CN114750137A - RBF network-based upper limb exoskeleton robot motion control method - Google Patents

RBF network-based upper limb exoskeleton robot motion control method Download PDF

Info

Publication number
CN114750137A
CN114750137A CN202210557091.3A CN202210557091A CN114750137A CN 114750137 A CN114750137 A CN 114750137A CN 202210557091 A CN202210557091 A CN 202210557091A CN 114750137 A CN114750137 A CN 114750137A
Authority
CN
China
Prior art keywords
exoskeleton robot
upper limb
joint
limb exoskeleton
robot
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.)
Pending
Application number
CN202210557091.3A
Other languages
Chinese (zh)
Inventor
唐昊
韩帅
王舒润
王彬
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.)
Hefei University of Technology
Original Assignee
Hefei University of Technology
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 Hefei University of Technology filed Critical Hefei University of Technology
Priority to CN202210557091.3A priority Critical patent/CN114750137A/en
Publication of CN114750137A publication Critical patent/CN114750137A/en
Pending legal-status Critical Current

Links

Images

Classifications

    • B—PERFORMING OPERATIONS; TRANSPORTING
    • B25—HAND TOOLS; PORTABLE POWER-DRIVEN TOOLS; MANIPULATORS
    • B25J—MANIPULATORS; CHAMBERS PROVIDED WITH MANIPULATION DEVICES
    • B25J9/00—Programme-controlled manipulators
    • B25J9/0006—Exoskeletons, i.e. resembling a human figure
    • B—PERFORMING OPERATIONS; TRANSPORTING
    • B25—HAND TOOLS; PORTABLE POWER-DRIVEN TOOLS; MANIPULATORS
    • B25J—MANIPULATORS; CHAMBERS PROVIDED WITH MANIPULATION DEVICES
    • B25J9/00—Programme-controlled manipulators
    • B25J9/16—Programme controls
    • B25J9/1602—Programme controls characterised by the control system, structure, architecture
    • B25J9/161—Hardware, e.g. neural networks, fuzzy logic, interfaces, processor
    • B—PERFORMING OPERATIONS; TRANSPORTING
    • B25—HAND TOOLS; PORTABLE POWER-DRIVEN TOOLS; MANIPULATORS
    • B25J—MANIPULATORS; CHAMBERS PROVIDED WITH MANIPULATION DEVICES
    • B25J9/00—Programme-controlled manipulators
    • B25J9/16—Programme controls
    • B25J9/1628—Programme controls characterised by the control loop
    • B25J9/163—Programme controls characterised by the control loop learning, adaptive, model based, rule based expert control
    • B—PERFORMING OPERATIONS; TRANSPORTING
    • B25—HAND TOOLS; PORTABLE POWER-DRIVEN TOOLS; MANIPULATORS
    • B25J—MANIPULATORS; CHAMBERS PROVIDED WITH MANIPULATION DEVICES
    • B25J9/00—Programme-controlled manipulators
    • B25J9/16—Programme controls
    • B25J9/1628—Programme controls characterised by the control loop
    • B25J9/1633—Programme controls characterised by the control loop compliant, force, torque control, e.g. combined with position control
    • B—PERFORMING OPERATIONS; TRANSPORTING
    • B25—HAND TOOLS; PORTABLE POWER-DRIVEN TOOLS; MANIPULATORS
    • B25J—MANIPULATORS; CHAMBERS PROVIDED WITH MANIPULATION DEVICES
    • B25J9/00—Programme-controlled manipulators
    • B25J9/16—Programme controls
    • B25J9/1656—Programme controls characterised by programming, planning systems for manipulators
    • B25J9/1664—Programme controls characterised by programming, planning systems for manipulators characterised by motion, path, trajectory planning

Landscapes

  • Engineering & Computer Science (AREA)
  • Robotics (AREA)
  • Mechanical Engineering (AREA)
  • Automation & Control Theory (AREA)
  • Physics & Mathematics (AREA)
  • Artificial Intelligence (AREA)
  • Evolutionary Computation (AREA)
  • Fuzzy Systems (AREA)
  • Mathematical Physics (AREA)
  • Software Systems (AREA)
  • Manipulator (AREA)

