the correction value setting calibration may be performed in the still state in the second and subsequent times as in the first time. good.

  Next, control processing for each correction value setting mode will be described with reference to FIGS. FIG. 21 is a flowchart showing a control procedure of initial calibration for performing initial setting. In the case of the initial calibration, as described above, the correction value is set by generating muscle strength so that the wearer 12 is seated against the motor load.

  As shown in FIG. 21, the control device 100 is driven by supplying a predetermined drive current to the drive motors 20, 22, 24, and 26 according to the stationary state (see FIG. 12) where the wearer 12 is seated in S <b> 31. Force (input torque) is applied as a load. Therefore, the wearer 12 generates muscle strength so as to antagonize the driving force of the drive motors 20, 22, 24, and 26 while in the seated state.

In next S32, the myoelectric potential signal of the wearer 12 is acquired from each myoelectric potential sensor 38a, 38b, 40a, 40b, 42a, 42b, 44a, 44b. In the next S33, a virtual torque is estimated by calculation based on the actually measured myoelectric potential signal (corresponding to the estimation means of claim 1) .

  Thereafter, the process proceeds to S34, and the input torque applied as a load is compared with the virtual torque. In S35, the ratio between the input torque and the virtual torque is obtained. In the next S36, the parameter for the load for each phase stored in the calibration database 148 is read, and the correction value (correction parameter) of the control signal supplied to the motor drivers 92 to 95 is multiplied by the above ratio. ) Then, it progresses to S37 and sets a correction parameter as a parameter of autonomous control.

  In this manner, the wearer 12 wearing the motion assisting wearing device 18 can automatically perform the calibration of the biological signal according to the state of the day while sitting, and the calibration can be performed as in the related art. For this purpose, the troublesome work of attaching the weight to the wearer as a load or attaching a coil spring instead of the weight is not required. Therefore, it is possible to greatly reduce labor and time required for calibration, and thereby, it is possible to further promote the practical use and spread of the wearable movement assist device 10.

  In addition, a correction value corresponding to the state of the wearer 12 is set without imposing an extra burden on the wearer 12 whose muscle strength has been weakened, and the myoelectric potential signal of the wearer 12 is set. It is possible to accurately apply the driving force based on the operation of the wearer 12.

  Therefore, when performing calibration, an assist force according to the intention of the wearer 12 is applied from the drive source, and the assist force does not become excessively large or small, and the operation of the wearer 12 is stably assisted. Thus, the reliability of the wearable movement assist device can be further increased.

  In particular, the wearer 12 can calibrate with peace of mind even in a situation where it is difficult to use the worn motion assisting wearing device 18 as expected, such as when the wearer 12 is a beginner. Therefore, even in the case of a physically handicapped person in which the wearer 12 cannot freely move, the calibration can be performed so as to avoid a disadvantageous physical movement of the wearer 12 without performing a special operation. It is possible to perform calibration so as to compensate for the physical weakness of the wearer 12.

  Next, the calibration in the reset mode 1 described above will be described with reference to FIG. FIG. 22 is a flowchart showing a reset calibration control procedure by one motion (one operation). When performing calibration by one motion, the wearer 12 moves the knee from the bent state to the extended state only once while sitting. The memory 130 stores in advance a reference myoelectric potential corresponding to the calibration operation.

  As shown in FIG. 22, the control device 100 confirms the presence or absence of detection signals from the knee joint angle sensors 74 and 76 in S41. When the wearer 12 detects the movement of the joint angle of the second joint 66 when the wearer 12 performs the knee expansion / contraction operation in the seated state as shown in FIG. , 76 is used to set the knee operating angle.

  Subsequently, the process proceeds to S43, and the reference myoelectric potential corresponding to the knee operating angle is read from the memory 130. Next, in S44, the measured value of the myoelectric potential of the wearer 12 is read from each myoelectric potential sensor 38a, 38b, 40a, 40b, 42a, 42b, 44a, 44b. In S45, the reference myoelectric potential is compared with the actually measured value of the myoelectric potential.

