CN109711527A - A kind of Robotic Manipulator method based on particle swarm optimization algorithm - Google Patents
A kind of Robotic Manipulator method based on particle swarm optimization algorithm Download PDFInfo
- Publication number
- CN109711527A CN109711527A CN201811594798.1A CN201811594798A CN109711527A CN 109711527 A CN109711527 A CN 109711527A CN 201811594798 A CN201811594798 A CN 201811594798A CN 109711527 A CN109711527 A CN 109711527A
- Authority
- CN
- China
- Prior art keywords
- particle
- control parameter
- task
- optimization algorithm
- swarm optimization
- 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
Links
Landscapes
- Manipulator (AREA)
- Feedback Control In General (AREA)
Abstract
The Robotic Manipulator method based on particle swarm optimization algorithm that the invention proposes a kind of, comprising: the task parameters of each task are set；It is iterated training using particle swarm optimization algorithm, generates optimal control parameter, comprising: carries out population initialization, impedance control parameter is generated by single particle；Real-time circuit is manipulated according to the control parameter, completes the manipulation task to robot；It evaluates to obtain the fitness of each particle as a result, passing through the control performance detected according to completion manipulation task, then more new individual history desired positions and particle group history desired positions, the speed of particle and position；If global optimum position meets minimum limit, stacking generation is trained to finish, exports optimal control parameter to task layer.The present invention use simple and convenient, training process light weight and quickly, maneuvering performance it is excellent, a variety of Robotic Manipulator tasks can be competent at.
Description
Technical field
The present invention relates to Industrial Robot Technology field, in particular to a kind of robot behaviour based on particle swarm optimization algorithm
Vertical method.
Background technique
After the fast development after many decades, whether theoretical or practice has all tended to for the motion control of robot
It is mature with interaction that is perfect, but when it come to arriving machine human and environment, especially expectation robot can crawl freely and
When manipulating object, all none satisfactory solution so far.Possessing reliable maneuvering capability is that robot can be true
Laboratory is just walked out, into the necessary condition in mass consumption market, this shows a kind of practical and handy Robotic Manipulator skill
Art is very urgent and real demand.
In Robotic Manipulator technology, bottom control frame generally uses variable element impedance control, and wherein most critical with
Most complicated part is then how to choose and determines optimal control parameter.
Current existing technical solution mainly includes following two:
1, control parameter is manually adjusted, under determining operating condition, by complicated modeling and optimization, obtaining one group can be real
The control parameter of existing specific operation function, when environmental characteristics variation (mass inertia of such as manipulation of objects changes, contact surface it is rigid
Degree change etc.), it needs to re-start parameter tuning.
2, adjust automatically control parameter, using machine learning algorithm (such as deep learning, enhancing learn) by training certainly
The dynamic control parameter obtained under different operating conditions.
But the major defect of above two technical approach is with deficiency: since control parameter is numerous and is mutually coupled,
It manually adjusts parameter and needs a large amount of deposit knowledge, not only ordinary consumer is difficult to complete with common laborer, professional technician
Need to expend a large amount of manpower and time cost, and each operating condition in impossible exhaustive reality.Due to deep learning meter
Huge, existing adjust automatically parameter algorithm (such as peg-in-hole assembly) when calculating simplest operating condition requires to count on a large scale
The magnanimity of calculation machine cluster calculates power, this is more difficult to realize in complicated consumption scene and industrial scene.
Summary of the invention
The purpose of the present invention aims to solve at least one of described technological deficiency.
For this purpose, it is an object of the invention to propose a kind of Robotic Manipulator method based on particle swarm optimization algorithm.
To achieve the goals above, the embodiment of the present invention provides a kind of Robotic Manipulator based on particle swarm optimization algorithm
Method, comprising:
Step S1 informs the task that robot needs to complete, i.e. constrained parameters of the setting task in each execution stage,
In include: original state position constraint and force constraint xinit、Finit, a series of intermediate states position constraint and force constraint
xexec1、Fexec1、xexec2、Fexec2... and the position constraint under completion status and force constraint xfini、Ffini；These constraints are for true
Determine jump condition of the robot in operation between locating each state and each state；
Step S2 is iterated training using particle swarm optimization algorithm, generates optimal control parameter, comprising:
Population initialization is carried out, impedance control parameter is generated by particle position；
Robot impedance control is carried out according to control parameter generated, completes manipulation task；
Result of manipulation corresponding to each particle is investigated, using the time needed for completing operation task as the adaptation of particle
Degree, thus more new individual history desired positions dpbestWith group history desired positions dgbest, and the speed of more new particle and position
It sets；
If global optimum position meets minimum limit, stacking generation is trained to finish, to task layer output optimum control ginseng
Number, regenerates impedance parameter by updated particle state if not up to termination condition, and carries out new round training.
