CN113995629A - Upper limb double-arm rehabilitation robot admittance control method and system based on mirror force field - Google Patents

Upper limb double-arm rehabilitation robot admittance control method and system based on mirror force field Download PDF

Info

Publication number
CN113995629A
CN113995629A CN202111295849.2A CN202111295849A CN113995629A CN 113995629 A CN113995629 A CN 113995629A CN 202111295849 A CN202111295849 A CN 202111295849A CN 113995629 A CN113995629 A CN 113995629A
Authority
CN
China
Prior art keywords
representing
subject
matrix
intention
force field
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Granted
Application number
CN202111295849.2A
Other languages
Chinese (zh)
Other versions
CN113995629B (en
Inventor
李智军
苏航
李国欣
康宇
刘碧珊
王昶茹
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Institute of Advanced Technology University of Science and Technology of China
Shanghai Robot Industrial Technology Research Institute Co Ltd
Original Assignee
Institute of Advanced Technology University of Science and Technology of China
Shanghai Robot Industrial Technology Research Institute Co Ltd
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Institute of Advanced Technology University of Science and Technology of China, Shanghai Robot Industrial Technology Research Institute Co Ltd filed Critical Institute of Advanced Technology University of Science and Technology of China
Priority to CN202111295849.2A priority Critical patent/CN113995629B/en
Publication of CN113995629A publication Critical patent/CN113995629A/en
Application granted granted Critical
Publication of CN113995629B publication Critical patent/CN113995629B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Images

Classifications

    • A—HUMAN NECESSITIES
    • A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
    • A61H—PHYSICAL THERAPY APPARATUS, e.g. DEVICES FOR LOCATING OR STIMULATING REFLEX POINTS IN THE BODY; ARTIFICIAL RESPIRATION; MASSAGE; BATHING DEVICES FOR SPECIAL THERAPEUTIC OR HYGIENIC PURPOSES OR SPECIFIC PARTS OF THE BODY
    • A61H1/00—Apparatus for passive exercising; Vibrating apparatus; Chiropractic devices, e.g. body impacting devices, external devices for briefly extending or aligning unbroken bones
    • A61H1/02—Stretching or bending or torsioning apparatus for exercising
    • A61H1/0274—Stretching or bending or torsioning apparatus for exercising for the upper limbs
    • A—HUMAN NECESSITIES
    • A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
    • A61H—PHYSICAL THERAPY APPARATUS, e.g. DEVICES FOR LOCATING OR STIMULATING REFLEX POINTS IN THE BODY; ARTIFICIAL RESPIRATION; MASSAGE; BATHING DEVICES FOR SPECIAL THERAPEUTIC OR HYGIENIC PURPOSES OR SPECIFIC PARTS OF THE BODY
    • A61H2201/00—Characteristics of apparatus not provided for in the preceding codes
    • A61H2201/50—Control means thereof
    • A61H2201/5007—Control means thereof computer controlled
    • A61H2201/501—Control means thereof computer controlled connected to external computer devices or networks
    • A—HUMAN NECESSITIES
    • A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
    • A61H—PHYSICAL THERAPY APPARATUS, e.g. DEVICES FOR LOCATING OR STIMULATING REFLEX POINTS IN THE BODY; ARTIFICIAL RESPIRATION; MASSAGE; BATHING DEVICES FOR SPECIAL THERAPEUTIC OR HYGIENIC PURPOSES OR SPECIFIC PARTS OF THE BODY
    • A61H2230/00—Measuring physical parameters of the user
    • A61H2230/08—Other bio-electrical signals
    • A61H2230/085—Other bio-electrical signals used as a control parameter for the apparatus
    • Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
    • Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
    • Y02P—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN THE PRODUCTION OR PROCESSING OF GOODS
    • Y02P90/00—Enabling technologies with a potential contribution to greenhouse gas [GHG] emissions mitigation
    • Y02P90/02—Total factory control, e.g. smart factories, flexible manufacturing systems [FMS] or integrated manufacturing systems [IMS]

Landscapes

  • Health & Medical Sciences (AREA)
  • Epidemiology (AREA)
  • Pain & Pain Management (AREA)
  • Physical Education & Sports Medicine (AREA)
  • Rehabilitation Therapy (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Animal Behavior & Ethology (AREA)
  • General Health & Medical Sciences (AREA)
  • Public Health (AREA)
  • Veterinary Medicine (AREA)
  • Rehabilitation Tools (AREA)

Abstract

The invention provides an admittance control method and system of an upper limb double-arm rehabilitation robot based on a mirror force field, comprising the following steps: step 1, modeling a human-computer tight coupling healthy lateral force field based on multi-sensing signal fusion to obtain the movement intention of healthy lateral of a subject; step 2: mapping a physiological signal and a force field of the healthy side based on a state space according to the movement intention of the healthy side of the subject to obtain a movement track and intention of the affected side of the subject; and step 3: and performing force field mirror image-based healthy and affected side synchronous coupling control according to the motion trail and intention of the affected side of the subject, and further controlling the motion of the exoskeleton. Aiming at the great clinical requirement of upper limb movement function reconstruction after clinical nerve displacement, the invention combines a force field control strategy of man-machine tight coupling with a mirror image rehabilitation strategy, explores a new mirror image force field rehabilitation strategy for guiding the motion of the affected side based on the patient healthy side force field information, is more natural, and improves the participation sense and the active rehabilitation ability of the patient.