Abstract

The invention belongs to the technical field of mechanical control, and particularly relates to an upper limb exoskeleton robot motion control method based on an RBF neural network. The method comprises the following steps: s1, constructing a mechanical structure of the upper limb exoskeleton robot; s2, acquiring the expected angle change of each joint of the robot and the expected angular speed change of each joint; s3, the data of step S2 is input to the PI controller, and the actual angle change of each joint and the magnitude of the torque for driving each joint are output. According to the invention, a plurality of RBF neural networks are applied to the design of the controller, so that the approaching speed and precision of the neural networks to the uncertain items of the upper limb exoskeleton dynamic model can be improved when the control problem with the uncertain model is solved, the steady-state error of track tracking is reduced, and the motion control performance of the upper limb exoskeleton robot is effectively improved and optimized.

Description

RBF network-based upper limb exoskeleton robot motion control method
Technical Field
The invention belongs to the technical field of mechanical control, and particularly relates to an upper limb exoskeleton robot motion control method based on an RBF neural network.
Background
Progress in human society is closely related to development of production tools, and the appearance of robots has greatly accelerated development of human productivity. Nowadays, robots are widely applied to industrial production, so that the industrial production efficiency is improved, and the labor cost in the industrial production is saved. The progress of science and technology continuously pushes the robot to develop towards intellectualization. Modern robots already have certain intelligence and can make corresponding responses under the influence of external environment. However, there is still an insurmountable gap between the level of machine intelligence and that of human intelligence. However, developing machine intelligence is an arbitrary and remote process, and in order to rapidly solve the problem that the robot intelligence is more and more required by more and more complex robot application environments, many researchers have proposed the idea of combining human "intelligence" with "physical strength" of the robot. In the background of this era, exoskeleton robots have come into force.
The upper limb exoskeleton robot motion system has the characteristics of nonlinearity, strong coupling, time-varying property, uncertainty and the like, so the upper limb exoskeleton robot system is a very complex multiple-input multiple-output (MIMO) nonlinear system. The control of the robot is mainly the position control of each joint or end effector, so that each joint or end effector can track a given trajectory with a desired dynamic quality or be stabilized at a specified position, i.e. the task of the designed control system is trajectory tracking control. In actual engineering, because physical quantities such as the load weight, the mass of the connecting rod, the length and the position of the mass center of the connecting rod of the robot are unknown, changed or only partial information is known, the change of the designed controller quality is possibly caused, and even the integral instability of a control system is caused. Therefore, the control problem of the nonlinear time-varying strong coupling MIMO robot is always a difficult problem in the control field.
Disclosure of Invention
In order to solve the problem of model uncertainty of the upper limb exoskeleton robot, the invention provides an upper limb exoskeleton robot motion control method based on RBF networks, so that uncertainty items existing in a robot dynamics model during modeling are approximated by using a plurality of RBF neural networks, the response speed of each joint of the robot is shortened, and the motion control performance of the upper limb exoskeleton robot is improved.
In order to achieve the purpose, the invention adopts the following technical scheme:
an upper limb exoskeleton robot motion control method based on an RBF network comprises the following steps:
s1, constructing a mechanical structure of the upper limb exoskeleton robot;
s2, acquiring the expected angle change of each joint of the robot and the expected angular speed change of each joint;
s3, the data of step S2 is input to the PI controller, and the actual angle change of each joint and the magnitude of the torque for driving each joint are output.
According to the further optimization of the technical scheme, the upper limb exoskeleton robot in the step S1 consists of a left mechanical arm, a right mechanical arm and a back base, wherein each mechanical arm comprises three rotary joints, and the two mechanical arms are identical in structure and symmetrical left and right; wherein, the No. 1 joint close to the back base at the shoulder is a passive joint which is not driven by a motor, and the No. 2 joint and the No. 3 joint are active joints.