  Next, in S46, the ratio between the reference myoelectric potential and the measured value of the myoelectric potential is obtained. In S47, a parameter corresponding to the knee motion angle stored in the calibration database 148 described above is read out, and the correction value (correction) of the control signal supplied to the motor drivers 92 to 95 is multiplied by this ratio. Parameter). Subsequently, the process proceeds to S48, and the correction parameter is set as a parameter for autonomous control.

  Thus, in the second and subsequent calibrations, the parameter K ′ is corrected by an operation (one motion) of rotating the knee while sitting without using the driving force of the driving motors 20, 22, 24, and 26. Therefore, the physical burden on the wearer 12 can be greatly reduced, and the preparation time required for calibration after the operation assisting wearing device 18 is worn can be shortened. Therefore, in the second and subsequent calibrations, walking can be started quickly.

  Next, the calibration in the reset mode 2 described above will be described with reference to FIG. In the reset mode 2, the wearer 12 performs the rising operation from the seated state (phases A1 to A4) as the reference operation, and then performs the seating operation (phases A4 to A1) again (see FIG. 9). .

  As shown in FIG. 23, the control device 100 confirms the presence / absence of detection signals from the angle sensors 70, 72, 74, 76 provided in the motion assisting wearing tool 18 in S 51. Then, when the wearer 12 performs the operation as shown in FIG. 9, the movement of the joint angles of the first joint 64 and the second joint 66 is detected by the detection signals from the angle sensors 70, 72, 74, 76. In S52, the task stored in the calibration database 148 is selected based on the detection signals from the angle sensors 70, 72, 74, 76, and the reference operation of the wearer 12 is set.

  In the next S 53, the reference myoelectric potential corresponding to the reference motion of the first joint 64 and the second joint 66 is read from the memory 130. Then, it progresses to S54 and the measured value of the myoelectric potential of the wearer 12 is read from each myoelectric potential sensor 38a, 38b, 40a, 40b, 42a, 42b, 44a, 44b. In S55, the reference myoelectric potential is compared with the actually measured value of the myoelectric potential.

  Next, in S56, a ratio between the reference myoelectric potential and the measured value of the myoelectric potential is obtained. In S57, a parameter corresponding to the knee movement angle stored in the calibration database 148 is read, and the correction value (correction) of the control signal supplied to the motor drivers 92 to 95 is multiplied by the above ratio. Parameter). Subsequently, the process proceeds to S58, and the correction parameter is set as a parameter for autonomous control.

  In the next S59, it is confirmed whether or not the calibration operation task is completed. In S59, when the phase of the calibration operation still remains, the process proceeds to S60 to update to the next phase and execute the processing after S53 again.

  If the calibration operation task is completed in S59, the current calibration process is terminated.

  In this way, the second and subsequent calibrations can correct the parameter K ′ without using the driving force of the driving motors 20, 22, 24, and 26, so the physical burden on the wearer 12 is greatly increased. In addition to being able to reduce, it is possible to shorten the preparation time required for calibration after mounting the motion assisting mounting tool 18.

  Accordingly, the calibration operation may be performed by the wearer 12 performing a bending / extending operation, or the surface myoelectric potential may be calibrated, or the knee bending / extending operation may be performed with the wearer 12 sitting on a chair. Since it is good also as an operation | movement and the calibration suitable for the individual can be performed, when the wearer 12 is a physically handicapped person, it is also possible to perform a calibration by arbitrary operation | movement which can be operated, An operation (task) can be set as a reference operation.

  In the above-described embodiment, the motion assisting device 10 configured to apply the assist force to the leg of the wearer 12 is described as an example. However, the present invention is not limited to this. For example, the motion assisting device 10 is configured to assist the motion of the arm. Of course, the present invention can also be applied to other motion assisting devices.

  In the above-described embodiment, the configuration in which the drive torque of the electric motor is transmitted as the assist force has been described. However, the present invention can be applied to a device that generates the assist force using a drive source other than the electric motor. is there.

