WO2015177634A1

WO2015177634A1 - Rehabilitation apparatus, control method, and control program

Info

Publication number: WO2015177634A1
Application number: PCT/IB2015/000990
Authority: WO
Inventors: Hitoshi Yamada; Masashi Yamashita
Original assignee: Toyota Jidosha Kabushiki Kaisha; Toyota School Foundation
Priority date: 2014-05-22
Filing date: 2015-05-20
Publication date: 2015-11-26
Also published as: JP2015221073A

Abstract

A rehabilitation apparatus includes: an operation unit that is operated by a subject of rehabilitation; an operation amount detection unit that detects an operation amount of the operation unit; a drive unit that drives the operation unit; a control unit that controls drive of the drive unit; a posture detection unit that detects a posture of the subject; an external force detection unit that detects an external force acting upon the operation unit; and a display unit that displays an operation target position of the operation unit. The control unit calculates a manipulability ellipse representing manipulability of a body of the subject based on the posture of the subject detected by the posture detection unit, calculates the operation target position based on the calculated manipulability ellipse, calculates a target value of the operation amount for the operation unit on the basis of the external force detected by the external force detection unit and the calculated operation target position, and controls the driving unit such that the operation amount detected by the operation amount detection unit follows the calculated target value of the operation amount.

Description

REHABILITATION APPARATUS, CONTROL METHOD, AND CONTROL PROGRAM

BACKGROUND OF THE INVENTION 1. Field of the Invention

[0001] The invention relates to a rehabilitation apparatus, a control method, and a control program for performing rehabilitation that allows a subject to restore a motor function. 2. Description of Related Art

[0002] People with impaired motor function try to restore the motor function by rehabilitation, and a variety of apparatuses have been developed for performing such a rehabilitation. For example, an apparatus for rehabilitation is available in which a load is applied to the upper extremity of a subject by driving means through a liquid clutch (see Japanese Patent Application Publication No. 2006-247280 (JP 2006-247280 A)).

[0003] However, in the body movement of a subject, depending on the posture, typically there is a direction in which the operations are difficult. A problem associated with the abovementioned rehabilitation apparatus is that since the posture of the subject is not taken into account, an excessive load is applied to the subject in certain postures.

SUMMARY OF THE INVENTION

[0004] The invention provides a rehabilitation apparatus, a control method, and a control program that enable optimum rehabilitation training corresponding to the posture of a subject.

[0005] A first aspect of the invention relates to a rehabilitation apparatus. The rehabilitation apparatus is equipped with an operation unit that is operated by a subject of rehabilitation; an operation amount detection unit that detects an operation amount of the operation unit; a drive unit that drives the operation unit; a control unit that controls drive of the drive unit; a posture detection unit that detects a posture of the subject; an external force detection unit that detects an external force acting upon the operation unit, and a display unit that displays an operation target position of the operation unit. The control unit calculates a manipulability ellipse representing manipulability of a body of the subject on the basis of the posture of the subject detected by the posture detection unit, calculates the operation target position on the basis of the calculated manipulability ellipse, calculates a target value of the operation amount for the operation unit on the basis of the external force detected by the external force detection unit and the calculated operation target position, and controls the driving unit such that the operation amount detected by the operation amount detection unit follows the calculated target value of the operation amount. In this aspect, the manipulability ellipse includes a manipulability ellipsoid. In this aspect, the control unit may calculate the operation target position in a two-dimensional coordinate system on a display screen of the display unit on the basis of a sine wave function of an amplitude including a ratio of a long diameter and a short diameter of the manipulability ellipse and an inclination of the manipulability ellipse. In this aspect, the control unit may calculate an operation target position p _m in a two-dimensional coordinate system in the display unit by using the expressions presented hereinbelow. In this aspect, the control unit may generate a virtual guideline connecting the operation target position with a point of origin of the two-dimensional coordinate system in the display unit and may perform admittance control on the drive unit such that a load is low in a direction of the virtual guideline and high in a direction perpendicular to the virtual guideline direction. In this aspect, the control unit may calculate the target value of the operation amount for the operation unit by using an admittance matrix in which at least one of a stiffness constant and a damping constant is set low in the direction of the virtual guideline and at least one of the stiffness constant and the damping constant is set high in the perpendicular direction. In this aspect, the control unit may transform the external force detected by the external force detection unit into an external force in a global coordinate system, perform rotation transformation of the transformed external force on the basis of a rotation angle of the virtual guideline, transform the rotation-transformed value by using the admittance matrix, perform inverse rotation transformation of the transformed value on the basis of the rotation angle of the virtual guideline, and transform the value obtained by inverse rotation transformation by using inverse kinematics, to calculate the target value of the operation amount for the operation unit. In this aspect, the control unit may calculate a target value 0_ref of the operation amount for the operation unit by using the expression presented hereinbelow. In this aspect, the control unit may gradually increase the set damping constant and/or the set stiffness constant according to a displacement of a present position on the display unit, which corresponds to the operation amount detected by the operation amount detection unit, from the virtual guideline in the perpendicular direction. In this aspect, the control unit may increase the damping constant and/or the stiffness constant by a predetermined amount when it is determined that a present position on the display unit, which corresponds to the operation amount detected by the operation amount detection unit, is outside of the manipulability ellipse. In this aspect, the posture detection unit may detect joint angles of the body of the subject, and the control unit may calculate the operational ellipse on the basis of the joint angles detected by the posture detection unit and a Jacobian matrix of the body. A second aspect of the invention relates to a control method for a rehabilitation apparatus. The control method for a rehabilitation apparatus includes: detecting an operation amount of an operation unit that is operated by a subject of rehabilitation; detecting a posture of the subject; detecting an external force acting upon the operation unit; displaying an operation target position of the operation unit; calculating a manipulability ellipse representing manipulability of a body of the subject on the basis of the detected posture of the subject; calculating the operation target position on the basis of the calculated manipulability ellipse; calculating a target value of an operation amount for the operation unit on the basis of the detected external force and the calculated operation target position; and controlling a driving unit for driving the operation unit, such that the detected operation amount follows the calculated target value of the operation amount. A third aspect of the invention relates to a control program for a rehabilitation apparatus. The control program for a rehabilitation apparatus includes: a process of calculating a manipulability ellipse representing manipulability of a body of a subject of rehabilitation on the basis of a posture of the subject; a process of calculating an operation target position on the basis of the calculated manipulability ellipse; a process of displaying the operation target position of an operation unit operated by the subject; a process of calculating a target value of an operation amount for the operation unit on the basis of an external force acting upon the operation unit and the calculated operation target position; and a process of controlling a driving unit for driving the operation unit, such that an operation amount of the operation unit follows the calculated target value of the operation amount.