Further, in the step S2, the progress population initialization, comprising: setting particle be and the impedance
The relevant n-dimensional vector of control parameter number, setting population scale 25 or so, speed maximum value parameter variation range 20% with
It is interior, the random initial position d for generating particleinitWith speed vinit。
Further, in the step S2, according to the state of each particle and task parameters determine impedance control parameter alpha,
β、γα、γβAnd desired trajectory parameter xd、Fd；
Impedance control rate corresponding to each particle is generated as the impedance control parameter and desired trajectory parameter
Wherein, FffFor adaptive feedforward torque；FdFor feedforward torque corresponding to ideal trajectory；K is impedance control rigidity；
D is damping matrix；E is deviation；α,γαFor the Studying factors and forgetting factor of torque feedforward；β,γβFor the Studying factors of rigidity
And forgetting factor；T is controller step-length；κ is adaptive tracing error；J is robot Jacobian matrix.
Further, using robot-operating time as particle fitness, each particle is calculated first under initial position
Fitness, selects the highest position of fitness as history desired positions, in each iteration, by the current fitness of particle with
The fitness of history desired positions is made comparisons, with more new individual history desired positions dpbestWith particle group history desired positions
dgbest。
Further, in the step S2, the speed of the particle and position, comprising:
vi=vi+c1·R1·(dpbest-di)+c2·R1·(dgbest-di) (4)
di=di+vi (5)
In formula, viFor particle rapidity, diFor particle position, that is, characterize the multi-C vector of impedance control parameter, R1、R2For 0~
Random number between 1 characterizes the randomness of Particles Moving, c1、c2For acceleration constant, individual cognition is characterized respectively and is recognized with group
Know the size influenced on particle.
Robotic Manipulator method according to an embodiment of the present invention based on particle swarm optimization algorithm, is controlled using adjust automatically
The strategy of parameter carries out robot impedance control parameter training using particle swarm optimization algorithm, obtains optimal control parameter.This hair
Bright realization is divided into three levels: outermost layer is task layer, and different tasks has determined different restrictions on the parameters, and is transmitted
To middle layer；Middle layer is training layer, and iteration can all call particle swarm optimization algorithm that epicycle training is calculated generated each time
Control parameter is passed to internal layer；Internal layer is Robotic Manipulator real-time control circuit, and the control parameter being passed to using middle layer is complete
At manipulation task.The present invention is by user open tasks layer input interface, by internal control algorithm and parameter tuning algorithm
Depth encapsulation, uses simple and convenient, need to only inform that robot needs the completing of the task.Training process light weight and fast
Speed, not needing huge hardware calculation power can be completed training, have the basis realized in reality scene.Also, maneuvering performance
It is excellent, a variety of Robotic Manipulator tasks can be competent at.
The additional aspect of the present invention and advantage will be set forth in part in the description, and will partially become from the following description
Obviously, or practice through the invention is recognized.
Detailed description of the invention
Above-mentioned and/or additional aspect of the invention and advantage will become from the description of the embodiment in conjunction with the following figures
Obviously and it is readily appreciated that, in which:
Fig. 1 is the flow chart according to the Robotic Manipulator method based on particle swarm optimization algorithm of the embodiment of the present invention；
Fig. 2 is the schematic diagram according to the Robotic Manipulator method based on particle swarm optimization algorithm of the embodiment of the present invention.
Specific embodiment
The embodiment of the present invention is described below in detail, the example of embodiment is shown in the accompanying drawings, wherein identical from beginning to end
Or similar label indicates same or similar element or element with the same or similar functions.It is retouched below with reference to attached drawing
The embodiment stated is exemplary, it is intended to is used to explain the present invention, and is not considered as limiting the invention.
As shown in Figs. 1-2, the Robotic Manipulator method based on particle swarm optimization algorithm of the embodiment of the present invention, including it is as follows
Step:
Step S1 informs the task that robot needs to complete, i.e. constrained parameters of the setting task in each execution stage,
In include: original state position constraint and force constraint xinit、Finit, a series of intermediate states position constraint and force constraint
xexec1、Fexec1、xexec2、Fexec2... and the position constraint under completion status and force constraint xfini、Ffini；These constraints are for true
Determine jump condition of the robot in operation between locating each state and each state.
