CN109711527B - Robot control method based on particle swarm optimization algorithm - Google Patents
Robot control method based on particle swarm optimization algorithm Download PDFInfo
- Publication number
- CN109711527B CN109711527B CN201811594798.1A CN201811594798A CN109711527B CN 109711527 B CN109711527 B CN 109711527B CN 201811594798 A CN201811594798 A CN 201811594798A CN 109711527 B CN109711527 B CN 109711527B
- Authority
- CN
- China
- Prior art keywords
- particle
- robot
- control parameters
- parameters
- task
- 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.)
- Active
Links
Landscapes
- Feedback Control In General (AREA)
- Manipulator (AREA)
Abstract
The invention provides a robot control method based on a particle swarm optimization algorithm, which comprises the following steps: setting task parameters of each task; iterative training is performed by using a particle swarm optimization algorithm to generate optimal control parameters, including: initializing a particle swarm, and generating impedance control parameters by single particles; operating the real-time loop according to the control parameters to finish the operation task of the robot; according to the result of completing the manipulation task, obtaining the fitness of each particle through the detected control performance evaluation, and then updating the best position of the individual history and the best position of the particle group history, and the speed and the position of the particle; and if the global optimal position meets the minimum limit, outputting optimal control parameters to the task layer after finishing the iteration of the training layer. The invention has the advantages of simple and convenient use, light and quick training process and excellent maneuvering performance, and can overcome any of various robot maneuvering tasks.
Description
Technical Field
The invention relates to the technical field of industrial robots, in particular to a robot operating method based on a particle swarm optimization algorithm.
Background
After decades of rapid development of motion control of robots, both theoretical and practical, have tended to be mature and perfect, but no satisfactory solution has been available to date once the interaction of the robot with the environment has been involved, in particular when it is desired that the robot be able to grasp and manipulate objects freely. Having reliable steering capability is a necessary condition for the robot to actually leave the laboratory and enter the mass consumer market, which indicates that a practical and useful robotic steering technique is a very urgent and realistic requirement.
In the robot manipulation technique, the underlying control framework generally employs variable parameter impedance control, and the most critical and complex part thereof is how to select and determine the optimal control parameters.
The prior technical proposal mainly comprises the following two types:
1. manually adjusting control parameters, under the determined working condition, obtaining a group of control parameters capable of realizing specific operation functions through complex modeling and optimization, and when environmental characteristics change (such as the change of mass inertia of an operation target, the change of rigidity of a contact surface and the like), setting the parameters again.
2. Automatically adjusting control parameters, and automatically obtaining the control parameters under different working conditions through training by using a machine learning algorithm (such as deep learning, reinforcement learning and the like).
However, the main defects and shortcomings of the two technical methods are as follows: because of the numerous and mutually coupled control parameters, manual adjustment of the parameters requires a large amount of reserve knowledge, which is not only difficult for ordinary consumers and ordinary workers to accomplish, but also requires a great deal of labor and time costs for professional technicians, and is not possible to exhaust every working condition in reality. Because of the huge calculation scale of deep learning, the existing automatic parameter adjustment algorithm needs massive calculation power of a large-scale computer cluster when calculating the simplest working condition (such as shaft hole assembly), which is more difficult to realize in complex consumption scenes and industrial scenes.
Disclosure of Invention
The object of the present invention is to solve at least one of the technical drawbacks.
Therefore, the invention aims to provide a robot operating method based on a particle swarm optimization algorithm.
To achieve the above object, an embodiment of the present invention provides a robot manipulation method based on a particle swarm optimization algorithm, including:
step S1, informing a robot of tasks to be completed, namely setting constraint parameters of the tasks in each execution stage, wherein the steps comprise: position constraint and force constraint x of initial state init 、F init Position constraint, force constraint x, for an intermediate series of states exec1 、F exec1 、x exec2 、F exec2 … position constraint and force constraint in the completed state x fini 、F fini The method comprises the steps of carrying out a first treatment on the surface of the These constraints are used to determine the machineThe transfer conditions between each state of the robot in the operation process;
step S2, iterative training is carried out by utilizing a particle swarm optimization algorithm, and optimal control parameters are generated, wherein the step comprises the following steps:
initializing a particle swarm, and generating impedance control parameters from the particle positions;
performing robot impedance control according to the generated control parameters to complete the manipulation task;
examining the manipulation results corresponding to the particles to obtain the fitness of the particles, and updating the historical best position d pbest And group history best location d gbest And updating the speed and position of the particles;
if the global optimal position meets the minimum limit, the training layer iterates, the optimal control parameters are output to the task layer, if the termination condition is not met, the impedance parameters are regenerated from the updated particle state, and a new round of training is performed.