In a further optimization of the present technical solution, the designing of the PI controller in step S3 includes the following steps:
s3.1, constructing an upper limb exoskeleton robot dynamic model;
s3.2, improving the upper limb exoskeleton robot dynamics model by using the RBF network;
and S3.3, designing a motion controller of the upper limb exoskeleton robot according to the improved dynamic model.
In a further optimization of the technical scheme, the step S3.1 of constructing the upper limb exoskeleton robot dynamics model is as follows:
the Lagrange dynamics method is used for establishing a dynamics equation of the upper limb exoskeleton robot as follows:
Figure BDA0003652621110000021
wherein D (q) is a positive definite inertia matrix of the upper limb exoskeleton robot,
Figure BDA0003652621110000022
is a matrix of centrifugal force and Copenese force, G (q) is a gravity matrix, q is an angle matrix when the upper limb exoskeleton robot joint moves,
Figure BDA0003652621110000023
is a matrix of the angular velocity and,
Figure BDA0003652621110000024
is an angular acceleration matrix and tau is a matrix of driving forces acting on the joints of the robot.
In a further optimization of the present technical solution, step S3.2 includes:
the input layer x (t) is the input variable of the neural network at the time t;
the hidden layer is formed by a non-linear excitation function hj(t) constitution:
Figure BDA0003652621110000025
wherein j is the node of the hidden layer, m is the number of nodes of the hidden layer, cjAs a central vector of the neural network, bjIs the width of the base band of the gaussian base function;
output layer yj(t) is expressed as:
Figure BDA0003652621110000031
wherein n is the number of output nodes, wjiRepresenting the connection weight of the output layer;
three RBF networks are utilized to respectively realize the pair D (q),
Figure BDA00036526211100000313
And G (q), the outputs of the three networks are respectively DNN(q)、
Figure BDA00036526211100000312
GNN(q)：
Figure BDA0003652621110000032
Figure BDA0003652621110000033
In the formula, ED、ECAnd EGRespectively a neural network pair D (q),
Figure BDA00036526211100000314
And modeling error of G (q), WD、WCAnd WGWeight, xi, for modelling neural networksD(q)、ΞC(z) and xiG(q) is the output of the hidden layer gaussian function,
Figure BDA0003652621110000034
definition of e (t) ═ qd(t)-q(t)、
Figure BDA0003652621110000035
e (t) is the angle tracking error, qd(t) is the desired angle of the robot joint, q (t) is the actual angle, Λ>And 0, substituting the initially established kinetic model (1) to obtain an improved kinetic model which is as follows:
Figure BDA0003652621110000036
in a further optimization of the present technical solution, the controller in step S3.3 is designed as follows:
the controller based on model estimation is designed as:
Figure BDA0003652621110000037
in the formula (I), the compound is shown in the specification,
Figure BDA0003652621110000038
respectively as a neural network modeling term DNN(q)、
Figure BDA0003652621110000039
GNN(q) the estimation of the (q),
Figure BDA00036526211100000310
are respectively WD、WC、WGThe estimated weight of (2);
the robust terms added to overcome the approximation error of the neural network are:
τr＝Krsgn(r) (8)
in the formula, KrSgn (r) is a sign function, is a constant greater than 0;
the master controller designed by using PI control is as follows:
Figure BDA00036526211100000311
in the formula, KpAs a proportionality coefficient in PI control, KiIs an integral coefficient in PI control;
in order to realize the self-adaptive approximation of the neural network, the weight self-adaptation laws of the three networks are respectively designed as follows:
Figure BDA0003652621110000041
in the formula, gammaD，ΓD，ΓGAnd updating step length matrixes for the weights, wherein the step length matrixes are symmetrical positive definite matrixes.
Different from the prior art, the beneficial effects of the technical scheme are that:
the invention provides a method for performing polynomial approximation on an uncertain item of a dynamic model by using a plurality of RBF neural networks, aiming at the problem that the dynamic model of an upper limb exoskeleton robot has uncertainty. Compared with the traditional control method, the neural network can approximate any nonlinear function, and has online parameter adjustment and self-learning capabilities for any function, so that the neural network control can be widely applied when the control problem with an uncertain model is solved, and the method has great significance for improving the motion control performance of the upper limb exoskeleton robot. Compared with a control method based on RBF network integral approximation, the method has the advantages that the response speed of each joint of the upper limb exoskeleton robot is greatly reduced, the approximation speed and precision of the neural network to uncertain items of the upper limb exoskeleton dynamic model are improved, the maximum steady-state error and the average steady-state error of trajectory tracking are reduced, and a good control effect is achieved.
