CN111983925A - Generalized dynamic prediction control method based on exoskeleton robot - Google Patents

Generalized dynamic prediction control method based on exoskeleton robot Download PDF

Info

Publication number
CN111983925A
CN111983925A CN202010892356.6A CN202010892356A CN111983925A CN 111983925 A CN111983925 A CN 111983925A CN 202010892356 A CN202010892356 A CN 202010892356A CN 111983925 A CN111983925 A CN 111983925A
Authority
CN
China
Prior art keywords
law
prediction
exoskeleton robot
control
double
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
CN202010892356.6A
Other languages
Chinese (zh)
Inventor
孙振兴
徐品进
张兴华
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.)
Nanjing Tech University
Original Assignee
Nanjing Tech University
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 Nanjing Tech University filed Critical Nanjing Tech University
Priority to CN202010892356.6A priority Critical patent/CN111983925A/en
Publication of CN111983925A publication Critical patent/CN111983925A/en
Pending legal-status Critical Current

Links

Images

Classifications

    • G—PHYSICS
    • G05—CONTROLLING; REGULATING
    • G05B—CONTROL OR REGULATING SYSTEMS IN GENERAL; FUNCTIONAL ELEMENTS OF SUCH SYSTEMS; MONITORING OR TESTING ARRANGEMENTS FOR SUCH SYSTEMS OR ELEMENTS
    • G05B13/00—Adaptive control systems, i.e. systems automatically adjusting themselves to have a performance which is optimum according to some preassigned criterion
    • G05B13/02—Adaptive control systems, i.e. systems automatically adjusting themselves to have a performance which is optimum according to some preassigned criterion electric
    • G05B13/04—Adaptive control systems, i.e. systems automatically adjusting themselves to have a performance which is optimum according to some preassigned criterion electric involving the use of models or simulators
    • G05B13/042—Adaptive control systems, i.e. systems automatically adjusting themselves to have a performance which is optimum according to some preassigned criterion electric involving the use of models or simulators in which a parameter or coefficient is automatically adjusted to optimise the performance

Landscapes

  • Engineering & Computer Science (AREA)
  • Health & Medical Sciences (AREA)
  • Artificial Intelligence (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Evolutionary Computation (AREA)
  • Medical Informatics (AREA)
  • Software Systems (AREA)
  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Automation & Control Theory (AREA)
  • Feedback Control In General (AREA)

Abstract

The invention discloses a generalized dynamic prediction control method based on an exoskeleton robot, and belongs to the field of exoskeleton robot systems. Aiming at the problem of nonparametric uncertainty in the exoskeleton robot, the invention provides a generalized dynamic prediction control method of a trigonometric system under the nonparametric uncertainty, and firstly, a nominal model of a cascade elastic actuator in the exoskeleton robot is obtained through Newton's law of motion; then, carrying out output prediction on the model, and designing a double-layer updating law to determine a proper prediction period of the system under non-parameter uncertainty; and finally, performing rolling time domain optimization on the performance indexes to obtain a control law. The method solves the problem that when the nominal model prediction state/output of a Model Prediction Control (MPC) method is uncertain, the precision can be changed, and finally the closed-loop control performance is deteriorated.