Step S2 is iterated training using particle swarm optimization algorithm, generates optimal control parameter, comprising:
Step S21 carries out population initialization, generates impedance control parameter by particle position.
Specifically, carry out population initialization, comprising: setting particle be n relevant to impedance control number of parameters tie up to
Amount is arranged population scale, the upper and lower bound of particle position and speed is determined according to task parameters, wherein population scale can be with
25 or so, speed maximum value is within parameter variation range 20%.Then the initial position d of particle is generated at randominitWith speed
Spend vinit。
Impedance control parameter alpha, β, γ are determined according to the state of each particle and task parametersα、γβAnd desired trajectory parameter
xd、FdDeng.
Impedance control rate corresponding to each particle is generated as impedance control parameter and desired trajectory parameter, specific as follows:
Wherein, FffFor adaptive feedforward torque；FdFor feedforward torque corresponding to ideal trajectory；K is impedance control rigidity；
D is damping matrix；E is deviation；α,γαFor the Studying factors and forgetting factor of torque feedforward；β,γβFor the Studying factors of rigidity
And forgetting factor；T is controller step-length；κ is adaptive tracing error；J is robot Jacobian matrix.
Step S22 carries out robot impedance control according to control parameter generated, completes manipulation task.
Step S23 investigates result of manipulation corresponding to each particle, using the time needed for completing operation task as particle
Fitness, thus more new individual history desired positions dpbestWith group history desired positions dgbest, and the speed of more new particle
With position.
In this step, using robot-operating time as particle fitness, each particle is calculated first in initial position
Under fitness, select the highest position of fitness as history desired positions, in each iteration, particle currently adapted to
Degree is made comparisons with the fitness of history desired positions, with more new individual history desired positions dpbestWith the best position of particle group history
Set dgbest。
In addition, speed and the position of more new particle, using following formula:
vi=vi+c1·R1·(dpbest-di)+c2·R1·(dgbest-di) (4)
di=di+vi (5)
In formula, viFor particle rapidity, diFor particle position, that is, characterize the multi-C vector of impedance control parameter, R1、R2For 0~
Random number between 1 characterizes the randomness of Particles Moving, c1、c2For acceleration constant, individual cognition is characterized respectively and is recognized with group
Know the size influenced on particle.
Step S24 trains stacking generation to finish if global optimum position meets minimum limit, most to task layer output
Excellent control parameter regenerates impedance parameter by updated particle state if not up to termination condition, and carries out new one
Wheel training.
Robotic Manipulator method according to an embodiment of the present invention based on particle swarm optimization algorithm, is controlled using adjust automatically
The strategy of parameter carries out robot impedance control parameter training using particle swarm optimization algorithm, obtains optimal control parameter.This hair
Bright realization is divided into three levels: outermost layer is task layer, and different tasks has determined different restrictions on the parameters, and is transmitted
To middle layer；Middle layer is training layer, and iteration can all call particle swarm optimization algorithm that epicycle training is calculated generated each time
Control parameter is passed to internal layer；Internal layer is Robotic Manipulator real-time control circuit, and the control parameter being passed to using middle layer is complete
At manipulation task.The present invention is by user open tasks layer input interface, by internal control algorithm and parameter tuning algorithm
Depth encapsulation, uses simple and convenient, need to only inform that robot needs the completing of the task.Training process light weight and fast
Speed, not needing huge hardware calculation power can be completed training, have the basis realized in reality scene.Also, maneuvering performance
It is excellent, a variety of Robotic Manipulator tasks can be competent at.
In the description of this specification, reference term " one embodiment ", " some embodiments ", " example ", " specifically show
The description of example " or " some examples " etc. means specific features, structure, material or spy described in conjunction with this embodiment or example
Point is included at least one embodiment or example of the invention.In the present specification, schematic expression of the above terms are not
Centainly refer to identical embodiment or example.Moreover, particular features, structures, materials, or characteristics described can be any
One or more embodiment or examples in can be combined in any suitable manner.
Although the embodiments of the present invention has been shown and described above, it is to be understood that above-described embodiment is example
Property, it is not considered as limiting the invention, those skilled in the art are not departing from the principle of the present invention and objective
In the case where can make changes, modifications, alterations, and variations to the above described embodiments within the scope of the invention.The scope of the present invention
By appended claims and its equivalent limit.