Drawings
FIG. 1 is a model diagram of a mechanical structure of an upper limb exoskeleton robot;
FIG. 2 is a block diagram of a control method of the present invention;
FIG. 3 is a schematic view of a control system according to the present invention;
fig. 4 is a graph of angle tracking of the upper limb exoskeleton robot joint 1;
fig. 5 is a graph of angle tracking of the upper limb exoskeleton robot joint 2;
fig. 6 is a graph of tracking error of the upper limb exoskeleton robot joint 1;
fig. 7 is a graph of tracking error of the upper limb exoskeleton robot joint 2.
Detailed Description
To explain technical contents, structural features, and objects and effects of the technical solutions in detail, the following detailed description is given with reference to the accompanying drawings in conjunction with the embodiments.
Referring to fig. 1, a model diagram of a mechanical structure of an upper limb exoskeleton robot is shown, the upper limb exoskeleton robot is composed of a left mechanical arm, a right mechanical arm and a back base, each mechanical arm comprises three rotary joints, and the two mechanical arms have the same structure and are symmetrical left and right. The model is a mechanical structure model of the three-degree-of-freedom upper limb exoskeleton robot built by utilizing Solidworks software, wherein a No. 1 joint close to a back base at a shoulder is a passive joint, the joint is not driven by a motor, and the design aim is to achieve adjustment of the planar posture of the robot and realize displacement of the whole mechanical arm in the horizontal direction. Joints No. 2 and No. 3 are active joints.
Constructing a dynamic model of the upper limb exoskeleton robot:
the dynamic equation of the upper limb exoskeleton robot established by the Lagrange dynamics method is as follows:
Figure BDA0003652621110000051
wherein D (q) is a positive definite inertia matrix of the upper limb exoskeleton robot,
Figure BDA0003652621110000052
is a matrix of centrifugal force and Copenese force, G (q) is a gravity matrix, q is an angle matrix when the upper limb exoskeleton robot joint moves,
Figure BDA0003652621110000053
in the form of a matrix of angular velocities,
Figure BDA0003652621110000054
is an angular acceleration matrix and tau is a matrix of driving forces acting on the joints of the robot.
Because the dynamic model has modeling errors, the model has uncertainty, and therefore, the inventor designs an RBF network to approximate the uncertainty direction. The RBF network is utilized to improve the upper limb exoskeleton robot dynamics model:
the input layer x (t) is the input variable of the neural network at the time t;
the hidden layer is formed by a non-linear excitation function hj(t) constitution:
Figure BDA0003652621110000055
wherein j is the node of the hidden layer, m is the node number of the hidden layer, cjAs a central vector of the neural network, bjIs the base band width of the gaussian basis function.
Output layer yj(t) is expressed as:
Figure BDA0003652621110000056
wherein n is the number of output nodes, wjiRepresenting the connection weights of the output layer.
Because the dynamic equation comprises a plurality of uncertain parameter items, the problems of too long approximation time and insufficient approximation precision can be caused when the RBF network is used for carrying out integral approximation on the dynamic equation. Fig. 2 is a block diagram of a control method. The invention respectively realizes the pair D (q) and the pair D (q) by utilizing three RBF networks,
Figure BDA0003652621110000057
And G (q), the outputs of the three networks are respectively DNN(q)、
Figure BDA0003652621110000058
GNN(q)；
Figure BDA0003652621110000059
Figure BDA0003652621110000061
In the formula, ED、ECAnd EGRespectively a neural network pair D (q),
Figure BDA0003652621110000062
And modeling error of G (q), WD、WCAnd WGWeights for modelling neural networks, xiD(q)、ΞC(z) and xiG(q) is the output of the hidden layer Gaussian function, neural network input
Figure BDA0003652621110000063
Figure BDA0003652621110000064
Defining a functional relation with respect to error
Figure BDA0003652621110000065
Wherein e (t) qd(t) -q (t), e (t) is angle tracking error, qd(t) desired angle of robot joint, q (t) actual angle, Λ>0 is a constant coefficient. Substituting the initially established kinetic model (1) to obtain an improved kinetic model which is as follows:
Figure BDA0003652621110000066
wherein E ═ ED+EC+EG。
Designing an upper limb exoskeleton robot motion controller according to the improved dynamic model:
the controller based on model estimation is designed as:
Figure BDA0003652621110000067
in the formula (I), the compound is shown in the specification,
Figure BDA0003652621110000068
respectively as a neural network modeling term DNN(q)、
Figure BDA0003652621110000069
GNN(q) the estimation of the value of (q),
Figure BDA00036526211100000610
are each WD、WC、WGThe estimated weight value of (2).
The robust terms added to overcome the approximation error of the neural network are:
τr＝Krsgn(r) (8)
in the formula, KrSgn (r) is a sign function, with a constant greater than 0.