1 is a block diagram showing a control system of a wearable movement assist device to which an embodiment of a calibration device according to the present invention is applied. It is the perspective view which looked at the state by which one Example of the mounting | wearing type movement assistance apparatus was mounted | worn from the front side. It is the perspective view which looked at the state by which one Example of the mounting | wearing type movement assistance apparatus was mounted | worn from the rear side. FIG. 6 is a left side view of the motion assisting wearing tool 18. FIG. 6 is a rear view of the motion assisting wearing tool 18. It is a block of each apparatus which comprises the movement assistance apparatus 10. FIG. It is a figure which shows an example of each task and phase. It is the figure which showed the calibration database 148 typically. It is a figure which shows the operation | movement process of phase A1-A4 as an example of operation | movement. It is a diagram showing the detection position of surface myoelectric potential e ₁ ~e _4, (A) figure as viewed from the front of the legs (B) is a view as seen from later legs. Is a diagram showing the detection position of surface myoelectric potential e ₁ ~e _4, (A) is a side view of the leg showing the surface myoelectric potential when bending the hip joint in the arrow direction, (B) an arrow direction knee joint It is a side view of the leg which shows the surface myoelectric potential at the time of bending in the direction. It is the schematic which shows the state of the flexor of the knee joint of the wearer 12 with which the movement assistance mounting tool 18 was mounted | worn. It is a graph which shows the input torque and virtual torque with respect to the extensor muscle of a right hip joint. It is a graph which shows the input torque and virtual torque with respect to the flexor muscle of a right hip joint. It is a graph which shows the difference of surface myoelectric potential and virtual torque when the wearer 12 performs the same operation | movement as a reference | standard predetermined operation | movement. It is a graph which shows the joint angle change of the hip joint accompanying the bending-extension operation | movement, and the joint angle change of a knee joint. It is a graph which shows the virtual torque of the bending operation of the hip joint accompanying the bending and extending operation, the virtual torque of the extending operation of the hip joint, the virtual torque of the bending operation of the knee joint, and the virtual torque of the extending operation of the knee joint. It is a graph which shows the surface myoelectric potential of the bending operation of the hip joint accompanying bending and extending operation, the reference virtual torque of the bending operation of the hip joint, and the estimated torque of the extending operation of the hip joint. It is a graph which shows the surface myoelectric potential of the hip joint extension operation | movement accompanying bending / extension operation | movement, the reference | standard virtual torque of the hip joint extension operation | movement, and the estimation torque of the bending operation of a hip joint. 4 is a flowchart for explaining a procedure of main control processing executed by the control device 100. It is a flowchart which shows the control procedure of the first calibration which performs the initial setting in a still state. It is a flowchart which shows the control procedure of the reset calibration by one motion (one operation | movement). It is a flowchart which shows the control procedure of the calibration by predetermined reference | standard operation | movement.

Explanation of symbols

DESCRIPTION OF SYMBOLS 10 Movement assistance apparatus 12 Wearer 20 Right thigh drive motor 22 Left thigh drive motor 24 Right knee drive motor 26 Left knee drive motor 30 Lumbar belt 32, 34 Battery 36 Control back 38a, 38b, 40a, 40b, 42a, 42b, 44a 44b EMG sensor 50a, 50b, 52a, 52b Reaction force sensor 54 Right leg auxiliary part 55 Left leg auxiliary part 56 1st frame 58 2nd frame 60 3rd frame 62 4th frame 64 1st joint 66 2nd joint 70 , 72, 74, 76 Angle sensor 78 First fastening belt 80 Second fastening belt 84 Wedge receiving portion 86 Power supply circuit 88 Input / output interface 100 Control devices 101 to 108 Differential amplifiers 111 to 114 Angle detection portions 121 to 124 Reaction force detection Unit 130 memory 140 drive source 142 physical phenomenon detection means 144 biological signal detection means 1 6 data storage means 148 the calibration database 150 the command signal database 152 phase identification unit 154 difference deriving means 156 parameter correcting means 158 power amplifying means 160 control means 162 calibration control unit 164 load generating means