Claims (5)
1. a kind of Robotic Manipulator method based on particle swarm optimization algorithm, which comprises the steps of:
Step S1 informs the task that robot needs to complete, i.e. constrained parameters of the setting task in each execution stage, wherein wrapping
It includes: the position constraint and force constraint x of original stateinit、Finit, a series of intermediate states position constraint and force constraint xexec1、
Fexec1、xexec2、Fexec2... and the position constraint under completion status and force constraint xfini、Ffini；These constraints are for determining machine
People's jump condition between locating each state and each state in operation；
Step S2 is iterated training using particle swarm optimization algorithm, generates optimal control parameter, comprising:
Population initialization is carried out, impedance control parameter is generated by particle position；
Robot impedance control is carried out according to control parameter generated, completes manipulation task；
Result of manipulation corresponding to each particle is investigated, using the time needed for completing operation task as the fitness of particle, from
And more new individual history desired positions dpbestWith group history desired positions dgbest, and the speed of more new particle and position；
If global optimum position meets minimum limit, stacking generation is trained to finish, exports optimal control parameter to task layer, such as
Fruit is not up to termination condition and then regenerates impedance parameter by updated particle state, and carries out new round training.
2. such as Robotic Manipulator method of the claim 1 based on particle swarm optimization algorithm, which is characterized in that in the step S2
In, the progress population initialization, comprising: setting particle is n-dimensional vector relevant to the impedance control number of parameters, if
Population scale is set 25 or so, speed maximum value is within parameter variation range 20%, the random initial position for generating particle
dinitWith speed vinit。
3. such as Robotic Manipulator method of the claim 1 based on particle swarm optimization algorithm, which is characterized in that in the step S2
In, impedance control parameter alpha, β, γ are determined according to the state of each particle and task parametersα、γβAnd desired trajectory parameter xd、Fd；
Impedance control rate corresponding to each particle is generated as the impedance control parameter and desired trajectory parameter
Wherein, FffFor adaptive feedforward torque；FdFor feedforward torque corresponding to ideal trajectory；K is impedance control rigidity；D is
Damping matrix；E is deviation；α,γαFor the Studying factors and forgetting factor of torque feedforward；β,γβFor the Studying factors and something lost of rigidity
Forget the factor；T is controller step-length；κ is adaptive tracing error；J is robot Jacobian matrix.
4. such as Robotic Manipulator method of the claim 1 based on particle swarm optimization algorithm, which is characterized in that in the step S2
In, using robot-operating time as particle fitness, fitness of each particle under initial position is calculated first, is selected suitable
The highest position of response is as history desired positions, in each iteration, by the current fitness of particle and the best position of history
The fitness set is made comparisons, with more new individual history desired positions dpbestWith particle group history desired positions dgbest。
5. such as Robotic Manipulator method of the claim 1 based on particle swarm optimization algorithm, which is characterized in that in the step S2
In, the speed of the particle and position are updated by following formula, obtain updated robot impedance control parameter:
vi=vi+c1·R1·(dpbest-di)+c2·R1·(dgbest-di)
di=di+vi
In formula, viFor particle rapidity, diFor particle position, that is, characterize the multi-C vector of impedance control parameter, R1、R2Between 0~1
Random number, characterize the randomness of Particles Moving, c1、c2For acceleration constant, individual cognition and group cognition are characterized respectively to grain
The size that son influences.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811594798.1A CN109711527B (en) | 2018-12-25 | 2018-12-25 | Robot control method based on particle swarm optimization algorithm |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811594798.1A CN109711527B (en) | 2018-12-25 | 2018-12-25 | Robot control method based on particle swarm optimization algorithm |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN109711527A true CN109711527A (en) | 2019-05-03 |
| CN109711527B CN109711527B (en) | 2023-07-28 |
Family
ID=66258086
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201811594798.1A Active CN109711527B (en) | 2018-12-25 | 2018-12-25 | Robot control method based on particle swarm optimization algorithm |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109711527B (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113985864A (en) * | 2021-08-17 | 2022-01-28 | 北京空间飞行器总体设计部 | Dynamically walking four-footed detection robot and control method |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20080243307A1 (en) * | 2007-03-26 | 2008-10-02 | Honda Research Institute Europe Gmbh | Apparatus and Method for Generating and Controlling the Motion of a Robot |