The general controller designed by using PI control is as follows:
Figure BDA00036526211100000611
in the formula, KpAs a proportionality coefficient in PI control, KiIs an integration coefficient in the PI control.
In order to realize the self-adaptive approximation of the neural network, the weight self-adaptation laws of the three networks are respectively designed as follows:
Figure BDA00036526211100000612
in the formula, gammaD，ΓC，ΓGAnd updating step length matrixes for the weights, wherein the step length matrixes are symmetrical positive definite matrixes.
And controlling the upper limb exoskeleton robot by using the controller to realize accurate tracking of the expected track. A control system is built by using the designed master controller, and the upper limb exoskeleton robot is subjected to track tracking control research. The input of the system is the expected angle change of each joint of the robot and the expected angular speed change of each joint, and the output is the actual angle change of each joint and the torque magnitude for driving each joint to move. The design effect of the control method can be verified by analyzing the data output by the system.
Examples
The process of the invention is illustrated below by means of a specific example. The simulation control system constructed by the control method of the invention is shown in fig. 3.
The inputs to the setup system are: the expected angle change of two active joints of the robot is qd＝[0.5 sin(πt) sin(πt)]The desired angular velocity is
Figure BDA0003652621110000071
For approximation
Figure BDA0003652621110000072
Figure BDA0003652621110000073
The gaussian based parameter settings are as follows: central vector c _ D of neural network of inertia matrixiIs set to [ -1.2-0.600.61.2]Central vector C _ C of neural network of centripetal force and Coriolis force moment matrixiIs set to [ -1.0-0.500.51.0]Gravity matrix neural network center vector c _ GiIs set to [ -1.0-0.500.51.0]. The controller parameters are as follows: kr＝0.1,Λ＝diag{5,5}。
In the simulation experiment, an RBF control method of integral approximation of an uncertain item is adopted as a comparison item. The angle tracking curves obtained for each joint are shown in fig. 4 and 5. The angle tracking error curves for each joint are shown in fig. 6 and 7.
By analyzing the simulation result, the control method can greatly shorten the response time of the control system of the upper limb exoskeleton robot with three degrees of freedom, the average steady-state error and the maximum steady-state error of the system are reduced to different degrees, the track tracking control effect is good, and the simulation result verifies the effectiveness of the control method.
In summary, the invention relates to a motion control method of an upper limb exoskeleton robot based on an RBF network. The method adopts a plurality of RBF networks to approach uncertain items existing in a dynamics model of the upper limb exoskeleton robot, adds a robust item in a controller to overcome approximation errors of the neural network, and designs a weight adaptive law to realize adaptive approximation of the neural network. The invention effectively improves the approaching speed and precision of the neural network to the dynamics model uncertainty, reduces the steady-state error of the track tracking, and effectively improves and optimizes the motion control performance of the upper limb exoskeleton robot.
It is noted that, herein, relational terms such as first and second, and the like may be used solely to distinguish one entity or action from another entity or action without necessarily requiring or implying any actual such relationship or order between such entities or actions. Also, the terms "comprises," "comprising," or any other variation thereof, are intended to cover a non-exclusive inclusion, such that a process, method, article, or terminal that comprises a list of elements does not include only those elements but may include other elements not expressly listed or inherent to such process, method, article, or terminal. Without further limitation, an element defined by the phrases "comprising … …" or "comprising … …" does not exclude the presence of additional elements in a process, method, article, or terminal that comprises the element. Further, herein, "greater than," "less than," "more than," and the like are understood to exclude the present numbers; the terms "above", "below", "within" and the like are to be understood as including the present number.
Although the embodiments have been described, once the basic inventive concept is obtained, other variations and modifications of these embodiments can be made by those skilled in the art, so that the above embodiments are only examples of the present invention, and not intended to limit the scope of the present invention, and all equivalent structures or equivalent processes using the contents of the present specification and drawings, or any other related technical fields, which are directly or indirectly applied thereto, are included in the scope of the present invention.