Description

Generalized dynamic prediction control method based on exoskeleton robot
Technical Field
The invention relates to the field of exoskeleton robot systems, in particular to a generalized dynamic prediction control method based on an exoskeleton robot.
Background
The exoskeleton robot is a human-computer electric device which is worn outside the body of an operator and integrates the technologies of advanced control, information coupling, mobile computing, communication and the like, can provide extra power or capacity for the wearer to enhance the functions of the human body, can complete specific functions and tasks under the control of the operator, and realizes the enhancement of the strength of the human body and the extension of sense organs. Exoskeleton robotics was first applied in the industrial field to provide support and assistance to operators, and has been gradually applied in the fields of fire rescue and nuclear detection with the development of control technology and the advancement of human-machine coupling technology. In recent years, exoskeleton robot technology has also been primarily applied to army for carrying individual loads.
The Chinese patent application, application number CN201710681749.0, published 2017, 11 and 28, discloses a control method of an upper limb exoskeleton rehabilitation robot based on a radial basis function neural network, and establishes a human upper limb musculoskeletal model; acquiring myoelectric signals of upper limb muscles and upper limb movement data, importing the movement data into an upper limb musculoskeletal model to obtain upper limb joint torque, constructing a radial basis function neural network, and giving out a neural network model; and identifying the movement intention of the patient, performing fusion analysis on the joint angular velocity, and using the result to identify the joint extension and flexion state of the training object to determine the movement intention of the limbs. The patent proposes a control method based on a neural network, but the method has too large calculation amount and is not easy to realize the controller on line.
The chinese patent application, application No. CN201610096527.8, published 2018, 4 and 6 discloses a control method for a lower limb exoskeleton robot. The control method is characterized in that an air bag pressure sensor is added on the basis of a traditional human body intention detection sensor for the lower limb exoskeleton, and the acting force of a human body and the exoskeleton is reflected by measuring a signal generated by pressing an air bag by a thigh of the human body, so that the human body movement intention is fed back, and the deviation of an exoskeleton control algorithm is corrected; the air bag sensor is used, and meanwhile, a flexible human-computer interface can be provided for a human body, so that the acting force of the human body and the exoskeleton robot is buffered. This patent proposes a fuzzy control method, but in this method, the exact method of the fuzzy set is complicated and relies on expert experience.
Disclosure of Invention
1. Technical problem to be solved
Aiming at the problem of nonparametric uncertainty in the exoskeleton robot, the invention provides a generalized dynamic prediction control method of a triangular system under nonparametric uncertainty, which can estimate the centralized influence of the system uncertainty by a double-layer updating law.
2. Technical scheme
The purpose of the invention is realized by the following technical scheme.
A generalized dynamic predictive control method based on an exoskeleton robot comprises the following steps:
step 1, writing a nominal model of a cascade elastic actuator in the exoskeleton robot by using a Newton's law of motion;
step 2, carrying out output prediction on the nominal model in the step 1, and designing a double-layer updating law to determine a proper prediction period of the system under nonparametric uncertainty;
and 3, performing rolling time domain optimization on the performance index based on the double-layer update rate in the step 2 to finally obtain a control law.
Further, in step 1, by using newton's law of motion, a nominal model of the tandem elastic actuators in the exoskeleton robot can be written as:
Figure BDA0002656581070000021
in the formula, qmAnd q islRespectively, motor angle and connecting rod angle, FmIs the motor torque; m ismAnd mlRespectively a motor inertia and a connecting rod inertia; k is the torsional spring stiffness; bmAnd blThe viscous friction coefficient of the motor and the viscous friction coefficient of the connecting rod are respectively.
Further, in step 2, a specific method for performing output prediction on the model is as follows:
first, let
Figure BDA0002656581070000022
And
Figure BDA0002656581070000023
wherein the content of the first and second substances,
Figure BDA00026565810700000211
is a reference signal for the link angle.
Then using the system to output
Figure BDA0002656581070000024
The prediction can be made during the prediction period by the following taylor expansion:
Figure BDA0002656581070000025
wherein the content of the first and second substances,
Figure BDA0002656581070000026
further, in step 2, the method for designing the double-layer update law is as follows:
step 201, let
Figure BDA0002656581070000027
Where ρ is1Are secondary design parameters. In rescaled coordinates, the system can be compressed as:
Figure BDA0002656581070000028
wherein the content of the first and second substances,
Figure BDA0002656581070000029
I∈R4×4the unit matrix is represented by a matrix of units,
Figure BDA00026565810700000210
Figure BDA0002656581070000031
wherein k is1,k2,k3,k4Is an optimum gain, is a constant related to the order only;
step 202, a double-layer updating law of the prediction period is provided, and the form of the double-layer updating law is as follows:
Figure BDA0002656581070000032
where ρ is1,ρ2,ρ3And ρ4Is a tunable parameter satisfying
Figure BDA0002656581070000033
Figure BDA0002656581070000034
Further, according to the two-layer update law, a prediction period can be derived:
T＝T(0)/L，T(0)＞0 (5)。
further, in step 3, the time domain optimization is performed on the performance index, and a method for obtaining the control law is as follows:
step 301, based on the nominal model, the performance index can be predicted as follows:
Figure BDA0002656581070000035
wherein the content of the first and second substances,
Figure BDA0002656581070000036
the partial derivative of U is calculated,
Figure BDA0002656581070000037
let
Figure BDA0002656581070000038
And is
Figure BDA0002656581070000039
Obtaining optimized control sequences
Figure BDA00026565810700000310
Step 302, using the first line of the control sequence to obtain the generalized dynamic predictive control law:
Figure BDA00026565810700000311
wherein, I ═ 1]∈R1×1Considering T2(i,j)＝pi,jT3+i+jAnd T3(i,j)＝qi,jT7+i+jThe control law can be simplified as follows:
Figure BDA00026565810700000312
wherein p isi,j，qi,jOnly constants related to the order, where k1,k2,k3,k4Is the optimum gain and is a constant related to the order only.
3. Advantageous effects
Compared with the prior art, the invention has the advantages that: a double-layer self-adaptive law is designed to estimate the centralized effect of system uncertainty instead of relying on the inherent robustness of a standard predictive controller or online/offline parameter identification; the scheme can also solve the problem that when the nominal model prediction state/output of a Model Prediction Control (MPC) method has uncertainty, the precision can be changed, and finally the performance of closed-loop control is deteriorated.
Drawings
FIG. 1 is a block diagram of an implementation of the proposed generalized dynamic predictive control method;
FIG. 2 is a schematic view of an exoskeleton robot actuator;
FIG. 3 shows reference values
Figure BDA0002656581070000041
Predicting the track tracking performance under the controller in a time-generalized dynamic manner;
FIG. 4 shows reference values
Figure BDA0002656581070000042
And predicting the track tracking performance under the controller in a time-generalized manner.
Detailed Description
The invention is described in detail below with reference to the drawings and specific examples.
Examples
The invention provides a generalized dynamic prediction control method based on an exoskeleton robot, which comprises the following steps:
step 1, writing a nominal model of a tandem elastic actuator SEA in the exoskeleton robot by using newton's law of motion, specifically, as shown in fig. 1, an uncertainty system can be written as:
Figure BDA0002656581070000043
wherein the content of the first and second substances,
Figure BDA0002656581070000044
u and y are the system state, control input and control output, respectively.
Figure BDA0002656581070000045
Is an unknown parameter vector and is assumed to be within a known range.
Figure BDA0002656581070000046
aAnd
Figure BDA0002656581070000047
are known constants.
Figure BDA0002656581070000048
Is an uncertain non-linear function.
In step 1 above, by using newton's law of motion, a nominal model of the tandem elastic actuators SEA in the exoskeleton robot can be written as:
Figure BDA0002656581070000049