Description

Upper limb double-arm rehabilitation robot admittance control method and system based on mirror force field
Technical Field
The invention relates to the technical field of admittance control of upper limb double-arm rehabilitation robots, in particular to an admittance control method and system of an upper limb double-arm rehabilitation robot based on a mirror force field.
Background
The peripheral nerve injury is a clinical frequently-occurring disease, tens of millions of patients suffering from new wounds are suffered from each year, the most serious peripheral nerve injury, such as brachial plexus injury, can cause complete paralysis of one upper limb, the life quality of the patients is seriously affected, and the treatment of the peripheral nerve injury is a worldwide problem. Currently, the best treatment method for avulsion of brachial plexus is recognized as nerve displacement. Rehabilitation therapy is the key to postoperative upper limb functional recovery, the cerebral cortex can be widely remodeled in the recovery process, and the remodeling result is important for clinical prognosis. Research shows that after the nerve displacement operation, the original functional area of the affected limb in the brain is reactivated through remodeling and the affected limb is effectively controlled. However, in clinical rehabilitation, the functional recovery of many patients has the problems of poor limb movement control, wrong movement pattern and the like, and the fundamental reason is that the peripheral nerve innervation and the passage of the affected limb are greatly changed after the nerve displacement.
Research proves that compared with single unilateral rehabilitation training, the rehabilitation training is more accordant with the natural motion mode of the upper limbs of the human body by guiding the affected side to complete the rehabilitation training through the healthy side, and the rehabilitation training is favorable for the neural plasticity of the half brain of the affected side and is more favorable for improving the rehabilitation effect of the motion function of the affected limb of the patient. The mirror image therapy is a treatment means for guiding the movement of the affected side by the side-healthy information which is mainly adopted in the traditional clinical practice. The mirror image therapy is also called as mirror image visual feedback therapy, and the mirror image therapy copies the healthy side moving picture to the affected side by utilizing the plane mirror imaging principle, so that the patient imagines the motion of the affected side, and a rehabilitation training treatment means is realized by combining visual illusion, visual feedback and virtual reality. In the mirror image treatment, after seeing the mirror image of the movement of the healthy side, the patient activates the mirror image neuron of the corresponding cerebral cortex, which is beneficial to recovering the movement function of the affected side. However, the mirror is not strong in immersion in the method of being used as a mirror image carrier, and the stability and the promotion of the clinical research result of the mirror image therapy are directly influenced. In addition, the rehabilitation training of the traditional mirror image therapy can only realize the control of the motion trail of the upper limb, and the stress state of the muscle group of the affected limb is ignored, so that the improvement of the clinical research effect of the mirror image therapy is directly influenced.
Chinese patent publication No. CN109091818A discloses a training method and system for a rope-drawn upper limb rehabilitation robot based on admittance control, which collects in real time an interaction force signal applied by the upper limb of a user to the rope-drawn rehabilitation robot and a kinematic signal of the upper limb when the user performs rehabilitation training movement of the upper limb joint; converting the interaction force signal into a motion parameter of an expected motion track through an admittance model, and determining a motion parameter of a target motion track according to the motion parameter of the expected motion track and the kinematic signal of the upper limb; the determined motion parameters are used as control quantities and converted into motor control quantities of the rope traction rehabilitation robot to control the corresponding motor output, so that the user can autonomously control the rehabilitation training action, and the active participation of the user is improved.
In view of the above-mentioned related technologies, the inventor believes that the above-mentioned method is a method for assisting the recovery of the affected side by performing visual feedback side-strengthening exercise with a mirror after the nerve repair, and the patient has weak participation and general recovery effect.
Disclosure of Invention
Aiming at the defects in the prior art, the invention aims to provide an upper limb double-arm rehabilitation robot method and system based on a mirror force field.
The invention provides an admittance control method of an upper limb double-arm rehabilitation robot based on a mirror force field, which comprises the following steps:
step 1: modeling a human-computer tight coupling healthy lateral force field based on multi-sensing signal fusion to obtain the movement intention of the healthy lateral of the subject;
step 2: mapping a physiological signal and a force field of the healthy side based on a state space according to the movement intention of the healthy side of the subject to obtain a movement track and intention of the affected side of the subject;
and step 3: and performing force field mirror image-based healthy and affected side synchronous coupling control according to the motion trail and intention of the affected side of the subject, and further controlling the motion of the exoskeleton.
Preferably, the step 1 includes predicting the movement intention of the subject in real time through a healthy lateral electromyography sensor, modeling the acting force in the interaction process as an impedance model, and predicting the joint state of the subject through the impedance model, wherein the impedance model is as shown in formula (1):
Figure BDA0003336561920000021
wherein u ishActing force of the upper limb double-arm rehabilitation robot in the interaction process with the testee; x is the actual position of the tail end of the upper limb double-arm rehabilitation robot; x is the number ofrThe expected position of the tail end of the upper limb double-arm rehabilitation robot; superscript symbol · representing the derivative of the corresponding state quantity with respect to time; l ish,1Is the position error gain; l ish,2Is the speed gain;
estimating the motor intention of the subject by equation (1), as shown in equation (2):
Figure BDA0003336561920000022
wherein the content of the first and second substances,
Figure BDA0003336561920000023
an estimated value representing the exercise intention of the healthy side of the subject, and a superscript symbol ^ represents an estimated value of the corresponding quantity;
Figure BDA0003336561920000024
an initial value representing an error gain at an arbitrary virtual target position;
Figure BDA0003336561920000025
an initial value representing the gain at any virtual target speed; the superscript v indicates that the value is given an arbitrary initial value based on the virtual target.
Preferably, the step 2 comprises the following steps:
step 2.1: modeling a healthy lateral force field and a physiological electromyographic signal of the subject according to the step 1, namely obtaining the movement intention of the healthy lateral of the subject
Figure BDA0003336561920000031
Step 2.2: the two arms of the affected side of the testee do rehabilitation motions along the same track, and the motion track and intention of the affected side of the testee are obtained through the mirror image principle.
Preferably, the step 3 includes combining the interaction force generated by the affected side during the interaction process with the established affected side motion trajectory and intention to control the motion of the exoskeleton through admittance control, wherein the affected side intention is expressed as:
Figure BDA0003336561920000032
wherein the content of the first and second substances,
Figure BDA0003336561920000033
a subject intent predicted for the robust model; tau isrThe original movement intention of the affected model; λ is a hyper-parameter that adjusts the ratio of the two weights.
Preferably, the admittance control in step 3 includes:
the dynamic equation of the interaction process of the upper limb double-arm rehabilitation robot and the testee is shown in formula (4):
Figure BDA0003336561920000034
wherein M and G respectively represent inertia of the upper limb exoskeleton robot and the human interaction system under a Cartesian space coordinate systemA sexual matrix and a gravity matrix; c represents a Coriolis force and centrifugal force matrix of the upper limb exoskeleton robot and the human interaction system under a Cartesian space coordinate system; f. ofdisIs a disturbance in the interactive system; u is the control input of the system; the superscript · represents the second derivative of the actual position of the rehabilitation robot tip over time, i.e. the acceleration;
suppose the actual position x of the end of the upper arm two-arm rehabilitation robot and the derivative of the actual position of the end of the upper arm two-arm rehabilitation robot with respect to time
Figure BDA0003336561920000035
Is obtained by measurement; let x1＝[q1,q2,…,qn]T,
Figure BDA0003336561920000036
Wherein q isiAnd
Figure BDA0003336561920000037
respectively representing the rotation angle and the angular speed of the ith joint, wherein i is more than or equal to 1 and less than or equal to n; x is the number of1Representing a position matrix formed by rotation angles of each joint of the robot; x is the number of2Representing a velocity matrix composed of angular velocities of the joints; superscript T denotes transpose; the dynamics of the interaction task is represented in the form:
Figure BDA0003336561920000038
defining a position error z1＝x1-xrVelocity error z2＝x2-α1，α1Is to z1The virtual control of (2) is as follows:
Figure BDA0003336561920000039
using Lyapunov functions
Figure BDA00033365619200000310
V1A function representing the constructed Lyapunov function form; symbol denotes matrix multiplication; the time is derived as follows:
Figure BDA00033365619200000311
order to
Figure BDA00033365619200000312
Wherein K1To gain the matrix, equation (7) is reset:
Figure BDA00033365619200000313
from equation (8):
Figure BDA0003336561920000041
defining Lyapunov functions
Figure BDA0003336561920000042
V2A function representing the constructed Lyapunov function form; the time is derived as follows:
Figure BDA0003336561920000043
when the parameters of the dynamics are known, the control is expressed in the form:
Figure BDA0003336561920000044
wherein, K2Representing a gain matrix;
approximating G, C and M terms of robot dynamics using a radial basis function neural network; the external disturbance is compensated by a disturbance observer; the self-adaptive control law is as follows:
Figure BDA0003336561920000045
wherein the content of the first and second substances,
Figure BDA0003336561920000046
the radial basis function network is a radial basis function neural network, W is a weight coefficient, Y (Z) is a dynamic regression matrix, namely the distance between a sample point and each radial basis center, and Z represents the input of the radial function network; order to
Figure BDA0003336561920000047
The high-order disturbance observer is in the form of:
Figure BDA0003336561920000048
wherein, KdRepresenting a gain matrix during disturbance observation;
Figure BDA0003336561920000049
representing an estimation error; y isd(Zd) Representing a dynamic regression matrix, Yd(Zd) Representing a dynamic regression matrix; zdRepresenting actual sampling points; wdRepresenting a weight coefficient;
Figure BDA00033365619200000410
Figure BDA00033365619200000411
the weight matrix is updated as follows:
Figure BDA00033365619200000412
Figure BDA00033365619200000413
Figure BDA00033365619200000414
Yd(Zd)Wd＝M-1(u+uh-C(x1,x2)x2-G(x1))-∈d
wherein, Yi(Z) representing an updated value of the dynamic regression quantity matrix; z is a radical of2iAn update indicative of a speed error; wiIndicating an update of the estimate; superscript symbol
Figure BDA00033365619200000415
An expected value representing a derivative of the weight; wdiRepresenting an updated value of a physical parameter; e represents an estimation error; e is the same asdRepresenting the desired estimation error; y (Z) W represents the output of the radial basis function; gamma-shapediAnd ΓdiTo update the rate, θiAnd thetadiAre weights.
The invention provides an admittance control system of an upper limb double-arm rehabilitation robot based on a mirror force field, which comprises the following modules:
module M1: modeling a human-computer tight coupling healthy lateral force field based on multi-sensing signal fusion to obtain the movement intention of the healthy lateral of the subject;
module M2: mapping a physiological signal and a force field of the healthy side based on a state space according to the movement intention of the healthy side of the subject to obtain a movement track and intention of the affected side of the subject;
module M3: and performing force field mirror image-based healthy and affected side synchronous coupling control according to the motion trail and intention of the affected side of the subject, and further controlling the motion of the exoskeleton.
Preferably, the module M1 includes predicting the movement intention of the subject in real time through a healthy lateral electromyography sensor, modeling the acting force during the interaction as an impedance model, and predicting the joint state of the subject through the impedance model, wherein the impedance model is as shown in formula (1):
Figure BDA0003336561920000051
wherein u ishActing force of the upper limb double-arm rehabilitation robot in the interaction process with the testee; x is the actual position of the tail end of the upper limb double-arm rehabilitation robot; x is the number ofrThe expected position of the tail end of the upper limb double-arm rehabilitation robot; superscript symbol · representing the derivative of the corresponding state quantity with respect to time; l ish,1Is the position error gain; l ish,2Is the speed gain;
estimating the motor intention of the subject by equation (1), as shown in equation (2):
Figure BDA0003336561920000052
wherein the content of the first and second substances,
Figure BDA0003336561920000053
an estimated value representing the exercise intention of the healthy side of the subject, and a superscript symbol ^ represents an estimated value of the corresponding quantity;
Figure BDA0003336561920000054
an initial value representing an error gain at an arbitrary virtual target position;
Figure BDA0003336561920000055
an initial value representing the gain at any virtual target speed; the superscript v indicates that the value is given an arbitrary initial value based on the virtual target.
Preferably, the module M2 includes the following modules:
module M2.1: modeling the healthy lateral force field and the physiological electromyographic signal of the subject according to the module M1, namely obtaining the exercise intention of the healthy lateral of the subject
Figure BDA0003336561920000056
Module M2.2: the two arms of the affected side of the testee do rehabilitation motions along the same track, and the motion track and intention of the affected side of the testee are obtained through the mirror image principle.
Preferably, the module M3 includes controlling the exoskeleton's motion through admittance control by combining the interaction force generated by the affected side during the interaction process with the established affected side motion trajectory and intent, wherein the affected side intent is represented as:
Figure BDA0003336561920000057
wherein the content of the first and second substances,
Figure BDA0003336561920000058
a subject intent predicted for the robust model; tau isrThe original movement intention of the affected model; λ is a hyper-parameter that adjusts the ratio of the two weights.
Preferably, the admittance control in the module M3 includes:
the dynamic equation of the interaction process of the upper limb double-arm rehabilitation robot and the testee is shown in formula (4):
Figure BDA0003336561920000061
m and G respectively represent an inertia matrix and a gravity matrix of the upper limb exoskeleton robot and the human interaction system under a Cartesian space coordinate system; c represents a Coriolis force and centrifugal force matrix of the upper limb exoskeleton robot and the human interaction system under a Cartesian space coordinate system; f. ofdisIs a disturbance in the interactive system; u is the control input of the system; the superscript · represents the second derivative of the actual position of the rehabilitation robot tip over time, i.e. the acceleration;
suppose the actual position x of the end of the upper arm two-arm rehabilitation robot and the derivative of the actual position of the end of the upper arm two-arm rehabilitation robot with respect to time
Figure BDA0003336561920000062
Is obtained by measurement; let x1＝[q1,q2,…,qn]T,
Figure BDA0003336561920000063
Wherein q isiAnd
Figure BDA0003336561920000064
respectively representing the rotation angle and the angular speed of the ith joint, wherein i is more than or equal to 1 and less than or equal to n; x is the number of1Representing a position matrix formed by rotation angles of each joint of the robot; x is the number of2Representing a velocity matrix composed of angular velocities of the joints; superscript T denotes transpose; the dynamics of the interaction task is represented in the form:
Figure BDA0003336561920000065
defining a position error z1＝x1-xrVelocity error z2＝x2-α1，α1Is to z1The virtual control of (2) is as follows:
Figure BDA0003336561920000066
using Lyapunov functions
Figure BDA0003336561920000067
V1A function representing the constructed Lyapunov function form; symbol denotes matrix multiplication; the time is derived as follows:
Figure BDA0003336561920000068
order to
Figure BDA0003336561920000069
Wherein K1To gain the matrix, equation (7) is reset:
Figure BDA00033365619200000610
from equation (8):
Figure BDA00033365619200000611
defining Lyapunov functions
Figure BDA00033365619200000612
V2A function representing the constructed Lyapunov function form; the time is derived as follows:
Figure BDA00033365619200000613
when the parameters of the dynamics are known, the control is expressed in the form:
Figure BDA00033365619200000614
wherein, K2Representing a gain matrix;
approximating G, C and M terms of robot dynamics using a radial basis function neural network; the external disturbance is compensated by a disturbance observer; the self-adaptive control law is as follows:
Figure BDA00033365619200000615
wherein the content of the first and second substances,
Figure BDA00033365619200000616
the radial basis function network is a radial basis function neural network, W is a weight coefficient, Y (Z) is a dynamic regression matrix, namely the distance between a sample point and each radial basis center, and Z represents the input of the radial function network; order to
Figure BDA0003336561920000071
The high-order disturbance observer is in the form of:
Figure BDA0003336561920000072
wherein, KdRepresenting a gain matrix during disturbance observation;
Figure BDA0003336561920000073
representing an estimation error; y isd(Zd) Representing a dynamic regression matrix, Yd(Zd) Representing a dynamic regression matrix; zdRepresenting actual sampling points; wdRepresenting a weight coefficient;
Figure BDA0003336561920000074
Figure BDA0003336561920000075
the weight matrix is updated as follows:
Figure BDA0003336561920000076
Figure BDA0003336561920000077
Figure BDA0003336561920000078
Yd(Zd)Wd＝M-1(u+uh-C(x1,x2)x2-G(x1))-∈d
wherein, Yi(Z) representing an updated value of the dynamic regression quantity matrix; z is a radical of2iAn update indicative of a speed error; wiIndicating an update of the estimate; superscript symbol
Figure BDA0003336561920000079
An expected value representing a derivative of the weight; wdiRepresenting an updated value of a physical parameter; e represents an estimation error; e is the same asdRepresenting the desired estimation error; y (Z) W represents the output of the radial basis function; gamma-shapediAnd ΓdiTo update the rate, θiAnd thetadiAre weights.
Compared with the prior art, the invention has the following beneficial effects:
1. aiming at the great clinical requirement of upper limb movement function reconstruction after clinical nerve displacement, the invention combines a force field control strategy of man-machine tight coupling with a mirror image rehabilitation strategy, explores a new mirror image force field rehabilitation strategy for guiding the motion of the affected side based on the patient healthy side force field information, and the method is more natural and improves the participation sense and the active rehabilitation capability of the patient;
2. aiming at the important clinical requirement of motor function reconstruction after the clinical nerve displacement, the invention explores a force field mirror image rehabilitation strategy for effectively guiding the affected side action based on the patient side-strengthening force field information based on the new technology in the engineering and medical fields, and newly explores a medical rehabilitation method, and the research result is an important breakthrough in the peripheral nerve rehabilitation research field, thereby not only driving the clinical treatment method in the peripheral nerve rehabilitation research field to be significantly innovated, but also providing a new technical means for exploring the mechanism of function reconstruction and brain function recovery after the nerve displacement, and having important academic value and clinical significance; the results will be available to hospitals, rehabilitation centers and communities to benefit patients in a shared manner;
3. the invention realizes the body-affected side force field mirror coupling of the upper limb rehabilitation robot by the admittance control method, and obtains the affected side rehabilitation training effect which accords with the real motion habit of the patient.
Drawings
Other features, objects and advantages of the invention will become more apparent upon reading of the detailed description of non-limiting embodiments with reference to the following drawings:
fig. 1 is a schematic diagram of an admittance control system and method of an upper limb double-arm rehabilitation robot based on a mirror force field according to the present invention;
FIG. 2 is a schematic diagram of the adjustment of the hyper-parameter λ in different rehabilitation training phases according to the present invention;
fig. 3 is a block diagram of an admittance control method of the present invention.
Detailed Description