| CN101436073A (en) * | 2008-12-03 | 2009-05-20 | 江南大学 | Wheeled mobile robot trace tracking method based on quantum behavior particle cluster algorithm |
| CN103869704A (en) * | 2014-04-08 | 2014-06-18 | 哈尔滨工业大学 | Method for coordination control over satellite arms of space robot based on expanded Jacobian matrix |
| CN104517297A (en) * | 2013-09-28 | 2015-04-15 | 沈阳新松机器人自动化股份有限公司 | Robot calibrate method based on particle swarm optimization |
| CN108068113A (en) * | 2017-11-13 | 2018-05-25 | 苏州大学 | 7-DOF humanoid arm flying object operation minimum acceleration trajectory optimization |
| CN108829137A (en) * | 2018-05-23 | 2018-11-16 | 中国科学院深圳先进技术研究院 | A kind of barrier-avoiding method and device of robot target tracking |
-
2018
- 2018-12-25 CN CN201811594798.1A patent/CN109711527B/en active Active
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20080243307A1 (en) * | 2007-03-26 | 2008-10-02 | Honda Research Institute Europe Gmbh | Apparatus and Method for Generating and Controlling the Motion of a Robot |
| CN101436073A (en) * | 2008-12-03 | 2009-05-20 | 江南大学 | Wheeled mobile robot trace tracking method based on quantum behavior particle cluster algorithm |
| CN104517297A (en) * | 2013-09-28 | 2015-04-15 | 沈阳新松机器人自动化股份有限公司 | Robot calibrate method based on particle swarm optimization |
| CN103869704A (en) * | 2014-04-08 | 2014-06-18 | 哈尔滨工业大学 | Method for coordination control over satellite arms of space robot based on expanded Jacobian matrix |
| CN108068113A (en) * | 2017-11-13 | 2018-05-25 | 苏州大学 | 7-DOF humanoid arm flying object operation minimum acceleration trajectory optimization |
| CN108829137A (en) * | 2018-05-23 | 2018-11-16 | 中国科学院深圳先进技术研究院 | A kind of barrier-avoiding method and device of robot target tracking |
Non-Patent Citations (1)
| Title |
|---|
| 周晓东等: "基于粒子群算法的阻抗控制在机械臂柔顺控制中的应用", 《空间控制技术与应用》 * |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113985864A (en) * | 2021-08-17 | 2022-01-28 | 北京空间飞行器总体设计部 | Dynamically walking four-footed detection robot and control method |
Also Published As
| Publication number | Publication date |
|---|---|
| CN109711527B (en) | 2023-07-28 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN108115681A (en) | Learning by imitation method, apparatus, robot and the storage medium of robot | |
| CN107272403A (en) | A kind of PID controller parameter setting algorithm based on improvement particle cluster algorithm | |
| CN105956297B (en) | Comprehensive evaluation and optimization method for redundant robot motion flexibility performance | |
| Zhao et al. | Impedance control and performance measure of series elastic actuators | |
| CN109702744A (en) | A method of the robot learning by imitation based on dynamic system model | |
| CN107102644A (en) | The underwater robot method for controlling trajectory and control system learnt based on deeply | |
| CN106874914A (en) | A kind of industrial machinery arm visual spatial attention method based on depth convolutional neural networks | |
| CN102184454B (en) | Granulator formula generation method based on neural network system | |
| Walke et al. | Bridgedata v2: A dataset for robot learning at scale | |
| Shahid et al. | Learning continuous control actions for robotic grasping with reinforcement learning | |
| CN106078748A (en) | A kind of robot based on hands, eye, arm coordinated manipulation captures the control method of object | |
| CN108527372A (en) | A kind of joint of robot self-adaptation control method of variation rigidity series elastic driver | |
| CN106774379A (en) | A kind of strong robust attitude control method of intelligent supercoil | |
| WO2018227820A1 (en) | Method and device for controlling manipulator movement, storage medium, and terminal device | |
| CN108427282A (en) | A kind of solution of Inverse Kinematics method based on learning from instruction | |
| CN112700060A (en) | Station terminal load prediction method and prediction device | |
| CN108555914A (en) | A kind of DNN Neural Network Adaptive Control methods driving Dextrous Hand based on tendon | |
| CN109711527A (en) | A kind of Robotic Manipulator method based on particle swarm optimization algorithm | |
| CN117093033A (en) | Resistance heating furnace temperature control system for optimizing PID parameters based on particle swarm optimization | |
| CN115179295A (en) | Robust dichotomy consistency tracking control method for multi-Euler-Lagrange system | |
| Zakaria et al. | Robotic control of the deformation of soft linear objects using deep reinforcement learning | |
| CN113190029B (en) | Adaptive gait autonomous generation method of four-footed robot based on deep reinforcement learning | |
| Luo et al. | Interactive generation of dynamically feasible robot trajectories from sketches using temporal mimicking | |
| CN110587611A (en) | Mechanical arm control method for television set assembly line | |
| Marvel et al. | Model-assisted stochastic learning for robotic applications |
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 |