Claims (6)

1. An upper limb exoskeleton robot motion control method based on an RBF network is characterized by comprising the following steps:
s1, constructing a mechanical structure of the upper limb exoskeleton robot;
s2, acquiring the expected angle change of each joint of the robot and the expected angular speed change of each joint;
s3, the data of step S2 is input to the PI controller, and the actual angle change of each joint and the magnitude of the torque for driving each joint are output.
2. An RBF network-based upper limb exoskeleton robot motion control method as claimed in claim 1, wherein said upper limb exoskeleton robot in step S1 is composed of two left and right robotic arms and a back base, each robotic arm comprises three rotational joints, and the two robotic arms are identical in structure and symmetrical left and right; wherein, the No. 1 joint close to the back base at the shoulder is a passive joint which is not driven by a motor, and the No. 2 joint and the No. 3 joint are active joints.
3. An RBF network-based upper extremity exoskeleton robot motion control method according to claim 1, wherein the design of PI controller in step S3 comprises the following steps:
s3.1, constructing an upper limb exoskeleton robot dynamic model;
s3.2, improving the upper limb exoskeleton robot dynamics model by using the RBF network;
and S3.3, designing a motion controller of the upper limb exoskeleton robot according to the improved dynamic model.
4. The RBF network-based upper extremity exoskeleton robot motion control method according to claim 3, wherein said step S3.1 is to construct an upper extremity exoskeleton robot dynamics model as follows:
the dynamic equation of the upper limb exoskeleton robot established by the Lagrange dynamics method is as follows:
Figure FDA0003652621100000011
wherein D (q) is a positive definite inertia matrix of the upper limb exoskeleton robot,
Figure FDA0003652621100000012
is a matrix of centrifugal force and Copenese force, G (q) is a gravity matrix, q is an angle matrix when the upper limb exoskeleton robot joint moves,
Figure FDA0003652621100000013
in the form of a matrix of angular velocities,
Figure FDA0003652621100000014
is angular accelerationAnd the matrix tau is a driving force matrix acting on the robot joint.
5. An RBF network based upper extremity exoskeleton robot motion control method as claimed in claim 4 wherein said step S3.2 comprises:
the input layer x (t) is the input variable of the neural network at the time t;
the hidden layer is formed by a non-linear excitation function hj(t) constitution:
Figure FDA0003652621100000021
wherein j is the node of the hidden layer, m is the number of nodes of the hidden layer, cjAs a central vector of the neural network, bjIs the width of the base band of the gaussian base function;
output layer yj(t) is expressed as:
Figure FDA0003652621100000022
wherein n is the number of output nodes, wjiRepresenting the connection weight of the output layer;
three RBF networks are utilized to respectively realize the pair D (q),
Figure FDA0003652621100000023
And G (q), the outputs of the three networks are respectively DNN(q)、
Figure FDA0003652621100000024
GNN(q)：
Figure FDA0003652621100000025
Figure FDA0003652621100000026
In the formula, ED、ECAnd EGRespectively a neural network pair D (q),
Figure FDA0003652621100000027
And modeling error of G (q), WD、WCAnd WGWeights for modelling neural networks, xiD(q)、ΞC(z) and xiG(q) is the output of the hidden layer gaussian function,
Figure FDA0003652621100000028
definition e (t) qd(t)-q(t)、
Figure FDA0003652621100000029
e (t) is the angle tracking error, qd(t) is the desired angle of the robot joint, q (t) is the actual angle, Λ>And 0, substituting the initially established kinetic model (1) to obtain an improved kinetic model which is as follows:
Figure FDA00036526211000000210
6. an RBF network based upper extremity exoskeleton robot motion control method according to claim 5 wherein in step S3.3 the controller is designed as follows:
the controller based on model estimation is designed as:
Figure FDA00036526211000000211
in the formula (I), the compound is shown in the specification,
Figure FDA00036526211000000212
respectively as a neural network modeling term DNN(q)、
Figure FDA00036526211000000213
GNN(q) the estimation of the value of (q),
Figure FDA00036526211000000214
are respectively WD、WC、WGThe estimated weight of (2);
the robust terms added to overcome the approximation error of the neural network are:
τr＝Krsgn(r) (8)
in the formula, KrSgn (r) is a sign function, is a constant greater than 0;
the master controller designed by using PI control is as follows:
Figure FDA0003652621100000031
in the formula, KpAs a proportionality coefficient in PI control, KiIs an integral coefficient in PI control;
in order to realize the self-adaptive approximation of the neural network, the weight self-adaptation laws of the three networks are respectively designed as follows:
Figure FDA0003652621100000032
in the formula, gammaD，ΓC，ΓGAnd updating the step size matrixes for the weights, wherein the step size matrixes are all symmetric positive definite matrixes.
CN202210557091.3A 2022-05-19 2022-05-19 RBF network-based upper limb exoskeleton robot motion control method Pending CN114750137A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202210557091.3A CN114750137A (en) 2022-05-19 2022-05-19 RBF network-based upper limb exoskeleton robot motion control method