The present invention will be described in detail with reference to specific examples. The following examples will assist those skilled in the art in further understanding the invention, but are not intended to limit the invention in any way. It should be noted that it would be obvious to those skilled in the art that various changes and modifications can be made without departing from the spirit of the invention. All falling within the scope of the present invention.
The embodiment of the invention discloses an admittance control method of an upper limb double-arm rehabilitation robot based on a mirror image force field, which comprises man-machine tight coupling side force field modeling based on multi-sensing signal fusion, health affected side physiological signals and force field mapping based on a state space, and health affected side synchronous coupling control based on force field mirror image, as shown in fig. 1 and fig. 2. The healthy side includes a healthy side and an affected side. The method comprises the following steps: step 1: and (3) modeling a human-computer tight coupling side-healthy force field based on multi-sensing signal fusion to obtain the movement intention of the side-healthy of the subject. The motor intention of the subject is predicted in real time through the healthy lateral electromyography sensor, acting force in the interaction process is modeled into an impedance model, and the joint state of the subject is predicted through the impedance model. The human-computer tight coupling side-healthy force field modeling based on multi-sensing signal fusion comprises the steps of estimating and predicting human movement intention through a side-healthy electromyographic sensor, modeling acting force in an interaction process into an impedance model, and predicting the joint state of a human by using the model, wherein the impedance model is as shown in a formula (1):
Figure BDA0003336561920000081
wherein u ishActing force in the interaction process of the upper limb double-arm rehabilitation robot and the testee, namely the interaction force between the upper limb double-arm rehabilitation robot and the testee; x is the actual position of the tail end of the upper limb double-arm rehabilitation robot, xrFor the desired position of the extremity of the upper arm dual-arm rehabilitation robot, the superscript symbol represents the derivative of the corresponding state quantity with respect to time, Lh,1As a position error gain, Lh,2Is the speed gain.
Estimating the motor intention of the subject by equation (1), as shown in equation (2):
Figure BDA0003336561920000082
wherein the content of the first and second substances,
Figure BDA0003336561920000083
an estimated value representing the exercise intention of the healthy side of the subject, and a superscript symbol ^ represents an estimated value of the corresponding quantity;
Figure BDA0003336561920000084
an initial value representing an error gain at an arbitrary virtual target position;
Figure BDA0003336561920000085
an initial value representing the gain at any virtual target speed; the superscript v indicates that the value is given an arbitrary initial value based on the virtual target.
Step 2: and mapping the physiological signal and the force field of the healthy side based on the state space according to the movement intention of the healthy side of the subject to obtain the movement track and intention of the affected side of the subject. The step 2 comprises the following steps: step 2.1: modeling a healthy lateral force field and a physiological electromyographic signal of the subject according to the step 1, namely obtaining the movement intention of the healthy lateral of the subject
Figure BDA0003336561920000091
Step 2.2: the two arms of the affected side of the testee do rehabilitation motions along the same track, and the motion track and intention of the affected side of the testee are obtained through the mirror image principle.
The specific process of the healthy side physiological signal and force field mapping based on the state space comprises the following steps: and (3) modeling a healthy side force field and a physiological electromyographic signal of the subject according to the method in the step 1, namely obtaining the exercise intention of the healthy side of the subject. In the upper arm rehabilitation training process, the healthy and affected side of the testee performs rehabilitation actions along the same track, and the movement track and intention of the affected side of the testee are obtained through the mirror image principle.
And step 3: as shown in fig. 1 and fig. 3, the healthy and sick side synchronous coupling control based on force field mirror image is carried out according to the motion trail and intention of the sick side of the subject, so as to control the motion of the exoskeleton.
And combining the interaction force generated by the affected side in the interaction process with the established affected side motion track and intention, and controlling the motion of the exoskeleton through admittance control. The healthy and affected side synchronous coupling control based on force field mirror image combines the interaction force generated by the affected side in the interaction process with the established affected side motion track and intention, and controls the motion of the exoskeleton by an admittance control method, wherein the affected side intention is expressed as:
Figure BDA0003336561920000092
wherein the content of the first and second substances,
Figure BDA0003336561920000093
a subject intent predicted for the robust model; tau isrThe original movement intention of the affected model, and lambda is a hyper-parameter for adjusting the weight ratio of the two. Can be through adjusting lambda at hemiplegia patient different stages of rehabilitation training, at the rehabilitation training initial stage, the interaction force influence between sick side and the rehabilitation robot is less, uses the mirror image of the movement intention of healthy side as the guide promptly lambda 1 at this moment, and the accessible reduces lambda along with the increase of rehabilitation training and increases the movement intention of sick side in the rehabilitation robot control process, and different rehabilitation training stages surpass parameter lambda adjustment schematic diagrams and are shown in fig. 2.
Force field mirror image-based health affected side synchronous coupling control, wherein the specific process of the admittance control method is as follows, namely admittance control comprises: the dynamic equation of the interaction process of the upper limb double-arm rehabilitation robot and the testee is shown in formula (4):
Figure BDA0003336561920000094
m and G respectively represent an inertia matrix and a gravity matrix of the upper limb exoskeleton robot and human interaction system under a Cartesian space coordinate system, and C represents a subject of the upper limb exoskeleton robot and human interaction system under the Cartesian space coordinate systemA matrix of force and centrifugal force; f. ofdisU is the control input of the system for the disturbance in the interactive system; u. ofhActing force of the upper limb double-arm rehabilitation robot in the interaction process with the testee; the superscript, here in particular, represents the first derivative of the actual position of the rehabilitation robot tip over time, i.e. the robot tip velocity, and the superscript represents the second derivative of the actual position of the rehabilitation robot tip over time, i.e. the acceleration.
Suppose the actual position x of the end of the upper arm two-arm rehabilitation robot and the derivative of the actual position of the end of the upper arm two-arm rehabilitation robot with respect to time
Figure BDA0003336561920000101
Is measured. Let x1＝[q1,q2,…,qn]T，
Figure BDA0003336561920000102
Wherein q isiAnd
Figure BDA0003336561920000103
respectively representing the rotation angle and the angular speed of the ith joint, wherein i is more than or equal to 1 and less than or equal to n; x is the number of1Representing a position matrix formed by rotation angles of each joint of the robot; x is the number of2Representing a velocity matrix composed of angular velocities of the joints; superscript T denotes transpose; the dynamics of the interaction task is represented in the form:
Figure BDA0003336561920000104
defining a position error z1＝x1-xrVelocity error z2＝x2-α1Wherein x isrFor the desired position of the end of the upper arm double-arm rehabilitation robot, i.e. the desired reference track, alpha1Is to z1The virtual control of (2) is as follows:
Figure BDA0003336561920000105
consider the use of the Lyapunov function
Figure BDA0003336561920000106
V1Function V representing the form of the constructed Lyapunov function1(ii) a Symbol denotes matrix multiplication; the time is derived as follows:
Figure BDA0003336561920000107
order to
Figure BDA0003336561920000108
Wherein K1Equation (7) is reset for the gain matrix. Namely, formula (7) is rewritten to obtain:
Figure BDA0003336561920000109
from equation (8):
Figure BDA00033365619200001010
defining Lyapunov functions
Figure BDA00033365619200001011
V2Function V representing the form of the constructed Lyapunov function2(ii) a The time is derived as follows:
Figure BDA00033365619200001012
when the parameters of the dynamics are all known, the control method is expressed in the form:
Figure BDA00033365619200001013
wherein, K2To representA gain matrix.
Due to interference fdisThe precise information of (a) is difficult to obtain and the items G, C, M, etc. of robot dynamics are also not readily available. The G, C and M terms of robot dynamics are approximated using a Radial Basis Function Neural Network (RBFNN). In addition, external disturbances are compensated by a disturbance observer; the self-adaptive control law is as follows:
Figure BDA00033365619200001014
wherein the content of the first and second substances,
Figure BDA00033365619200001015
the radial basis function network input is a radial basis function neural network RBFNN, W is a weight coefficient, Y (Z) is a dynamic regression matrix, namely the distance between a sample point and each radial basis center, and Z represents the radial function network input; order to
Figure BDA00033365619200001016
The high-order disturbance observer is in the form of:
Figure BDA00033365619200001017
wherein, KdRepresenting a gain matrix during disturbance observation;
Figure BDA0003336561920000111
representing an estimation error; y isd(Zd) Representing a dynamic regression matrix, YdRepresenting a dynamic regression matrix; zdRepresenting an actual sample data set; wdRepresenting a weight coefficient;
Figure BDA0003336561920000112
Figure BDA0003336561920000113
the weight matrix is updated as follows:
Figure BDA0003336561920000114
Figure BDA0003336561920000115
Figure BDA0003336561920000116
Yd(Zd)Wd＝M-1(u+uh-C(x1，x2)x2-G(x1))-∈d
wherein, Yi(Z) representing an updated value of the dynamic regression quantity matrix; z is a radical of2iAn update indicative of a speed error; wiIndicating an update of the estimate; superscript symbol
Figure BDA0003336561920000117
An expected value representing a derivative of the weight; wdiRepresenting an updated value of a physical parameter; e represents an estimation error; e is the same asdRepresenting the desired estimation error; y (Z) W represents the output of the radial basis function; gamma-shapediAnd ΓdiTo update the rate, θiAnd thetadiAre weights. The characteristics of the regressor are also utilized in the disturbance observer.
The admittance control method is as shown in fig. 3, the reference track of the rehabilitation robot is reconstructed based on the physiological signal at the sick side of the state space and the force field mapping, and the control method can be suitable for people with different skill levels and different forces without off-line model adjustment, and ensures the robustness of the controller. The control scheme consists of an inner loop and an outer loop. The former can handle unknown mass and moment of inertia in robot dynamics, the latter is to adjust the interaction model taking into account human subject intent.