Further, in the step S2, the initializing the particle swarm includes: setting particles as n-dimensional vectors related to the number of the impedance control parameters, setting population scale to be about 25, setting the maximum speed to be within 20% of the parameter variation range, and randomly generating initial positions d of the particles init And velocity v init 。
Further, in said step S2, impedance control parameters α, β, γ are determined based on the state of each particle and the task parameters α 、γ β Desired trajectory parameter x d 、F d ；
Generating impedance control rates corresponding to the particles according to the impedance control parameters and the expected track parameters
Wherein F is ff Is self-adaptive feedforward moment; f (F) d The feedforward moment corresponding to the ideal track; k is impedance control stiffness; d is a damping matrix; e is the deviation; alpha, gamma α Learning factors and forgetting factors which are torque feedforward; beta, gamma β Learning factors and forgetting factors for stiffness; t is the step length of the controller; kappa is the adaptive tracking error; j is a robot jacobian matrix.
Further, taking the robot operation time as particle fitness, firstly calculating fitness of each particle at an initial position, selecting a position with highest fitness as a historical best position, and comparing the current fitness of the particle with the fitness of the historical best position in each iteration to update an individual historical best position d pbest With particle population history best location d gbest 。
Further, in the step S2, the speed and the position of the particle include:
v i ＝v i +c 1 ·R 1 ·(d pbest -d i )+c 2 ·R 1 ·(d gbest -d i ) (4)
d i ＝d i +v i (5)
in the formula, v i For particle velocity, d i For particle position, i.e. a multidimensional vector characterizing impedance control parameters, R 1 、R 2 A random number between 0 and 1, c representing the randomness of the movement of the particles 1 、c 2 And (3) the acceleration constant is used for respectively characterizing the influence of individual cognition and population cognition on particles.
According to the robot control method based on the particle swarm optimization algorithm, the strategy of automatically adjusting the control parameters is adopted, and the particle swarm optimization algorithm is utilized for carrying out robot impedance control parameter training to obtain the optimal control parameters. The implementation of the invention is divided into three levels: the outermost layer is a task layer, and different tasks determine different parameter constraints and transmit the different parameter constraints to the middle layer; the middle layer is a training layer, and each iteration calls a particle swarm optimization algorithm to calculate and obtain control parameters generated by the training of the round, and the control parameters are transmitted to the inner layer; the inner layer is a robot control real-time control loop, and the control task is completed by using the control parameters transmitted from the middle layer. According to the invention, only the task layer input interface is opened for a user, and the internal control algorithm and the parameter adjustment algorithm are deeply packaged, so that the method is simple and convenient to use, and only the task to be completed by the robot is required to be informed. The training process is light and quick, can be completed without huge hardware calculation force, and has the foundation of being realized in a real scene. And, the handling performance is excellent, can surpass arbitrary multiple robot manipulation task.
Additional aspects and advantages of the invention will be set forth in part in the description which follows and, in part, will be obvious from the description, or may be learned by practice of the invention.
Drawings
The foregoing and/or additional aspects and advantages of the invention will become apparent and may be better understood from the following description of embodiments taken in conjunction with the accompanying drawings in which:
fig. 1 is a flowchart of a robot manipulation method based on a particle swarm optimization algorithm according to an embodiment of the present invention;
fig. 2 is a schematic diagram of a robot manipulation method based on a particle swarm optimization algorithm according to an embodiment of the present invention.
Detailed Description
Embodiments of the present invention are described in detail below, examples of which are illustrated in the accompanying drawings, wherein like or similar reference numerals refer to like or similar elements or elements having like or similar functions throughout. The embodiments described below by referring to the drawings are illustrative and intended to explain the present invention and should not be construed as limiting the invention.
As shown in fig. 1-2, a robot manipulation method based on a particle swarm optimization algorithm according to an embodiment of the present invention includes the following steps:
step S1, informing the machineThe task that the person needs to complete, namely, set constraint parameters of the task in each execution stage, includes: position constraint and force constraint x of initial state init 、F init Position constraint, force constraint x, for an intermediate series of states exec1 、F exec1 、x exec2 、F exec2 … position constraint and force constraint in the completed state x fini 、F fini The method comprises the steps of carrying out a first treatment on the surface of the These constraints are used to determine the states in which the robot is in during operation and the transition conditions between the states.
Step S2, iterative training is carried out by utilizing a particle swarm optimization algorithm, and optimal control parameters are generated, wherein the step comprises the following steps:
step S21, initializing the particle swarm, and generating impedance control parameters from the particle positions.