[0006] According to the aspects of the invention, it is possible to provide a rehabilitation apparatus, a control method, and a control program that enable optimum rehabilitation training corresponding to the posture of a subject.

BRIEF DESCRIPTION OF THE DRAWINGS

[0007] Features, advantages, and technical and industrial significance of exemplary embodiments of the invention will be described below with reference to the accompanying drawings, in which like numerals denote like elements, and wherein:

FIG 1 is a block diagram illustrating the schematic system configuration of the rehabilitation apparatus according to an embodiment of the invention;

FIG 2 illustrates the operation of the grip lever unit;

FIG. 3 is a block diagram illustrating the configuration of the assist control system of the control device according to the embodiment of the invention;

FIG 4 illustrates the effect of admittance control according to the embodiment of the invention;

FIG 5 illustrates the shape of the manipulability ellipse obtained when the upper extremity of the subject is rotationally operated in the lateral direction in a horizontal plane;

FIG 6 depicts a virtual guideline in the case of an arm posture (a) at the left end in

FIG 5;

FIG 7 depicts a virtual guideline in the case of an arm posture (b) at the right end in FIG 5; and

FIG 8 is a flowchart illustrating the control processing flow of the rehabilitation apparatus according to the embodiment of the invention.

DETAILED DESCRIPTION OF EMBODIMENTS

[0008] The embodiments of the invention are described below with reference to the appended drawings. FIG 1 is a block diagram illustrating the schematic system configuration of a rehabilitation apparatus according to an embodiment of the invention. The rehabilitation apparatus 1 of the present embodiment includes a grip lever 2 operated by a subject, a camera 3 that captures the image of the posture of the upper extremity of the subject, first and second servo motors 4, 5 providing an operation torque to the grip lever 2, first and second rotation sensors 6, 7 that detect an operation amount of the grip lever 2, a force sensor 8 that detects an external force acting upon the grip lever 2, a control device 9 that controls the first and second servo motors 4, 5, and a display device 10 that displays various types of operation information.

[0009] The grip lever 2 is a specific example of operation means and performs operations allowing the subject to perform the rehabilitation of the upper extremity (FIG 2). The grip lever 2 is constituted by an articulated robot arm having a plurality of joints. For example, the grip lever 2 has a housing 21, a first joint shaft 22 provided rotatably at the housing 21, a first link 23 linked to the first joint shaft 22, a second joint shaft 24 linked to the first link 23, a second link 25 linked to the second joint shaft 24, and a handle 26 that is linked to a distal end of the second link 25 and gripped by the subject.

[0010] The first link 23 rotates about the first joint shaft 22, and the second link 25 rotates about the second joint shaft 24. With such a configuration, the handle 26 can move to any position on a two-dimensional plane. This configuration of the grip lever 2 is merely exemplary and not limiting. For example, the grip lever 2 may be configured to have three or more joints so that the handle 26 could be moved to any position on a two-dimensional plane or in a three-dimensional space. The subject holds the handle 26 and performs the rehabilitation training by moving the handle 26 in the indicated direction.

[0011] The first and second rotation sensors 6, 7 are specific examples of operation amount detection means and detect the operation amount of the grip lever 2. The first rotation sensor 6 detects the rotation angle of the first joint shaft 22 of the grip lever 2. The second rotation sensor 7 detects the rotation angle of the second joint shaft 24 of the grip lever 2. The first and second rotation sensors 6, 7 are each constituted, for example, by a potentiometer, a rotary encoder, or the like.

[0012] The first and second rotation sensors 6, 7 are connected through an analog/digital (AID) converter 11 to the control device 9. The first and second rotation sensors 6, 7 output rotation angle signals corresponding to the detected rotation angles of the first and second joint shafts 22, 24 of the grip lever 2 to the control device 9.

[0013] The first and second servo motors 4, 5 are specific examples of driving means and have a function of providing an operation torque to the handle 26 of the grip lever 2. The drive shaft of the first servo motor 4 is linked to the first joint shaft 22 of the grip lever 2. The drive shaft of the second servo motor 5 is linked to the second joint shaft 24 of the grip lever 2. The first and second servo motors 4, 5 are each, for example, an alternating current (AC) servo motor and incorporate a reduction mechanism.

[0014] The first and second servo motors 4, 5 are connected through a servo amplifier 12 and a digital/analog (D/A) converter 13 to the control device 9. The first and second servo motors 4, 5 provide a torque to the handle 26 of the grip lever 2 through the first and second joint shafts 22, 24 in response to a control signal transmitted from the control device 9.

[0015] The force sensor 8 is a specific example of external force detection means and detects an external force acting upon the handle 26 when the grip lever 2 is operated. The force sensor 8 is provided, for example, at the base of the handle 26 of the grip lever 2. The force sensor 8 is connected through an AID converter 11 to the control device 9. The force sensor 8 outputs a force value signal corresponding to the detected external force to the control device 9.