The embodiment of the invention discloses an admittance control system of an upper limb double-arm rehabilitation robot based on a mirror force field, which comprises the following modules: module M1: and (3) modeling a human-computer tight coupling side-healthy force field based on multi-sensing signal fusion to obtain the movement intention of the side-healthy of the subject. Predicting the movement intention of the subject in real time through a healthy side electromyography sensor, modeling acting force in the interaction process into an impedance model, and predicting the joint state of the subject through the impedance model, wherein the impedance model is shown as a formula (1):
Figure BDA0003336561920000118
wherein u ishActing force of the upper limb double-arm rehabilitation robot in the interaction process with the testee; x is the actual position of the tail end of the upper limb double-arm rehabilitation robot; x is the number ofrThe expected position of the tail end of the upper limb double-arm rehabilitation robot; superscript symbol · representing the derivative of the corresponding state quantity with respect to time; l ish,1Is the position error gain; l ish,2Is the speed gain;
estimating the motor intention of the subject by equation (1), as shown in equation (2):
Figure BDA0003336561920000119
wherein the content of the first and second substances,
Figure BDA00033365619200001110
an estimated value representing the exercise intention of the healthy side of the subject, and a superscript symbol ^ represents an estimated value of the corresponding quantity;
Figure BDA00033365619200001111
an initial value representing an error gain at an arbitrary virtual target position;
Figure BDA00033365619200001112
an initial value representing the gain at any virtual target speed; the superscript v indicates that the value is given an arbitrary initial value based on the virtual target.
Module M2: and mapping the physiological signal and the force field of the healthy side based on the state space according to the movement intention of the healthy side of the subject to obtain the movement track and intention of the affected side of the subject. Module M2The system comprises the following modules: module M2.1: modeling the healthy lateral force field and the physiological electromyographic signal of the subject according to the module M1, namely obtaining the exercise intention of the healthy lateral of the subject
Figure BDA0003336561920000121
Module M2.2: the two arms of the affected side of the testee do rehabilitation motions along the same track, and the motion track and intention of the affected side of the testee are obtained through the mirror image principle.
Module M3: and performing force field mirror image-based healthy and affected side synchronous coupling control according to the motion trail and intention of the affected side of the subject, and further controlling the motion of the exoskeleton. And combining the interaction force generated by the affected side in the interaction process with the established affected side motion track and intention, and controlling the motion of the exoskeleton through admittance control, wherein the affected side intention is expressed as:
Figure BDA0003336561920000122
wherein the content of the first and second substances,
Figure BDA0003336561920000123
a subject intent predicted for the robust model; tau isrThe original movement intention of the affected model; λ is a hyper-parameter that adjusts the ratio of the two weights.
The admittance control includes: the dynamic equation of the interaction process of the upper limb double-arm rehabilitation robot and the testee is shown in formula (4):
Figure BDA0003336561920000124
m and G respectively represent an inertia matrix and a gravity matrix of the upper limb exoskeleton robot and the human interaction system under a Cartesian space coordinate system; c represents a Coriolis force and centrifugal force matrix of the upper limb exoskeleton robot and the human interaction system under a Cartesian space coordinate system; f. ofdisIs a disturbance in the interactive system; u is the control input of the system; the superscript · denotes the second derivative of the actual position of the rehabilitation robot tip over time, i.e. the acceleration.
Suppose the actual position x of the end of the upper arm two-arm rehabilitation robot and the derivative of the actual position of the end of the upper arm two-arm rehabilitation robot with respect to time
Figure BDA0003336561920000125
Is obtained by measurement; let x1＝[q1,q2,…,qn]T,
Figure BDA0003336561920000126
Wherein q isiAnd
Figure BDA0003336561920000127
respectively representing the rotation angle and the angular speed of the ith joint, wherein i is more than or equal to 1 and less than or equal to n; x is the number of1Representing a position matrix formed by rotation angles of each joint of the robot; x is the number of2Representing a velocity matrix composed of angular velocities of the joints; superscript T denotes transpose; the dynamics of the interaction task is represented in the form:
Figure BDA0003336561920000128
defining a position error z1＝x1-xrVelocity error z2＝x2-α1，α1Is to z1The virtual control of (2) is as follows:
Figure BDA0003336561920000129
using Lyapunov functions
Figure BDA00033365619200001210
V1A function representing the constructed Lyapunov function form; symbol denotes matrix multiplication; the time is derived as follows:
Figure BDA00033365619200001211
order to
Figure BDA00033365619200001212
Wherein K1To gain the matrix, equation (7) is reset:
Figure BDA0003336561920000131
from equation (8):
Figure BDA0003336561920000132
defining Lyapunov functions
Figure BDA0003336561920000133
V2A function representing the constructed Lyapunov function form; the time is derived as follows:
Figure BDA0003336561920000134
when the parameters of the dynamics are known, the control is expressed in the form:
Figure BDA0003336561920000135
wherein, K2Representing a gain matrix;
approximating G, C and M terms of robot dynamics using a radial basis function neural network; the external disturbance is compensated by a disturbance observer; the self-adaptive control law is as follows:
Figure BDA0003336561920000136
wherein the content of the first and second substances,
Figure BDA0003336561920000137
the radial basis function network is a radial basis function neural network, W is a weight coefficient, Y (Z) is a dynamic regression matrix, namely the distance between a sample point and each radial basis center, and Z represents the input of the radial function network; order to
Figure BDA0003336561920000138
The high-order disturbance observer is in the form of:
Figure BDA0003336561920000139
wherein, KdRepresenting a gain matrix during disturbance observation;
Figure BDA00033365619200001310
representing an estimation error; y isd(Zd) Representing a dynamic regression matrix, Yd(Zd) Representing a dynamic regression matrix; zdRepresenting actual sampling points; wdRepresenting a weight coefficient;
Figure BDA00033365619200001311
Figure BDA00033365619200001312
the weight matrix is updated as follows:
Figure BDA00033365619200001313
Figure BDA00033365619200001314
Figure BDA00033365619200001315
Yd(Zd)Wd＝M-1(u+uh-C(x1,x2)x2-G(x1))-∈d
wherein, Yi(Z) representing an updated value of the dynamic regression quantity matrix; z is a radical of2iAn update indicative of a speed error; wiIndicating an update of the estimate; superscript symbol
Figure BDA00033365619200001316
An expected value representing a derivative of the weight; wdiRepresenting an updated value of a physical parameter; e represents an estimation error; e is the same asdRepresenting the desired estimation error; y (Z) W represents the output of the radial basis function; gamma-shapediAnd ΓdiTo update the rate, θiAnd thetadiAre weights.
The invention discloses a novel mirror image force field rehabilitation strategy for guiding the motion of the affected side based on the healthy side force field information of a patient, which is explored by the invention. Aiming at the great clinical requirement of upper limb motor function reconstruction after clinical nerve displacement, the invention combines a force field control strategy and a mirror image rehabilitation strategy which are closely coupled by a human machine.
The invention carries out new exploration on a medical rehabilitation method, the research result is a great breakthrough in the peripheral nerve rehabilitation research field, the invention not only drives the clinical treatment method in the peripheral nerve rehabilitation research field to be great innovation, provides a new technical means for exploring the functional reconstruction and brain function recovery mechanism after the nerve displacement, but also has great academic value and clinical significance.
The invention relates to the technical field of human-computer interaction, artificial intelligence and interaction control, and the method comprises the steps of human-computer tight coupling healthy lateral force field modeling based on multi-sensor signal fusion, healthy lateral physiological signals and force field mapping based on state space, and healthy lateral synchronous coupling control based on force field mirror images. The traditional method is to use a mirror to perform visual feedback lateral movement strengthening after the nerve repair operation to assist the recovery of the affected side, and the patient has weak participation feeling and general recovery effect. Different from the existing rehabilitation means. Aiming at the great clinical requirement of upper limb movement function reconstruction after clinical nerve displacement, the invention combines a force field control strategy of man-machine tight coupling with a mirror image rehabilitation strategy, explores a new mirror image force field rehabilitation strategy for guiding the motion of the affected side based on the patient side-healthy force field information, and the method is more natural and improves the participation sense and the active rehabilitation ability of the patient.
Those skilled in the art will appreciate that, in addition to implementing the system and its various devices, modules, units provided by the present invention as pure computer readable program code, the system and its various devices, modules, units provided by the present invention can be fully implemented by logically programming method steps in the form of logic gates, switches, application specific integrated circuits, programmable logic controllers, embedded microcontrollers and the like. Therefore, the system and various devices, modules and units thereof provided by the invention can be regarded as a hardware component, and the devices, modules and units included in the system for realizing various functions can also be regarded as structures in the hardware component; means, modules, units for performing the various functions may also be regarded as structures within both software modules and hardware components for performing the method.
The foregoing description of specific embodiments of the present invention has been presented. It is to be understood that the present invention is not limited to the specific embodiments described above, and that various changes or modifications may be made by one skilled in the art within the scope of the appended claims without departing from the spirit of the invention. The embodiments and features of the embodiments of the present application may be combined with each other arbitrarily without conflict.