as shown in FIG. 2, wherein q ismAnd q islRespectively is a motor included angle and a connecting rod included angle; fmIs the motor torque; m ismAnd mlRespectively a motor inertia and a connecting rod inertia; k is the torsional spring stiffness; bmAnd blThe viscous friction coefficient of the motor and the viscous friction coefficient of the connecting rod are respectively.
Step 2, carrying out output prediction on the model in the step 1, and designing a double-layer updating law to determine a proper prediction period of the system under the condition of non-parameter uncertainty, wherein the specific process is as follows:
first, the output prediction is performed on the system:
let
Figure BDA0002656581070000051
And
Figure BDA0002656581070000052
wherein the content of the first and second substances,
Figure BDA0002656581070000053
is a reference signal for the link angle.
Then using the system to output
Figure BDA0002656581070000054
During the prediction period (0. ltoreq. tau. ltoreq.T) can be predicted by the following Taylor expansion:
Figure BDA0002656581070000055
wherein the content of the first and second substances,
Figure BDA0002656581070000056
then, a double-layer update law is designed:
let
Figure BDA0002656581070000057
Where ρ is1Are secondary design parameters. In rescaled coordinates, the system can be compressed as:
Figure BDA0002656581070000058
wherein the content of the first and second substances,
Figure BDA0002656581070000059
Figure BDA00026565810700000510
and I ∈ R4×4Representing an identity matrix.
A two-level update law for the prediction period is proposed, which is of the form:
Figure BDA00026565810700000511
where ρ is1,ρ2,ρ3And ρ4Is a tunable parameter satisfying
Figure BDA00026565810700000512
Figure BDA00026565810700000513
Finally, according to the two-layer update law, a prediction period can be obtained:
T＝T(0)/L，T(0)＞0 (5)
by a two-layer update law, the collective effect of all system uncertainties can be estimated, rather than identifying all parameters.
And 3, based on the double-layer update rate in the step 2, performing rolling time domain optimization on the performance index to obtain a control law, wherein the specific process is as follows:
first, based on a nominal model, a performance index can be predicted as
Figure BDA0002656581070000061
Wherein the content of the first and second substances,
Figure BDA0002656581070000062
then, find out
Figure BDA0002656581070000063
For the partial derivative of U the number of the partial derivatives,
Figure BDA0002656581070000064
let
Figure BDA0002656581070000065
And is
Figure BDA0002656581070000066
Obtaining optimized control sequences
Figure BDA0002656581070000067
And finally, obtaining the achievable generalized dynamic prediction control law by adopting the first line of the control sequence:
Figure BDA0002656581070000068
in the formula (I), the compound is shown in the specification,
Figure BDA0002656581070000069
taking into account T2(i,j)＝pi,jT3+i+jAnd T3(i,j)＝qi,jT7+i+jThe control law can be simplified as follows:
Figure BDA00026565810700000610
in the formula, pi,j，qi,jOnly a constant related to the order.
After software simulation, the reference angles in FIG. 3 and FIG. 4 are respectively
Figure BDA00026565810700000611
In the meantime, the generalized dynamic predictive control method and the generalized predictive control method have the trajectory tracking performance, the dotted line is the ideal tracking performance, and the solid line is the actual tracking performance. It can be seen that the closer the dotted line and the solid line in the graph are, the better the tracking performance is, and the comparison between fig. 4 and fig. 3 shows that the tracking performance of the generalized predictive control method is not as good as that of the generalized dynamic predictive control method.
The invention and its embodiments have been described above schematically, without limitation, and the invention can be embodied in other specific forms without departing from the spirit or essential characteristics thereof. The representation in the drawings is only one of the embodiments of the invention, the actual construction is not limited thereto, and any reference signs in the claims shall not limit the claims concerned. Therefore, if a person skilled in the art receives the teachings of the present invention, without inventive design, a similar structure and an embodiment to the above technical solution should be covered by the protection scope of the present patent. Furthermore, the word "comprising" does not exclude other elements or steps, and the word "a" or "an" preceding an element does not exclude the presence of a plurality of such elements. Several of the elements recited in the product claims may also be implemented by one element in software or hardware. The terms first, second, etc. are used to denote names, but not any particular order.