Claims (5)

Detecting means for detecting a potential as a biological signal generated in accordance with the operation of the wearer;
An operation assisting wearing device having a drive source for providing auxiliary power to the wearer;
A calibration device for a wearable motion assisting device having control means for controlling the drive source by performing arithmetic processing so as to generate auxiliary power according to a detection signal detected by the detecting means,
The detection means detects a potential as a biological signal when generating a muscle force against a driving force as a load applied stepwise from the driving source, and the wearer generates in response to the detection signal. Estimating means for estimating a virtual torque to be
Calibration means for updating a parameter for correcting the auxiliary power by the arithmetic processing based on the relationship between the virtual torque estimated by the estimation means and the detection signal ;
A calibration device for a wearable movement assist device, comprising: The calibration means includes
When the wearer wearing the motion assisting wearing device is in a stationary state, the detection signal detected by the detection unit and the virtual torque due to the muscle force according to the muscle strength against the input torque as the load input from the drive source First storage means for storing the correspondence as initial setting data;
Comparing the input torque as the load input from the drive source and the estimated torque due to the muscular force corresponding to the detection signal detected by the detection means in the process in which the wearer performs a predetermined operation to counter the load A second storage means for storing the correspondence relationship between the differences as reset data;
The virtual torque corresponding to the detection signal and the input torque coincide with each other based on the correspondence relationship between the detection signal stored in the first storage means and the virtual torque each time the movement assisting wearing device is mounted on the wearer. First parameter setting means for setting parameters as follows:
Second parameter setting means for setting a parameter such that the estimated torque corresponding to the detection signal and the input torque match based on the difference between the input torque stored in the second storage means and the estimated torque;
The calibration apparatus for a wearable movement assist apparatus according to claim 1, comprising: A phase specifying means for specifying a phase constituting a partial operation included in the predetermined operation performed by the wearer;
The second parameter setting means sets the parameter so that the estimated torque corresponding to the detection signal matches the input torque based on the correspondence relationship between the estimated torque and the input torque in the phase specified by the phase specifying means. The calibration apparatus for a wearable movement assist apparatus according to claim 2, wherein the apparatus is a calibration apparatus. The correspondence stored in the first storage means is such that the auxiliary power has a positive correlation with the detection signal,
The wearable operation assistance according to claim 2, wherein the correspondence stored in the second storage means is a relationship between a change in the detection signal and a change in the auxiliary power in the predetermined operation. Device calibration device. An operation assistance wearing device having a drive source that generates auxiliary power corresponding to a detection signal obtained from a detection unit that detects a potential as a biological signal generated in accordance with the operation of the wearer is attached to the wearer. A calibration program for a wearable motion assisting device that causes a computer to execute calibration for associating the biological signal and the auxiliary power with a required relationship each time,
In the computer,
When the wearer wearing the motion assisting wearing device is in a stationary state, the detection signal detected by the detection unit and the virtual torque due to the muscle force according to the muscle strength against the input torque as the load input from the drive source A first procedure for storing the correspondence as initial setting data;
The detection means detects a potential as a biological signal when generating a muscle force against a driving force as a load applied stepwise from the driving source, and the wearer generates in response to the detection signal. A second procedure for estimating the virtual torque to be
An input torque as the load inputted from the drive source, before comparing the Ki推 constant torque, and a third step of storing the correspondence between both by the difference as resetting data,
Each time the motion assisting wearing device is worn by the wearer, the virtual torque corresponding to the detection signal matches the input torque based on the correspondence between the detection signal stored in the first procedure and the virtual torque. A fourth procedure for setting parameters in
A fifth procedure for setting parameters so that the estimated torque corresponding to the detection signal and the input torque match based on the difference between the input torque stored in the third procedure and the estimated torque;
A program for calibration of a wearable motion assisting device, characterized in that

Images