[0016] The camera 3 is a specific example of posture detection means and captures the image of the posture of the upper extremity of the subject when the subject operates the grip lever 2. For example, a plurality of markers M is attached to a hand (wrist joint, elbow joint, shoulder joint, etc.) of the subject. The camera 3 captures the image of each image marker M of the hand of the subject. The camera 3 outputs the captured image to the control device 9.

[0017] The control device 9 is a specific example of control means and controls the first and second servo motors 4, 5. The control device 9 calculates a torque command value (the target value of the operation amount) for the first and second servo motors 4, 5, on the basis of the force value signal outputted from the force sensor 8 and the rotation angle signals outputted from the first and second rotation sensors 6, 7. The control device 9 generates a control signal corresponding to the calculated torque command value and outputs the generated control signal to the first and second servo motors 4, 5. The first and second servo motors 4, 5 provide a torque to the grip lever 2 in response to the control signals from the control device 9.

[0018] The control device 9 is configured, for example, of hardware around a microcomputer constituted by a central processing unit (CPU) 9a performing computational processing, control processing, and the like, a memory 9b such as a read only memory (ROM) or random access memory (RAM) that stores a computational program and control program to be executed by the CPU 9a, and an interface unit (I/F) 9c exchanging signals with the outside. The CPU 9a, the memory 9b, and the interface 9c are connected to each other by a data bus 9d.

[0019] The display device 10 is a specific example of display means and displays various types of operation information relating to the operations performed by the subject. The display device 10 is connected to the control device 9 and displays various type of operation information on the basis of information outputted from the control device 9. The display device 10 is configured, for example, of a liquid crystal display device or an organic EL display device.

[0020] For example, the display device 10 displays a target mark □ corresponding to the present position of the handle of the grip lever 2, which is outputted from the control device 9, and a target mark O which is the operation target for the subject. The target marks are displayed simultaneously on a display screen in a two-dimensional coordinate system. The target mark O serves as a target position of handle operation for performing the rehabilitation training of the upper extremity of the subject. The subject operates the handle 26 of the grip lever 2 in a manner such that the target mark O displayed on the display screen of the display device 10 is caused to follow the target mark □ , which is the tracking task therefor. As a result, rehabilitation training is performed such that enables the restoration of the desired joint movement. The above-described rehabilitation training method is merely an example and is not limiting.

[0021] However, in the body movement of a subject, depending on the posture, typically there is a direction in which the operations are difficult. A problem associated with the conventional rehabilitation apparatus is that since the posture of the subject is not taken into account, an excessive load is applied to the subject in certain postures.

[0022] By contrast, in the rehabilitation apparatus 1 according to the present embodiment, the control device 9 calculates a manipulability ellipse representing manipulability of the body of the subject on the basis of the image of the posture of the subject which is captured by the camera 3, calculates the operation target position of the handle 26 of the grip lever 2 on the basis of the calculated manipulability ellipse, and displays the calculated operation target position on the display device 10. The control device 9 then performs admittance control on the basis of the external force detected by the force sensor 8 and the calculated operation target position, calculates control target values of the first and second joint shafts 22, 24 of the grip lever 2, and controls the first and second servo motors 4, 5 such that the rotation angles of the first and second joint shafts 22, 24, which have been detected by the first and second rotation sensors 6, 7, follow the calculated control target values.

[0023] As a result, the operation target position is set with consideration for the manipulability ellipse which is based on the posture of the subject. Therefore, no excessive load is applied to the subject. Furthermore, guidance admittance control is performed such that the subject can easily operate the handle 26 of the grip lever 2 in the direction of the operation target position. As a result, the load applied to the subject can be suitably reduced. Thus, optimum rehabilitation training corresponding to the posture of the upper extremity of the subject can be performed. [0024] In order to realize the abovementioned control, the control device 9 performs assist control for assisting the operation of the handle 26 of the grip lever 2, which is performed by the subject, on the basis of the rotation angles of the first and second joint shafts 22, 24 of the grip lever 2 which has been detected by the first and second rotation sensors 6, 7, the external force acting upon the handle 26 of the grip lever 2 which has been detected by the force sensor 8, and the image of the posture of the upper extremity of the subject which has been captured by the camera 3. The amplitude of the operation target value (target mark) moving along a virtual guideline on the display screen of the display device 10 is restricted by the size of the manipulability ellipse. The control device 9 performs the below-described higher-order control system and lower-order control system after the assist control.

[0025] In the higher-order control system, the control device 9 calculates the manipulability ellipse on the basis of the image of the upper extremity posture of the subject which has been captured by the camera 3. Further, in the higher-order control system, the control device 9 performs admittance control with two degrees of freedom and calculates the virtual guideline for guiding the handle 26 to the operation target position on the basis of the external force acting upon the handle 26 of the grip lever 2 which has been detected by the force sensor 8.

[0026] In the lower-order control system, the control device 9 performs position control that causes the rotation angles of the first and second joint shafts 22, 24 of the grip lever 2 to follow the control target values of the first and second joint shafts 22, 24 which are along the virtual guidelines calculated in the higher-order control system. In the position control, the control device 9 performs feedback control based on proportional-integral-derivative (PID) control that causes the feedback of the rotation angles of the first and second joint shafts 22, 24 of the grip lever 2 and feedforward control based on inertia compensation and friction compensation, and calculates torque command values for the first and second servo motors 4, 5.

[0027] The higher-order control system mentioned hereinabove is described below in the greater detail. FIG 3 is a block diagram illustrating the schematic system configuration of the control device according to the present embodiment. The control device 9 according to the present embodiment, has a circular Hough transform unit 91, a joint angle transform unit 92, a manipulability ellipse calculation unit 93, a target value calculation unit 94, a rotation transform unit 95, and admittance control unit 96, and an inverse kinematics unit 97.