Claims (6)

1. A generalized dynamic predictive control method based on an exoskeleton robot is characterized by comprising the following steps:
step 1, obtaining a nominal model of the cascade elastic actuator of the exoskeleton robot through Newton's law of motion;
step 2, carrying out output prediction on the nominal model in the step 1, and designing a double-layer updating law for determining a prediction period of the system under the condition of non-parameter uncertainty;
and 3, performing rolling time domain optimization on the performance index based on the double-layer update rate in the step 2 to obtain a control law.
2. The method of claim 1, wherein in step 1, the nominal model of the cascaded elastic actuators in the exoskeleton robot can be written as:
Figure FDA0002656581060000011
in the formula, qmAnd q islRespectively is a motor included angle and a connecting rod included angle; fmIs the motor torque; m ismAnd mlRespectively a motor inertia and a connecting rod inertia; k is the torsional spring stiffness; bmAnd blThe viscous friction coefficient of the motor and the viscous friction coefficient of the connecting rod are respectively.
3. The method of claim 1, wherein in step 2, the method of output prediction of the model is as follows:
let
Figure FDA0002656581060000012
And
Figure FDA0002656581060000013
wherein x is the system state, u is the control input,
Figure FDA0002656581060000014
is a reference signal of the link angle;
then using the system to output
Figure FDA0002656581060000015
During the prediction period (0. ltoreq. tau. ltoreq.T) can be predicted by the following Taylor expansion:
Figure FDA0002656581060000016
wherein the content of the first and second substances,
Figure FDA0002656581060000017
4. the method for generalized dynamic predictive control based on exoskeleton robots as claimed in claim 1, wherein in step 2, the method for designing the double-layer update law is as follows:
step 201, let
Figure FDA0002656581060000018
Where ρ is1Is an auxiliary design parameter;
in rescaled coordinates, the system can be compressed as:
Figure FDA0002656581060000019
in the formula (I), the compound is shown in the specification,
Figure FDA00026565810600000110
I∈R4×4the unit matrix is represented by a matrix of units,
Figure FDA00026565810600000111
L(0)＝1，
Figure FDA0002656581060000021
wherein k is1，k2，k3，k4Is an optimum gain, is a constant related to the order only;
step 202, a double-layer updating law of the prediction period is provided, and the form of the double-layer updating law is as follows:
Figure FDA0002656581060000022
where ρ is1，ρ2，ρ3And ρ4Is an adjustable parameter and satisfies:
Figure FDA0002656581060000023
Figure FDA0002656581060000024
5. the method of claim 4, wherein according to the two-level update law, a prediction cycle is derived:
T＝T(0)/L，T(0)＞0 (5)。
6. the method as claimed in claim 1, wherein in step 3, the time domain optimization of the performance index is performed to obtain the control law as follows:
step 301, based on the nominal model, the performance index can be predicted as follows:
Figure FDA0002656581060000025
wherein the content of the first and second substances,
Figure FDA0002656581060000026
to find
Figure FDA0002656581060000027
For the partial derivative of U the number of the partial derivatives,
Figure FDA0002656581060000028
let
Figure FDA0002656581060000029
And is
Figure FDA00026565810600000210
Obtaining optimized control sequences
Figure FDA00026565810600000211
Step 302, using the first line of the control sequence to obtain the generalized dynamic predictive control law:
Figure FDA00026565810600000212
wherein I is ═ 1]∈R1×1Considering T2(i，j)＝pi，jT3+i+jAnd T3(i，j)＝qi，jT7+i+jThe control law can be simplified as follows:
Figure FDA0002656581060000031
in the formula, pi，j，qi，jI, j is a constant related to the order, where k1，k2，k3，k4Is the optimum gain and is a constant related to the order only.
CN202010892356.6A 2020-08-28 2020-08-28 Generalized dynamic prediction control method based on exoskeleton robot Pending CN111983925A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202010892356.6A CN111983925A (en) 2020-08-28 2020-08-28 Generalized dynamic prediction control method based on exoskeleton robot

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202010892356.6A CN111983925A (en) 2020-08-28 2020-08-28 Generalized dynamic prediction control method based on exoskeleton robot

Publications (1)

Publication Number Publication Date
CN111983925A true CN111983925A (en) 2020-11-24

Family

ID=73440399

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202010892356.6A Pending CN111983925A (en) 2020-08-28 2020-08-28 Generalized dynamic prediction control method based on exoskeleton robot

Country Status (1)

Country Link
CN (1) CN111983925A (en)

Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JP2001134320A (en) * 1999-11-01 2001-05-18 Honda Motor Co Ltd Lane follow-up controller
CN107942671A (en) * 2017-12-04 2018-04-20 国网山东省电力公司电力科学研究院 A kind of improved underwater robot Work machine arm generalized forecast control method
CN111290273A (en) * 2020-02-18 2020-06-16 湖州和力机器人智能科技有限公司 Position tracking optimization control method based on exoskeleton robot flexible actuator
CN111522243A (en) * 2020-05-20 2020-08-11 河北工业大学 Robust iterative learning control strategy for five-degree-of-freedom upper limb exoskeleton system

Patent Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JP2001134320A (en) * 1999-11-01 2001-05-18 Honda Motor Co Ltd Lane follow-up controller
CN107942671A (en) * 2017-12-04 2018-04-20 国网山东省电力公司电力科学研究院 A kind of improved underwater robot Work machine arm generalized forecast control method
CN111290273A (en) * 2020-02-18 2020-06-16 湖州和力机器人智能科技有限公司 Position tracking optimization control method based on exoskeleton robot flexible actuator
CN111522243A (en) * 2020-05-20 2020-08-11 河北工业大学 Robust iterative learning control strategy for five-degree-of-freedom upper limb exoskeleton system

Non-Patent Citations (2)

* Cited by examiner, † Cited by third party
Title
YUNDA YAN 等: "Generalized Dynamic Predictive Control for Nonparametric Uncertain Systems With Application to Series Elastic Actuators", 《IEEE TRANSACTIONS ON INDUSTRIAL INFORMATICS》 *
张高巍 等: "可穿戴五自由度上肢外骨骼机器人固定时间控制", 《控制理论与应用》 *

Similar Documents

Publication Publication Date Title
Li et al. Asymmetric bimanual control of dual-arm exoskeletons for human-cooperative manipulations
Huang et al. Echo state network based predictive control with particle swarm optimization for pneumatic muscle actuator
Daachi et al. A radial basis function neural network adaptive controller to drive a powered lower limb knee joint orthosis
Babaiasl et al. Sliding mode control of an exoskeleton robot for use in upper-limb rehabilitation
Yin et al. Personalised control of robotic ankle exoskeleton through experience-based adaptive fuzzy inference
Chen et al. Control and implementation of 2-DOF lower limb exoskeleton experiment platform
CN114102600B (en) Multi-space fusion human-machine skill migration and parameter compensation method and system
CN114467097A (en) Method for learning parameters of a neural network, for generating trajectories of an exoskeleton and for setting the exoskeleton in motion
Bao et al. Model predictive control of a feedback-linearized hybrid neuroprosthetic system with a barrier penalty
Wang et al. Nonlinear disturbance observer based flexible-boundary prescribed performance control for a lower limb exoskeleton
Yang et al. Adaptive neural network-based sliding mode control for a hydraulic rotary drive joint
Cao et al. A visual servo-based predictive control with echo state gaussian process for soft bending actuator
Ning et al. Human-robot interaction control for robot driven by variable stiffness actuator with force self-sensing
Jin et al. Motion prediction of human wearing powered exoskeleton
Tang et al. Direct force control of upper-limb exoskeleton based on fuzzy adaptive algorithm
Jäntsch et al. Adaptive neural network dynamic surface control: An evaluation on the musculoskeletal robot anthrob
Long et al. Development of a lower extremity wearable exoskeleton with double compact elastic module: preliminary experiments
CN113733117A (en) Reconfigurable robot human intention identification optimal control method and device
Qu et al. Research on fuzzy adaptive impedance control of lower extremity exoskeleton
CN111983925A (en) Generalized dynamic prediction control method based on exoskeleton robot
Soltani Sharif Abadi et al. Fixed‐time observer‐based controller for the human–robot collaboration with interaction force estimation
CN114851171B (en) Gait track tracking control method of lower limb exoskeleton rehabilitation robot
CN111781841B (en) Limited time model prediction control method based on exoskeleton robot
Cao et al. Adaptive sliding mode impedance control in lower limbs rehabilitation robotic
CN111796525B (en) Model prediction control method based on exoskeleton robot mechanical arm

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
RJ01 Rejection of invention patent application after publication
RJ01 Rejection of invention patent application after publication

Application publication date: 20201124