Claims (10)

1. An admittance control method of an upper limb double-arm rehabilitation robot based on a mirror image force field is characterized by comprising the following steps:
step 1: modeling a human-computer tight coupling healthy lateral force field based on multi-sensing signal fusion to obtain the movement intention of the healthy lateral of the subject;
step 2: mapping a physiological signal and a force field of the healthy side based on a state space according to the movement intention of the healthy side of the subject to obtain a movement track and intention of the affected side of the subject;
and step 3: and performing force field mirror image-based healthy and affected side synchronous coupling control according to the motion trail and intention of the affected side of the subject, and further controlling the motion of the exoskeleton.
2. The mirror force field-based upper-arm rehabilitation robot admittance control method according to claim 1, wherein the step 1 comprises predicting the movement intention of the subject in real time through a healthy lateral electromyography sensor, modeling the acting force during the interaction as an impedance model, and predicting the joint state of the subject through the impedance model, wherein the impedance model is as shown in formula (1):
Figure FDA0003336561910000011
wherein u ishActing force of the upper limb double-arm rehabilitation robot in the interaction process with the testee; x is the actual position of the tail end of the upper limb double-arm rehabilitation robot; x is the number ofrThe expected position of the tail end of the upper limb double-arm rehabilitation robot; superscript symbol · representing the derivative of the corresponding state quantity with respect to time; l ish,1Is the position error gain; l ish,2Is the speed gain;
estimating the motor intention of the subject by equation (1), as shown in equation (2):
Figure FDA0003336561910000012
wherein the content of the first and second substances,
Figure FDA0003336561910000013
an estimated value representing the exercise intention of the healthy side of the subject, and a superscript symbol ^ represents an estimated value of the corresponding quantity;
Figure FDA0003336561910000014
an initial value representing an error gain at an arbitrary virtual target position;
Figure FDA0003336561910000015
an initial value representing the gain at any virtual target speed; the superscript v indicates that the value is given an arbitrary initial value based on the virtual target.
3. The mirror force field based upper arm rehabilitation robot admittance control method according to claim 2, wherein the step 2 comprises the steps of:
step 2.1: modeling a healthy lateral force field and a physiological electromyographic signal of the subject according to the step 1, namely obtaining the movement intention of the healthy lateral of the subject
Figure FDA0003336561910000016
Step 2.2: the two arms of the affected side of the testee do rehabilitation motions along the same track, and the motion track and intention of the affected side of the testee are obtained through the mirror image principle.
4. The mirror force field based admittance control method for an upper arm rehabilitation robot of claim 3, wherein the step 3 comprises combining the interaction force generated by the affected side during the interaction process with the established affected side motion trajectory and intention to control the motion of the exoskeleton through admittance control, wherein the affected side intention is expressed as:
Figure FDA0003336561910000021
wherein the content of the first and second substances,
Figure FDA0003336561910000022
a subject intent predicted for the robust model; tau isrThe original movement intention of the affected model; λ is a hyper-parameter that adjusts the ratio of the two weights.
5. The mirror force field-based admittance control method for an upper-arm dual-arm rehabilitation robot of claim 4, wherein the admittance control in the step 3 comprises:
the dynamic equation of the interaction process of the upper limb double-arm rehabilitation robot and the testee is shown in formula (4):
Figure FDA0003336561910000023
m and G respectively represent an inertia matrix and a gravity matrix of the upper limb exoskeleton robot and the human interaction system under a Cartesian space coordinate system; c represents a Coriolis force and centrifugal force matrix of the upper limb exoskeleton robot and the human interaction system under a Cartesian space coordinate system; f. ofdisIs a disturbance in the interactive system; u is the control input of the system; the superscript · represents the second derivative of the actual position of the rehabilitation robot tip over time, i.e. the acceleration;
suppose the actual position x of the end of the upper arm two-arm rehabilitation robot and the derivative of the actual position of the end of the upper arm two-arm rehabilitation robot with respect to time
Figure FDA0003336561910000024
Is obtained by measurement; let x1＝[q1,q2,…,qn]T,
Figure FDA0003336561910000025
Wherein q isiAnd
Figure FDA0003336561910000026
respectively representing the rotation angle and the angular speed of the ith joint, wherein i is more than or equal to 1 and less than or equal to n; x is the number of1Representing a position matrix formed by rotation angles of each joint of the robot; x is the number of2Representing a velocity matrix composed of angular velocities of the joints; superscript T denotes transpose; the dynamics of the interaction task is represented in the form:
Figure FDA0003336561910000027
defining a position error z1＝x1-xrVelocity error z2＝x2-α1，α1Is to z1The virtual control of (2) is as follows:
Figure FDA0003336561910000028
using Lyapunov functions
Figure FDA0003336561910000029
V1A function representing the constructed Lyapunov function form; symbol denotes matrix multiplication; the time is derived as follows:
Figure FDA00033365619100000210
order to
Figure FDA00033365619100000211
Wherein K1To gain the matrix, equation (7) is reset:
Figure FDA00033365619100000212
from equation (8):
Figure FDA00033365619100000213
defining Lyapunov functions
Figure FDA00033365619100000214
V2A function representing the constructed Lyapunov function form; the time is derived as follows:
Figure FDA0003336561910000031
when the parameters of the dynamics are known, the control is expressed in the form:
Figure FDA0003336561910000032
wherein, K2Representing a gain matrix;
approximating G, C and M terms of robot dynamics using a radial basis function neural network; the external disturbance is compensated by a disturbance observer; the self-adaptive control law is as follows:
Figure FDA0003336561910000033
wherein the content of the first and second substances,
Figure FDA0003336561910000034
the radial basis function network is a radial basis function neural network, W is a weight coefficient, Y (Z) is a dynamic regression matrix, namely the distance between a sample point and each radial basis center, and Z represents the input of the radial function network; order to
Figure FDA0003336561910000035
The high-order disturbance observer is in the form of:
Figure FDA0003336561910000036
wherein, KdRepresenting a gain matrix during disturbance observation;
Figure FDA0003336561910000037
representing an estimation error; y isd(Zd) Representing a dynamic regression matrix, Yd(Zd) Representing a dynamic regression matrix; zdRepresenting actual sampling points; wdRepresenting a weight coefficient;
Figure FDA0003336561910000038
Figure FDA0003336561910000039
the weight matrix is updated as follows:
Figure FDA00033365619100000310
Figure FDA00033365619100000311
Figure FDA00033365619100000312
Yd(Zd)Wd＝M-1(u+uh-C(x1,x2)x2-G(x1))-∈d
wherein, Yi(Z) representing an updated value of the dynamic regression quantity matrix; z is a radical of2iAn update indicative of a speed error; wiIndicating an update of the estimate; superscript symbol
Figure FDA00033365619100000313
An expected value representing a derivative of the weight; wdiRepresenting an updated value of a physical parameter; e represents an estimation error; e is the same asdRepresenting the desired estimation error; y (Z) W represents the output of the radial basis function; gamma-shapediAnd ΓdiTo update the rate, θiAnd thetadiAre weights.
6. The utility model provides an upper limbs both arms rehabilitation robot admittance control system based on mirror image force field which characterized in that includes following module:
module M1: modeling a human-computer tight coupling healthy lateral force field based on multi-sensing signal fusion to obtain the movement intention of the healthy lateral of the subject;
module M2: mapping a physiological signal and a force field of the healthy side based on a state space according to the movement intention of the healthy side of the subject to obtain a movement track and intention of the affected side of the subject;
module M3: and performing force field mirror image-based healthy and affected side synchronous coupling control according to the motion trail and intention of the affected side of the subject, and further controlling the motion of the exoskeleton.
7. The mirror force field based upper arm rehabilitation robot admittance control system of claim 6, wherein the module M1 comprises predicting the movement intention of the subject in real time through a side-healthy electromyography sensor, modeling the acting force during the interaction as an impedance model, and predicting the joint state of the subject through the impedance model, wherein the impedance model is as shown in formula (1):
Figure FDA0003336561910000041
wherein u ishActing force of the upper limb double-arm rehabilitation robot in the interaction process with the testee; x is the actual position of the tail end of the upper limb double-arm rehabilitation robot; x is the number ofrThe expected position of the tail end of the upper limb double-arm rehabilitation robot; superscript symbol · representing the derivative of the corresponding state quantity with respect to time; l ish,1Is the position error gain; l ish,2Is the speed gain;
estimating the motor intention of the subject by equation (1), as shown in equation (2):
Figure FDA0003336561910000042
wherein the content of the first and second substances,
Figure FDA0003336561910000043
an estimated value representing the exercise intention of the healthy side of the subject, and a superscript symbol ^ represents an estimated value of the corresponding quantity;
Figure FDA0003336561910000044
an initial value representing an error gain at an arbitrary virtual target position;
Figure FDA0003336561910000045
an initial value representing the gain at any virtual target speed; the superscript v indicates that the value is given an arbitrary initial value based on the virtual target.
8. The mirror force field based upper arm rehabilitation robot admittance control system of claim 7, wherein the module M2 comprises the following modules:
module M2.1: modeling the healthy lateral force field and the physiological electromyographic signal of the subject according to the module M1, namely obtaining the exercise intention of the healthy lateral of the subject
Figure FDA0003336561910000049
Module M2.2: the two arms of the affected side of the testee do rehabilitation motions along the same track, and the motion track and intention of the affected side of the testee are obtained through the mirror image principle.
9. The mirror force field based admittance control system of an upper arm rehabilitation robot of claim 8, wherein the module M3 comprises an admittance control unit for combining the interaction force generated by the affected side during the interaction process with the established affected side motion trajectory and intention to control the motion of the exoskeleton, wherein the affected side intention is expressed as:
Figure FDA0003336561910000046
wherein the content of the first and second substances,
Figure FDA0003336561910000047
a subject intent predicted for the robust model; tau isrThe original movement intention of the affected model; lambda is a super parameter for adjusting the weight ratio of the twoAnd (4) counting.
10. The mirror force field based upper arm rehabilitation robot admittance control system of claim 9, wherein the admittance control in module M3 comprises:
the dynamic equation of the interaction process of the upper limb double-arm rehabilitation robot and the testee is shown in formula (4):
Figure FDA0003336561910000048
m and G respectively represent an inertia matrix and a gravity matrix of the upper limb exoskeleton robot and the human interaction system under a Cartesian space coordinate system; c represents a Coriolis force and centrifugal force matrix of the upper limb exoskeleton robot and the human interaction system under a Cartesian space coordinate system; f. ofdisIs a disturbance in the interactive system; u is the control input of the system; the superscript · represents the second derivative of the actual position of the rehabilitation robot tip over time, i.e. the acceleration;
suppose the actual position x of the end of the upper arm two-arm rehabilitation robot and the derivative of the actual position of the end of the upper arm two-arm rehabilitation robot with respect to time
Figure FDA0003336561910000051
Is obtained by measurement; let x1＝[q1,q2,…,qn]T,
Figure FDA0003336561910000052
Wherein q isiAnd
Figure FDA0003336561910000053
respectively representing the rotation angle and the angular speed of the ith joint, wherein i is more than or equal to 1 and less than or equal to n; x is the number of1Representing a position matrix formed by rotation angles of each joint of the robot; x is the number of2Representing a velocity matrix composed of angular velocities of the joints; superscript T denotes transpose; the dynamics of the interaction task is represented in the form:
Figure FDA0003336561910000054
defining a position error z1＝x1-xrVelocity error z2＝x2-α1，α1Is to z1The virtual control of (2) is as follows:
Figure FDA0003336561910000055
using Lyapunov functions
Figure FDA0003336561910000056
V1A function representing the constructed Lyapunov function form; symbol denotes matrix multiplication; the time is derived as follows:
Figure FDA0003336561910000057
order to
Figure FDA0003336561910000058
Wherein K1To gain the matrix, equation (7) is reset:
Figure FDA0003336561910000059
from equation (8):
Figure FDA00033365619100000510
defining Lyapunov functions
Figure FDA00033365619100000511
V2A function representing the constructed Lyapunov function form; the time is derived as follows:
Figure FDA00033365619100000512
when the parameters of the dynamics are known, the control is expressed in the form:
Figure FDA00033365619100000513
wherein, K2Representing a gain matrix;
approximating G, C and M terms of robot dynamics using a radial basis function neural network; the external disturbance is compensated by a disturbance observer; the self-adaptive control law is as follows:
Figure FDA00033365619100000514
wherein the content of the first and second substances,
Figure FDA00033365619100000515
the radial basis function network is a radial basis function neural network, W is a weight coefficient, Y (Z) is a dynamic regression matrix, namely the distance between a sample point and each radial basis center, and Z represents the input of the radial function network; order to
Figure FDA00033365619100000516
The high-order disturbance observer is in the form of:
Figure FDA0003336561910000061
wherein, KdRepresenting a gain matrix during disturbance observation;
Figure FDA0003336561910000062
representing an estimation error; y isd(Zd) Representing a dynamic regression matrix, Yd(Zd) Representing a dynamic regression matrix; zdRepresenting actual sampling points; wdRepresenting a weight coefficient;
Figure FDA0003336561910000063
Figure FDA0003336561910000064
the weight matrix is updated as follows:
Figure FDA0003336561910000065
Figure FDA0003336561910000066
Figure FDA0003336561910000067
Yd(Zd)Wd＝M-1(u+uh-C(x1,x2)x2-G(x1))-∈d
wherein, Yi(Z) representing an updated value of the dynamic regression quantity matrix; z is a radical of2iAn update indicative of a speed error; wiIndicating an update of the estimate; superscript symbol
Figure FDA0003336561910000068
An expected value representing a derivative of the weight; wdiRepresenting an updated value of a physical parameter; e represents an estimation error; e is the same asdRepresenting the desired estimation error; y (Z) W represents the output of the radial basis function; gamma-shapediAnd ΓdiTo update the rate, θiAnd thetadiAre weights.
CN202111295849.2A 2021-11-03 2021-11-03 Mirror image force field-based upper limb double-arm rehabilitation robot admittance control method and system Active CN113995629B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202111295849.2A CN113995629B (en) 2021-11-03 2021-11-03 Mirror image force field-based upper limb double-arm rehabilitation robot admittance control method and system