[0028] The circular Hough transform unit 91 performs circular Hough transform of the captured image of each image mark which is outputted from the camera 3 and calculates the center position p_Cj of each image mark. p_Cj = (imgr_ay, Rcir), wherein im^y is a gray scale image, and R_Cj_r is the radius of the image mark. p_cj = [uj, Vj, 1]^T is the center position of the image mark in the image coordinate system.

[0029] The joint angle transform unit 92 transforms the center position p_cj of each image mark into a center position p_j of a global coordinate system (XY coordinate system) by a conventional transformation matrix M_pers.

[Formula 1]

[0030] In the expression above, p_j = [p_jx, p_jy] is the center position of each image mark in the global coordinate system, in which j = 1, 2, 3 represent respectively, the shoulder joint, elbow joint, and wrist joint of the upper extremity of the subject. The shoulder joint, elbow joint, and wrist joint of the upper extremity are assumed to move in planes of the same height, and height information thereon is omitted.

[0031] Here, a shoulder joint angle qi and an elbow joint angle q₂ of the upper extremity can be represented by the following expressions based on the geometric relationship.

[Formula 2]

q₂ = tan ¹ ((p_3y - p_2y )/ (p_3x - p_2x )) - q_x

[0032] Therefore, the Jacobian matrix of the upper extremity can be represented by the following expression. In this expression, \,\ stands for the length of the upper arm of the upper extremity, and L₂ stands for the length of the forearm of the upper extremity. Formula 3]

[0033] The joint angle transform unit 92 calculates the Jacobian matrix of the upper extremity by using the expressions presented hereinabove. The manipulability ellipse calculation unit 93 performs singular value decomposition of the Jacobian matrix by using the following expression and calculates the size ai_ong of a long axis, the size a_Sh0rt of a short axis, and orientation (first row vector of U_h) of the manipulability ellipse representing the manipulability of the subject's fingers.

[Formula 4]

[0034] The manipulability ellipse calculation unit 93 calculates a radius ratio λ_β1ΐί_Ρ and an inclination 6_eiii_p of the manipulability ellipse by using the following expressions.

[Formula 5]

[0035] On the basis of the radius ratio λ_ειΐίρ and inclination 9_eiii_p of the manipulability ellipse calculated by the manipulability ellipse calculation unit 93, the target value calculation unit 94 calculates a final target value pf_in = [prx, pry]^Tof the tracking task by using the following expression: pf_m = ( _eiii_p, 6_eiii_p).

[0036] However, in the tracking task, the final target value pf_in is the operation target position (for example, the two-dimensional coordinate position of a target mark) which is to be taken visually by the subject as an operation target on the display screen of the display device 10.

[0037] The target value calculation unit 94 outputs the calculated final target value pen to the display device 10. The display device 10 displays the final target value Pfin, which has been outputted from the target value calculation unit 94, as the target mark for the operation target together with the target mark representing the present position of the handle 26 of the grip lever 2 in the two-dimensional coordinate system of the display screen.

[0038] Further, for example, the target value calculation unit 94 calculates the position of the target mark corresponding the present position of the handle 26 of the grip lever 2 by using forward kinematics on the basis of the rotation angles of the first and second joint shafts 22, 24 of the grip lever 2 which are outputted from the first and second rotation sensors 6, 7.

[0039] The final target value p_fi_n can be represented by the following expression. In the following expression, R(9_rot_k) is a rotation matrix. 0_rot_k = 0 + 7i(k - l)/8 (k = 1, ..., 8) is a rotation angle of the main axis (virtual guideline which is a line connecting pf_in with the point of origin of the two-dimensional coordinate system) of the final target value pg_n.

[Formula 6]

wherein

[0040] In the expression above, prxo is the signal of the final target value which can be represented by the following sine wave, ρ^ο = A_rsin(2nft). The amplitude A_r of this sine wave changes depending not only on the rotation angle 0_rot _k of the main axis of the final target value pf_m, but also on the radius of the manipulability ellipse.

[0041] Thus, the amplitude A_r of this sine wave changes significantly in the long-axis direction (direction in which the upper extremity of the subject moves easily) of the manipulability ellipse. Therefore, the final target value p_fi_n changes significantly in the long-axis direction of the manipulability ellipse on the display screen of the display device 10. Meanwhile, the amplitude A_r of this sine wave changes little in the short-axis direction (direction in which the upper extremity of the subject moves with difficulty) of the manipulability ellipse. Therefore, the final target value ρ_ί¾ changes little in the short-axis direction of the manipulability ellipse on the display screen of the display device 10. Thus, the final target value ρ«_π is generated such that ensures a tracking task reasonable for the subject.

[0042] In the expression above, the frequency of the sine wave is, for example, f = 0.3 Hz, and the amplitude A_r can be represented by the following expression.

[Formula 7]

[0043] In the expression above, a long diameter R_ei and a short diameter of the operational ellipse can be calculated using the following expression. In the following expression, Re„ is the nominal radius of the manipulability ellipse which is a design parameter that can be set in advance at a design stage.

[0044] The rotation angle e_rot__k = 0 + 7t(k - l)/8 (k = 1, 8) is not limiting. For example, the rotation angle 9_rot_k = 0 + 7t(k - 1 )/n (k = 1 , ... , n) may be also used. By increasing n, it is possible to guide the handle 26 of the grip lever 2 more smoothly. The target value calculation unit 94 outputs the calculated rotation angle 0_rot_k of the main axis of the final target value pa, to the rotation transform unit 95.