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202210557091.3A CN114750137A (en) 2022-05-19 2022-05-19 RBF network-based upper limb exoskeleton robot motion control method

Publications (1)

Publication Number Publication Date
CN114750137A true CN114750137A (en) 2022-07-15

Family

ID=82334958

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202210557091.3A Pending CN114750137A (en) 2022-05-19 2022-05-19 RBF network-based upper limb exoskeleton robot motion control method

Country Status (1)

Country Link
CN (1) CN114750137A (en)

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN115755592A (en) * 2023-01-10 2023-03-07 常熟理工学院 Multi-mode control method for adjusting motion state of three-degree-of-freedom exoskeleton and exoskeleton
TWI808852B (en) * 2022-08-01 2023-07-11 崑山科技大學 Method for stable control of six-axis robotic arm by deep learning

Citations (11)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN107045557A (en) * 2016-11-01 2017-08-15 长春工业大学 The reconfigurable manipulator non-singular terminal sliding formwork power position control method of constraint oriented
WO2018000854A1 (en) * 2016-06-29 2018-01-04 深圳光启合众科技有限公司 Human upper limb motion intention recognition and assistance method and device
CN108942924A (en) * 2018-06-25 2018-12-07 南京理工大学 Model uncertainty mechanical arm motion control method based on multilayer neural network
WO2019213164A1 (en) * 2018-04-30 2019-11-07 Vanderbilt University A control method for a robotic system
CN110812104A (en) * 2018-08-07 2020-02-21 深圳二十一天健康科技有限公司 Arm exoskeleton rehabilitation system based on virtual reality
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
CN112180729A (en) * 2020-10-09 2021-01-05 电子科技大学中山学院 Self-adaptive trajectory tracking control method for mechanical arm
CN112462606A (en) * 2020-11-10 2021-03-09 哈尔滨工业大学 Flexible joint dynamic parameter identification method based on self-adaptive control
CN112775937A (en) * 2021-01-07 2021-05-11 合肥工业大学 Self-adaptive robust control method and device for under-actuated hydraulic single-leg power-assisted exoskeleton
CN112947071A (en) * 2021-01-28 2021-06-11 河北工业大学 Backstepping-based lower limb exoskeleton control method
CN114415508A (en) * 2022-01-10 2022-04-29 西安文理学院 Underwater mechanical arm control method based on block approximation RBF sliding mode control