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202111295849.2A CN113995629B (en) 2021-11-03 2021-11-03 Mirror image force field-based upper limb double-arm rehabilitation robot admittance control method and system

Publications (2)

Publication Number Publication Date
CN113995629A true CN113995629A (en) 2022-02-01
CN113995629B CN113995629B (en) 2023-07-11

Family

ID=79926998

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202111295849.2A Active CN113995629B (en) 2021-11-03 2021-11-03 Mirror image force field-based upper limb double-arm rehabilitation robot admittance control method and system

Country Status (1)

Country Link
CN (1) CN113995629B (en)

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN116138909A (en) * 2023-04-24 2023-05-23 北京市春立正达医疗器械股份有限公司 Intelligent control method and system for dental implant robot

Citations (13)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN104586608A (en) * 2015-02-05 2015-05-06 华南理工大学 Wearable assistance finger based on myoelectric control and control method thereof
CN104881038A (en) * 2015-04-22 2015-09-02 哈尔滨工业大学 Unmanned underwater vehicle (UUV) track tracking control optimization method under environmental interference
CN107397649A (en) * 2017-08-10 2017-11-28 燕山大学 A kind of upper limbs exoskeleton rehabilitation robot control method based on radial base neural net
CN108324503A (en) * 2018-03-16 2018-07-27 燕山大学 Healing robot self-adaptation control method based on flesh bone model and impedance control
CN108785997A (en) * 2018-05-30 2018-11-13 燕山大学 A kind of lower limb rehabilitation robot Shared control method based on change admittance
US20190060155A1 (en) * 2017-08-22 2019-02-28 New Jersey Institute Of Technology Exoskeleton with admittance control
CN111281743A (en) * 2020-02-29 2020-06-16 西北工业大学 Self-adaptive flexible control method for exoskeleton robot for upper limb rehabilitation
CN111631923A (en) * 2020-06-02 2020-09-08 中国科学技术大学先进技术研究院 Neural network control system of exoskeleton robot based on intention recognition
EP3705105A1 (en) * 2019-03-08 2020-09-09 Syco di Menga Giuseppe & C. S.A.S. Control system for a haptic lower limb exoskeleton for rehabilitation or walking, with improved equilibrium control, man-machine interface
US20210045640A1 (en) * 2019-08-16 2021-02-18 Poltorak Technologies, LLC Device and method for medical diagnostics
CN112743540A (en) * 2020-12-09 2021-05-04 华南理工大学 Hexapod robot impedance control method based on reinforcement learning
US20210141395A1 (en) * 2017-08-02 2021-05-13 South China University Of Technology Stable flight control method for multi-rotor unmanned aerial vehicle based on finite-time neurodynamics
CN113478462A (en) * 2021-07-08 2021-10-08 中国科学技术大学 Method and system for controlling intention assimilation of upper limb exoskeleton robot based on surface electromyogram signal

Patent Citations (13)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN104586608A (en) * 2015-02-05 2015-05-06 华南理工大学 Wearable assistance finger based on myoelectric control and control method thereof
CN104881038A (en) * 2015-04-22 2015-09-02 哈尔滨工业大学 Unmanned underwater vehicle (UUV) track tracking control optimization method under environmental interference
US20210141395A1 (en) * 2017-08-02 2021-05-13 South China University Of Technology Stable flight control method for multi-rotor unmanned aerial vehicle based on finite-time neurodynamics
CN107397649A (en) * 2017-08-10 2017-11-28 燕山大学 A kind of upper limbs exoskeleton rehabilitation robot control method based on radial base neural net
US20190060155A1 (en) * 2017-08-22 2019-02-28 New Jersey Institute Of Technology Exoskeleton with admittance control
CN108324503A (en) * 2018-03-16 2018-07-27 燕山大学 Healing robot self-adaptation control method based on flesh bone model and impedance control
CN108785997A (en) * 2018-05-30 2018-11-13 燕山大学 A kind of lower limb rehabilitation robot Shared control method based on change admittance
EP3705105A1 (en) * 2019-03-08 2020-09-09 Syco di Menga Giuseppe & C. S.A.S. Control system for a haptic lower limb exoskeleton for rehabilitation or walking, with improved equilibrium control, man-machine interface
US20210045640A1 (en) * 2019-08-16 2021-02-18 Poltorak Technologies, LLC Device and method for medical diagnostics
CN111281743A (en) * 2020-02-29 2020-06-16 西北工业大学 Self-adaptive flexible control method for exoskeleton robot for upper limb rehabilitation
CN111631923A (en) * 2020-06-02 2020-09-08 中国科学技术大学先进技术研究院 Neural network control system of exoskeleton robot based on intention recognition
CN112743540A (en) * 2020-12-09 2021-05-04 华南理工大学 Hexapod robot impedance control method based on reinforcement learning
CN113478462A (en) * 2021-07-08 2021-10-08 中国科学技术大学 Method and system for controlling intention assimilation of upper limb exoskeleton robot based on surface electromyogram signal

Non-Patent Citations (2)

* Cited by examiner, † Cited by third party
Title
刘健: "《基于表面肌电信号的上肢外骨骼控制研究》", 中国优秀硕士学位论文全文数据库信息科技辑, no. 02 *
李智军，康宇等: "《面向人机智能融合医疗康复的辅助机器人关键技术及其应用》", 中国知网 *

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN116138909A (en) * 2023-04-24 2023-05-23 北京市春立正达医疗器械股份有限公司 Intelligent control method and system for dental implant robot
CN116138909B (en) * 2023-04-24 2023-10-27 北京市春立正达医疗器械股份有限公司 Intelligent control method and system for dental implant robot

Also Published As

Publication number Publication date
CN113995629B (en) 2023-07-11

Similar Documents

Publication Publication Date Title
US10959863B2 (en) Multi-dimensional surface electromyogram signal prosthetic hand control method based on principal component analysis
CN109549821B (en) Exoskeleton robot power-assisted control system and method based on myoelectricity and inertial navigation signal fusion
WO2021068542A1 (en) Force feedback technology-based robot system for active and passive rehabilitation training of upper limbs
CN111956452B (en) Control method and device for upper limb rehabilitation robot
CN106074089B (en) A kind of both hands tracking mode finger gymnastic robot system
CN105963100A (en) Patient movement demand-based assistance lower limb rehabilitation robot self-adaptation control method
Bai et al. A novel human-robot cooperative method for upper extremity rehabilitation
Esfahlani et al. An adaptive self-organizing fuzzy logic controller in a serious game for motor impairment rehabilitation
CN109481226A (en) A kind of both hands tracking mode multiple degrees of freedom software finger gymnastic robot and application method
CN115363907A (en) Rehabilitation decision-making method based on virtual reality rehabilitation training system
KOIKE et al. Estimation of arm posture in 3D-space from surface EMG signals using a neural network model
CN113995629A (en) Upper limb double-arm rehabilitation robot admittance control method and system based on mirror force field
CN109692104A (en) Medical rehabilitation ectoskeleton interactive mode synchronous control system and method
Jones et al. Development of a massage robot for medical therapy
Peng et al. Experimental study of robot-assisted exercise training for knee rehabilitation based on a practical EMG-driven model
CN113262088B (en) Multi-degree-of-freedom hybrid control artificial hand with force feedback and control method
Zhu et al. Underactuated rehabilitation robotics for hand function
CN117697717A (en) Exoskeleton physical man-machine two-way interaction simulation system
CN112951360B (en) Bilateral finger cooperative motion control method and system
CN113925742B (en) Control method and control system of target-driven upper limb exoskeleton rehabilitation robot
CN115416003A (en) On-demand auxiliary control method for lower limb exoskeleton of old people
CN114795604A (en) Lower limb prosthesis coordination control method and system based on non-zero sum game
Ahmadian et al. ℒ 1–ℬℒ Adaptive Controller Design for Wrist Rehabilitation Robot
Zhang et al. Design on subsection based mix position controller for lower limb rehabilitation robot
Halder et al. An Overview of Artificial Intelligence-based Soft Upper Limb Exoskeleton for Rehabilitation: A Descriptive Review

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination
GR01 Patent grant
GR01 Patent grant