[0045] The admittance control with two degrees of freedom which creates a virtual guideline and serves for guiding the operation of the grip lever 2 will be explained below in greater detail. The rotation transform unit 95 transforms the force value signal outputted from the force sensor 8 into a force value signal in the global coordinate system by rotation transformation using the following expression on the basis of the rotation angle θι of the first joint shaft 22, which is outputted from the first rotation sensor 6, and the rotation angle θ₂ of the second joint shaft 24, which is outputted from the second rotation sensor 7. In this case, the force value signal (force vector) prior to the rotation transformation is f_exto = [fex₀feyo]^T, and the force value signal (force vector) after the rotation transformation is f_ext = [fe_Xfe_y . Here, f_ext= R(e_x)f_exto9_x = θι + θ₂ - π/2.

[0046] The admittance control unit 96 calculates the control target value p_ref which is the operation target position of the handle 26 of the grip lever 2 in the work space by using the following expression on the basis of the force value signal f_ext transformed by the rotation transform unit 95.

[0047] The admittance control unit 96 performs the rotation transformation R(9rot_k) by using the rotation angle e_rot_k = 0 + 7c(k - l)/8 of the main axis of the final target value p_fin with respect to the force value signal f_ext from the rotation transform unit 95, as represented by the following expression. Then, the admittance control unit 96 performs the transformation (admittance control) with the admittance matrix Gad_m(s) after the rotation transformation R(6_rot _k) and performs the inverse rotation transformation R^T(6_rot _k) to calculate the control target value p_ref.

[Formula 9]

[0048] In the expression above, O_amn is a damping constant with respect to the main axis direction (virtual guideline direction), and K_a(lml is a stiffness constant in the main axis direction. D_adm2 is a damping constant with respect to the auxiliary axis direction (direction perpendicular to the virtual guideline direction), and K_ad_m2 is a stiffness constant in the auxiliary axis direction.

[0049] For example, the damping constant Damdi and/or the stiffness constant Kadmi in the main axis direction is set low, and the damping constant ϋ_3Π1(12 and/or the stiffness constant K_ad_m2 in the auxiliary axis direction is set high. As a result, when the handle 26 of the grip lever 2 is operated in the main axis direction, a low load is obtained, and when the handle is operated in the auxiliary axis direction, a high load is obtained. Therefore, when the subject operates the handle 26 of the grip lever 2 in the virtual guideline direction, the subject feels a low load, and where the handle 26 of the grip lever 2 is operated in the direction deviating from the virtual guideline, the subject feels a high load. Thus, the operation of the handle 26 of the grip lever 2 can be naturally guided in the direction of the operation target position (control target position p_ref) along the virtual guideline. Therefore, the subject can operate the handle 26 of the grip lever 2 easily toward the operation target position and the load can be suitably reduced.

[0050] The inverse kinematics unit 97 transforms the control target value p_ref in the work space, which is outputted from the admittance control unit 96, into the control target value 9_refin the joint space by using the following inverse kinematics function IK(): e_ref = IK(p_ref)

[0051] The abovementioned lower-order control system is explained in detail below. In the lower-order control system, the control device 9 performs position control which causes the control target value 0_ref = [θ_Γι, Θ_γ2] of the joint space which has been calculated in the higher-order control system to follow the rotation angles θι, θ₂ of the first and second joint shafts 22, 24 of the grip lever 2. The lower-order control system includes a feedback portion constituted by PID control and a feedforward portion constituted by inertia compensation and friction compensation.

[0052] The control device 9 calculates torque command values xi, τ₂ for the first and second servo motors 4, 5 by using the following expressions, such as to cause the rotation angles θ_ΐ5 θ₂ of the first and second joint shafts 22, 24 of the grip lever 2, which have been detected by the first and second rotation sensors 6, 7, to follow the control target value 6_ref outputted from the inverse kinematics unit 97 of the higher-order control system. The control device 9 generates control signals corresponding to the calculated torque command values xi, τ₂ and outputs the generated control signals to the first and second servo motors 4, 5, thereby controlling the first and second servo motors 4, 5.

[Formula 10]

= κ_{ρ η} -θ,)+κ,

+¾ø, +¾, sgn(0,)

[0053] In the expression above, i = 1, 2. Here, i = 1 represents the first joint shaft 22, and i = 2 represents the second joint shaft 24. I_m with a hat symbol on top denotes an estimated value of an inertia momentum of the handle 26 of the grip lever 2, B_m with a hat symbol on top denotes an estimated value of a viscous friction term coefficient, and D_m with a hat symbol on top denotes an estimated value of a dynamic friction coefficient. I_m, B_m, and D_m, each with a hat symbol on top, are off-line identified by a least square method for inertia compensation and friction compensation. K_p, Kj, and K_<j respectively represent a proportional gain, an integral gain, and a differential gain of PID control.

[0054] The control device 9 can perform the lower-order control including the inertia compensation portion, friction compensation portion, and feedback portion, which is realized by the PID control, on the basis of the expression above. As a result of including the inertia compensation portion and friction compensation portion in the lower-order control system, it is possible to use inexpensive servo motors. Therefore, the cost can be reduced.

[0055] The control device 9 calculates the present position p = [p_xpy]^T of the handle 26 of the grip lever 2 in the work space by using the forward kinematics represented by the expression below from the rotation angle Θ = [θιθ₂]^τ of the first and second joint shafts 22, 24 of the grip lever 2, and outputs the calculated present position to the display device 10. The display device 10 displays a target mark at a position corresponding to the present position p = [p_xp_y]^T of the handle 26 of the grip lever 2 in the two-dimensional coordinate system of the display screen, p = FK(9).

[0056] FIG 4 illustrates the effect of admittance control according to the present embodiment. However, the manipulability ellipse is not taken into account in this admittance control. As depicted in FIG. 4, the work space is divided by virtual guide lines into equal segments of 22.5 (deg) in the circumferential direction, and it can be confirmed that the handles of the grip lever 2 can be smoothly operated along the virtual guidelines.