Patent Citations (11)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO2018000854A1 (en) * 2016-06-29 2018-01-04 深圳光启合众科技有限公司 Human upper limb motion intention recognition and assistance method and device
CN107045557A (en) * 2016-11-01 2017-08-15 长春工业大学 The reconfigurable manipulator non-singular terminal sliding formwork power position control method of constraint oriented
WO2019213164A1 (en) * 2018-04-30 2019-11-07 Vanderbilt University A control method for a robotic system
CN108942924A (en) * 2018-06-25 2018-12-07 南京理工大学 Model uncertainty mechanical arm motion control method based on multilayer neural network
CN110812104A (en) * 2018-08-07 2020-02-21 深圳二十一天健康科技有限公司 Arm exoskeleton rehabilitation system based on virtual reality
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
CN112180729A (en) * 2020-10-09 2021-01-05 电子科技大学中山学院 Self-adaptive trajectory tracking control method for mechanical arm
CN112462606A (en) * 2020-11-10 2021-03-09 哈尔滨工业大学 Flexible joint dynamic parameter identification method based on self-adaptive control
CN112775937A (en) * 2021-01-07 2021-05-11 合肥工业大学 Self-adaptive robust control method and device for under-actuated hydraulic single-leg power-assisted exoskeleton
CN112947071A (en) * 2021-01-28 2021-06-11 河北工业大学 Backstepping-based lower limb exoskeleton control method
CN114415508A (en) * 2022-01-10 2022-04-29 西安文理学院 Underwater mechanical arm control method based on block approximation RBF sliding mode control

Non-Patent Citations (2)

* Cited by examiner, † Cited by third party
Title
李庆玲等: "外骨骼式上肢康复机器人力辅助控制", 哈尔滨工程大学学报, no. 02, pages 50 - 54 *
马莉丽等: "基于模型分块逼近的三关节机器人鲁棒滑模控制", 西安理工大学学报, no. 04, pages 63 - 68 *

Cited By (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
TWI808852B (en) * 2022-08-01 2023-07-11 崑山科技大學 Method for stable control of six-axis robotic arm by deep learning
CN115755592A (en) * 2023-01-10 2023-03-07 常熟理工学院 Multi-mode control method for adjusting motion state of three-degree-of-freedom exoskeleton and exoskeleton
CN115755592B (en) * 2023-01-10 2023-05-12 常熟理工学院 Multi-mode control method for adjusting motion state of three-degree-of-freedom exoskeleton and exoskeleton

Similar Documents

Publication Publication Date Title
CN112904728B (en) Mechanical arm sliding mode control track tracking method based on improved approach law
CN110275436B (en) RBF neural network self-adaptive control method of multi-single-arm manipulator
Soltanpour et al. Voltage based sliding mode control of flexible joint robot manipulators in presence of uncertainties
CN114750137A (en) RBF network-based upper limb exoskeleton robot motion control method
CN110977988B (en) Multi-joint mechanical arm impedance control method based on finite time command filtering
CN110434858B (en) Force/position hybrid control method of multi-mechanical-arm system based on command filtering
CN107263481B (en) A kind of class brain learning control method of multi-freedom robot
CN106826807B (en) Sliding mode variable structure control method of three-degree-of-freedom wrist structure
Hu et al. A reinforcement learning neural network for robotic manipulator control
CN110842913B (en) Adaptive sliding mode iterative learning control method of single-joint mechanical arm
CN115256386B (en) Uncertain mechanical arm neural self-adaptive control method considering tracking error constraint
Barhaghtalab et al. Dynamic analysis, simulation, and control of a 6-DOF IRB-120 robot manipulator using sliding mode control and boundary layer method
Wang et al. Udwadia-Kalaba approach based distributed consensus control for multi-mobile robot systems with communication delays
CN115179295A (en) Robust dichotomy consistency tracking control method for multi-Euler-Lagrange system
Jing et al. A recursive dynamic modeling and control for dual-arm manipulator with elastic joints
CN113219825B (en) Single-leg track tracking control method and system for four-leg robot
CN108227493B (en) Robot trajectory tracking method
Jiang Impedance control of flexible robot arms with parametric uncertainties
Truong et al. Adaptive neural sliding mode control for 3-DOF planar parallel manipulators
Tao et al. Application of neural network with real-time training to robust position/force control of multiple robots
Nguyen et al. Genetic algorithm tuned fuzzy logic controller for a robot arm with two-link flexibility and two-joint elasticity
CN111590561B (en) Robustness preset performance control method for distributed mechanical arm system
CN108132602A (en) Solid brewing yeast machine tool hand neural networks sliding mode self-adaptation control method
Wai et al. Tracking control design for robot manipulator via fuzzy neural network
Kalshetti et al. Self-adaptive grey wolf optimization based adaptive fuzzy aided sliding mode control for robotic manipulator

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