[0057] FIG 5 illustrates the shape of the manipulability ellipse obtained when the upper extremity of the subject is rotationally operated in the lateral direction in a horizontal plane. As depicted in FIG 5, in response to the circular-arc movement of the subject's fingers, the long axis of the manipulability ellipse is disposed in the circumferential direction of the circular arc. Therefore, it is clear that the shape of the calculated manipulability ellipse is good. It is also clear that the short axis of the manipulability ellipse is arranged in the radial direction of the circular arc and that the radial direction is the direction in which the movement is difficult. In the case of such a posture of the upper extremity, the rehabilitation apparatus 1 of the present embodiment sets the tracking task of operating the grip lever 2 to be large in the circumferential direction of the circular-arc movement of the fingers and small in the radial direction.

[0058] FIG 6 depicts a virtual guideline in the case of an arm posture (a) at the left end in FIG 5. FIG 7 depicts a virtual guideline in the case of an arm posture (b) at the right end in FIG. 5. It is clear from FIGS. 6 and 7 that the control target value _fin along the virtual guideline is restricted by the size of the manipulability ellipse, and the induced operation of the grip lever 2 is long in the direction in which the arm is easy to move (long-axis direction of the manipulability ellipse) and short in the direction in which the arm is difficult to move (short-axis direction of the manipulability ellipse). Thus, it is clear, that the subject can perform reasonable rehabilitation.

[0059] A control method for the rehabilitation device of the present embodiment will be described hereinbelow in detail. FIG. 8 is a flowchart illustrating the control processing flow of the rehabilitation apparatus of the present embodiment. The control processing depicted in FIG 8 is repeatedly executed for each predetermined period of time.

[0060] The camera 3 captures the image of each image mark M of the subject's arm and outputs the captured images to the control device 9. The circular Hough transform unit 91 performs Hough transform of the captured images of the image marks M, which have been outputted from the camera 3, calculates the center position p_CJ of each image mark, and outputs the calculated center positions to the joint angle transform unit 92 (step SI 01).

[0061] The joint angle transform unit 92 transforms the center position p_cj of each image mark, which have been outputted from the circular Hough transform unit 91, into center positions p_j in the global coordinate system by using a transformation matrix, calculates the Jacobian matrix of the upper extremity, and outputs the calculated Jacobian matrix to the manipulability ellipse calculation unit 93 (step SI 02).

[0062] The manipulability ellipse calculation unit 93 performs singular value decomposition of the Jacobian matrix outputted from the joint angle transform unit 92, calculates the size ai_ong of the long axis, the size a_Sh_ort of the short axis, and the orientation (first row vector of U_h) of the manipulability ellipse, and calculates the radius ratio λ_ει_ΐίΡ and inclination 9_ellip of the manipulability ellipse (step SI 03). The manipulability ellipse calculation unit 93 outputs the calculated radius ratio _em_p and inclination 9_eiii_P of the manipulability ellipse to the target value calculation unit 94.

[0063] On the basis of the radius ratio λ_βΜ_Ρ and inclination 9_eiii_p of the manipulability ellipse outputted from the manipulability ellipse calculation unit 93, the target value calculation unit 94 calculates the final target value p_fm = [p_ra, p^,]⁷ of the tracking task and outputs the calculated final target value to the display device 10 (step SI 04).

[0064] The display device 10 displays the final target value p_fin, which has been outputted from the target value calculation unit 94, as the target mark for the operation target together with the target mark representing the present position of the handle 26 of the grip lever 2 in the two-dimensional coordinate system of the display screen (S105).

[0065] The rotation transform unit 95 transforms the force value signal f_exto outputted from the force sensor 8 into the force value signal f_ext in the global coordinate system by rotation transformation on the basis of the rotation angle 0_! of the first joint shaft 22, which is outputted from the first rotation sensor 6, and the rotation angle θ₂ of the second joint shaft 24, which is outputted from the second rotation sensor 7, and outputs the obtained force value signal to the admittance control unit 96 (step SI 06).

[0066] The admittance control unit 96 calculates the control target value p_ref, which is the position of the handle 26 of the grip lever 2 in the work space, on the basis of the force value signal f_ext in the global coordinate system which has been outputted from the rotation transform unit 95, and outputs the calculated control target value to the inverse kinematics unit 97 (step SI 07).

[0067] The inverse kinematics unit 97 transforms the control target value p_ref in the work space, which is outputted from the admittance control unit 96, into the control target value 0_ref in the joint space by using the inverse kinematics function IK() (step SI 08).

[0068] The control device 9 calculates the torque command values τ_ΐ5 τ₂ for the first and second servo motors 4, 5 such as to cause the rotation angles θι, θ₂ of the first and second joint shafts 22, 24 of the grip lever 2, which have been detected by the first and second rotation sensors 6, 7, to follow the control target value 6_ref outputted from the inverse kinematics unit 97 of the higher-order control system (step SI 09). The control device 9 generates control signals corresponding to the calculated torque command values ti, τ₂ and outputs the generated control signals to the first and second servo motors 4, 5, thereby controlling the first and second servo motors 4, 5.

[0069] In the rehabilitation apparatus 1 of the present embodiment, the manipulability ellipse representing the operation ability of the subject's fingers is calculated, the operation target position of the handle 26 of the grip lever unit 2 is calculated on the basis of the calculated manipulability ellipse, and the calculated operation target position is displayed on the display device 10. Further, in the rehabilitation apparatus 1, the admittance control is performed on the basis of the external force detected by the force sensor 8 and the calculated operation target position, the target values of the operation amounts of the first and second joint shafts 22, 24 of the grip lever 2 are calculated, and the first and second servo motors 4, 5 are controlled such that the rotation angles of the first and second joint shafts 22, 24, which have been detected by the first and second rotation sensors 6, 7, follow the calculated target values of the operation amounts.

[0070] As a result, the operation target position is set with consideration for the manipulability ellipse which is based on the upper extremity posture of the subject. Therefore, no excessive load is applied to the subject. Furthermore, guidance admittance control is performed such that the subject can easily operate the handle 26 of the grip lever 2 in the direction of the operation target position. As a result, the load applied to the subject can be suitably reduced. Thus, optimum rehabilitation training corresponding to the posture of the subject can be performed.

[0071] The invention is not limited to the embodiments and can be changed, as appropriate, without departing from the essence thereof. [0072] In the embodiment above, the control device 9 calculates the manipulability ellipse on the basis of the captured image of the upper extremity of the subject which is outputted from the camera 3, but such a feature is not limiting, and the manipulability ellipse can be calculated using any sensor. For example, the control device 9 may calculate the manipulability ellipse on the basis of a signal outputted from an inertia sensor mounted on the upper extremity of the subject. For example, the control device 9 calculates joint angles of the upper extremity on the basis of the signal outputted from the inertia sensor and the kinematic model of the joint angles which has been set in advance, and calculates the manipulability ellipse on the basis of the calculated joint angles and the Jacobian matrix of the upper extremity.

[0073] Further, the control device 9 may calculate the manipulability ellipse, for example, on the basis of a signal outputted form a depth sensor (KINECT®). For example, the control device 9 calculates the joint angles of the upper extremity on the basis of depth information outputted from the depth sensor and the kinematic model of joint angle which has been set in advance, and calculates the manipulability ellipse on the basis of the calculated joint angles and the Jacobian matrix of the upper extremity.

[0074] The control device 9 calculates the virtual guideline by performing admittance control with two degrees of freedom on the basis of the external force acting upon the handle of the grip lever 2 which has been detected by the force sensor 8, but such a feature is not limiting. The control device 9 may calculate the virtual guideline by performing impedance control with two degrees of freedom on the basis of the external force acting upon the handle of the grip lever 2 detected by the force sensor 8 in the upper-level control system.

[0075] In the embodiment, the grip lever 2 may have a configuration which has three joint shafts and in which the handle 26 moves to any position in a three-dimensional space. In this case, the manipulability ellipse calculation unit 93 performs singular value decomposition of the Jacobian matrix of the upper extremity, calculates the size of the long axis, the size of the short axis, and the orientation of the manipulability ellipse, which indicates the manipulability of the subject's fingers, and calculates the radius ratio λ_εΐϋ_ρ and inclination 0_eiii_p of the manipulability ellipse. The target value calculation unit 94 calculates the final target value p_fi_n of the tracking task on the basis of the radius ratio λ_εΐΗ_Ρ and inclination 8_eiii_P of the manipulability ellipse outputted by the manipulability ellipse calculation unit 93. Meanwhile, the admittance control unit 96 calculates the control target value p_ref by performing rotation transformation R(9_rot k), then the transformation (admittance control) with a 3-rows x 3-columns admittance matrix, and the inverse rotation transformation R^T(6_rot_k)-

[0076] In the embodiment, the admittance control unit 96 of the control device 9 may gradually increase the set damping constant D_adm2 and/or stiffness constant K_adm2 according to the displacement of the present position of the handle 26 of the grip lever 2 from the virtual guideline in the auxiliary shaft direction. As a result, the handle 26 of the grip lever 2 can be gently guided along the virtual guideline.

[0077] In the embodiment, the admittance control unit 96 of the control device 9 may rapidly increase the damping constants D_admi, Dadm2 and/or stiffness constant K_admi , Kadm2 by a predetermined amount when it is determined that the present position of the handle 26 of the grip lever 2 is outside of the manipulability ellipse. As a result, wherein the position of the handle 26 of the grip lever 2 is displaced to the outside of the manipulability ellipse, the operation of the handle 26 is abruptly fixed. Therefore, the handle can be prevented from displacing to the outside of the manipulability ellipse, and the posture unsuitable for the subject can be forcibly suppressed.

[0078] In the embodiment, the rehabilitation apparatus 1 performs the rehabilitation training of the upper extremity of the subject, but such feature is not limiting, and the rehabilitation training for the lower extremity of the subject may be also performed. The control device 9 calculates the manipulability ellipse on the basis of the image of the lower extremity posture of the subject which has been captured by the camera 3. The control device 9 performs the admittance control on the basis of the external force of the lower extremity which has been detected by the force sensor 8, and calculates the virtual guideline of the operation target position.

[0079] In the invention, for example, the processes depicted in FIG. 8 can be also realized by causing the CPU 9a to execute a computer program.

[0080] The program can be stored by using non-transitory computer-readable media of various types and provided to the computer. The non-transitory computer-readable media include tangible storage media of various types. Examples of the non-transitory computer-readable media include a magnetic recording medium (for example, a flexible disk, a magnetic tape, and a hard disk drive), a magneto-optical recording medium, for example, a magneto-optical disk), a CD-ROM, a CD-recordable (CD-R), a CD-rewritable (CD-R/W), a semiconductor memory (for example, a mask ROM, a programmable ROM (PROM), an erasable PROM (EPROM), a flash ROM, and a RAM).

[0081] The program may be provided to the computer by transitory computer-readable media of various types. Examples of transitory computer-readable media include electric signals, optical signals, and electromagnetic waves. The transitory computer-readable medium can provide the program to the computer via a wire communication path using electric wires or optical fibers, or via a wireless communication path.

Claims

CLAIMS:

1. A rehabilitation apparatus comprising:

an operation unit that is operated by a subject of rehabilitation;

an operation amount detection unit that detects an operation amount of the operation unit;

a drive unit that drives the operation unit;

a control unit that controls drive of the drive unit;

a posture detection unit that detects a posture of the subject;

an external force detection unit that detects an external force acting upon the operation unit; and

a display unit that displays an operation target position of the operation unit, wherein the control unit calculates a manipulability ellipse representing manipulability of a body of the subject based on the posture of the subject detected by the posture detection unit, calculates the operation target position based on the calculated manipulability ellipse, calculates a target value of the operation amount for the operation unit based on the external force detected by the external force detection unit and the calculated operation target position, and controls the driving unit such that the operation amount detected by the operation amount detection unit follows the calculated target value of the operation amount.

2. The rehabilitation apparatus according to claim 1, wherein

the control unit calculates the operation target position in a two-dimensional coordinate system on a display screen of the display unit based on a sine wave function of an amplitude including a ratio of a long diameter and a short diameter of the manipulability ellipse and an inclination of the manipulability ellipse.

3. The rehabilitation apparatus according to claim 2, wherein

the control unit calculates an operation target position p_fm in a two-dimensional coordinate system in the display unit by using the following expressions:

llip - ^

en en

ilip + ¹ ^Rel - ^Ren +

D __ D _ £

es en en wherein 6_rot __k is a rotation angle of a virtual guideline which is a line connecting the operation target position with a point of origin of the two-dimensional coordinate system in the display unit, λ_ειΐί_ρ is a ratio of the long diameter and the short diameter of the manipulability ellipse, 9_eiii_p is an inclination of the manipulability ellipse, and Ren is a nominal radius of the manipulability ellipse.

4. The rehabilitation apparatus according to claim 3, wherein

the control unit generates the virtual guideline connecting the operation target position with a point of origin of the two-dimensional coordinate system in the display unit and performs admittance control on the drive unit such that a load is low in a direction of the virtual guideline and high in a direction perpendicular to a virtual guideline direction.

5. The rehabilitation apparatus according to claim 1, wherein the control unit generates a virtual guideline connecting the operation target position with a point of origin of a two-dimensional coordinate system in the display unit and performs admittance control on the drive unit such that a load is low in a direction of the virtual guideline and high in a direction perpendicular to a virtual guideline direction.

6. The rehabilitation apparatus according to claim 2, wherein

the control unit generates a virtual guideline connecting the operation target position with a point of origin of the two-dimensional coordinate system in the display unit and performs admittance control on the drive unit such that a load is low in a direction of the virtual guideline and high in a direction perpendicular to a virtual guideline direction.

7. The rehabilitation apparatus according to any one of claims 4 to 6, wherein the control unit calculates the target value of the operation amount for the operation unit by using an admittance matrix in which at least one of a stiffness constant and a damping constant is set low in the direction of the virtual guideline and at least one of the stiffness constant and the damping constant is set high in a perpendicular direction.

8. The rehabilitation apparatus according to any one of claims 4 to 7, wherein the control unit transforms the external force detected by the external force detection unit into an external force in a global coordinate system, performs rotation transformation of the transformed external force based on a rotation angle of the virtual guideline, transforms the rotation-transformed value by using the admittance matrix, performs inverse rotation transformation of the transformed value based on the rotation angle of the virtual guideline, and transforms the value obtained by inverse rotation transformation by using inverse kinematics, to calculate the target value of the operation amount for the operation unit.

9. The rehabilitation apparatus according to any one of claims 4 to 7, wherein the control unit calculates a target value G_ref of the operation amount for the operation unit by using the following expression:

^GadrX^S) =

wherein G_rot_k is a rotation angle of the virtual guideline which is a line connecting the operation target position with the point of origin of the two-dimensional coordinate system in the display unit, rotation transformation R(6_rotjc) is a rotation transformation matrix, Gadm(s) is the admittance matrix, R^T(0_rot_k) is ^an inverse rotation transformation matrix, f_ext is the external force in a global coordinate system, D_admi is a damping constant in the direction of the virtual guideline, K_admi is a stiffness constant in the direction of the virtual guideline, D_adm2 is a damping constant in the perpendicular direction, and K_ad_m2 is a stiffness constant in the perpendicular direction.

10. The rehabilitation apparatus according to claim 7, wherein

the control unit gradually increases the set damping constant or the set stiffness constant according to a displacement of a present position on the display unit, which corresponds to the operation amount detected by the operation amount detection unit, from the virtual guideline in the perpendicular direction.

11. The rehabilitation apparatus according to claim 7, wherein

the control unit increases the damping constant and/or the stiffness constant by a predetermined amount when it is determined that a present position on the display unit, which corresponds to the operation amount detected by the operation amount detection unit, is outside of the manipulability ellipse.

12. The rehabilitation apparatus according to any one of claims 1 to 11, wherein the posture detection unit detects joint angles of the body of the subject, and the control unit calculates an operational ellipse based on the joint angles detected by the posture detection unit and a Jacobian matrix of the body.

13. A control method comprising:

detecting an operation amount of an operation unit that is operated by a subject of rehabilitation;

detecting a posture of the subject;

detecting an external force acting upon the operation unit;

displaying an operation target position of the operation unit;

calculating a manipulability ellipse representing manipulability of a body of the subject based on the detected posture of the subject;

calculating the operation target position based on the calculated manipulability ellipse;

calculating a target value of an operation amount for the operation unit based on the detected external force and the calculated operation target position; and

controlling a driving unit for driving the operation unit, such that the detected operation amount follows the calculated target value of the operation amount.

14. A control program comprising:

a process of calculating a manipulability ellipse representing manipulability of a body of a subject of rehabilitation based on a posture of the subject;

a process of calculating an operation target position based on the calculated manipulability ellipse;

a process of displaying the operation target position of an operation unit operated by the subject;

a process of calculating a target value of an operation amount for the operation unit based on an external force acting upon the operation unit and the calculated operation target position; and

a process of controlling a driving unit for driving the operation unit, such that an operation amount of the operation unit follows the calculated target value of the operation amount.