Specifically, the particle swarm initialization is performed, including: setting particles as n-dimensional vectors related to the number of impedance control parameters, setting population scale, and determining upper and lower limits of particle positions and speeds according to task parameters, wherein the population scale can be about 25, and the maximum speed is within 20% of the parameter variation range. Then randomly generating the initial position d of the particle init And velocity v init 。
Determining impedance control parameters alpha, beta, gamma based on the state of each particle and the task parameters α 、γ β Desired trajectory parameter x d 、F d Etc.
The impedance control rate corresponding to each particle is generated by the impedance control parameter and the expected track parameter, and the specific steps are as follows:
wherein F is ff Is self-adaptive feedforward moment; f (F) d The feedforward moment corresponding to the ideal track; k is impedance control stiffness; d is a damping matrix; e is the deviation; alpha, gamma α Learning factors and forgetting factors which are torque feedforward; beta, gamma β Learning factors and forgetting factors for stiffness; t is the step length of the controller; kappa is the adaptive tracking error; j is a robot jacobian matrix.
And S22, performing robot impedance control according to the generated control parameters to finish the manipulation task.
Step S23, examining the manipulation results corresponding to each particle, and updating the individual historic best position d by taking the time required for completing the manipulation task as the fitness of the particle pbest And group history best location d gbest And updates the speed and position of the particles.
In the step, the robot operation time is taken as the particle fitness, the fitness of each particle under the initial position is calculated firstly, the position with the highest fitness is selected as the historical best position, and in each iteration, the current fitness of the particle is compared with the fitness of the historical best position so as to update the individual historical best position d pbest With particle population history best location d gbest 。
In addition, the speed and position of the particles are updated using the following formula:
v i ＝v i +c 1 ·R 1 ·(d pbest -d i )+c 2 ·R 1 ·(d gbest -d i ) (4)
d i ＝d i +v i (5)
in the formula, v i For particle velocity, d i For particle position, i.e. a multidimensional vector characterizing impedance control parameters, R 1 、R 2 A random number between 0 and 1, c representing the randomness of the movement of the particles 1 、c 2 And (3) the acceleration constant is used for respectively characterizing the influence of individual cognition and population cognition on particles.
And step S24, if the global optimal position meets the minimum limit, outputting optimal control parameters to a task layer after the iteration of the training layer is finished, and if the termination condition is not met, regenerating impedance parameters from the updated particle state and performing a new training round.
According to the robot control method based on the particle swarm optimization algorithm, the strategy of automatically adjusting the control parameters is adopted, and the particle swarm optimization algorithm is utilized for carrying out robot impedance control parameter training to obtain the optimal control parameters. The implementation of the invention is divided into three levels: the outermost layer is a task layer, and different tasks determine different parameter constraints and transmit the different parameter constraints to the middle layer; the middle layer is a training layer, and each iteration calls a particle swarm optimization algorithm to calculate and obtain control parameters generated by the training of the round, and the control parameters are transmitted to the inner layer; the inner layer is a robot control real-time control loop, and the control task is completed by using the control parameters transmitted from the middle layer. According to the invention, only the task layer input interface is opened for a user, and the internal control algorithm and the parameter adjustment algorithm are deeply packaged, so that the method is simple and convenient to use, and only the task to be completed by the robot is required to be informed. The training process is light and quick, can be completed without huge hardware calculation force, and has the foundation of being realized in a real scene. And, the handling performance is excellent, can surpass arbitrary multiple robot manipulation task.
In the description of the present specification, a description referring to terms "one embodiment," "some embodiments," "examples," "specific examples," or "some examples," etc., means that a particular feature, structure, material, or characteristic described in connection with the embodiment or example is included in at least one embodiment or example of the present invention. In this specification, schematic representations of the above terms do not necessarily refer to the same embodiments or examples. Furthermore, the particular features, structures, materials, or characteristics described may be combined in any suitable manner in any one or more embodiments or examples.
Although embodiments of the present invention have been shown and described above, it will be understood that the above embodiments are illustrative and not to be construed as limiting the invention, and that variations, modifications, alternatives, and variations may be made in the above embodiments by those skilled in the art without departing from the spirit and principles of the invention. The scope of the invention is defined by the appended claims and equivalents thereof.
Claims (3)
1. The robot control method based on the particle swarm optimization algorithm is characterized by comprising the following steps:
step S1, informing a robot of tasks to be completed, namely setting constraint parameters of the tasks in each execution stage, wherein the steps comprise: position constraint and force constraint x of initial state init 、F init Position constraint, force constraint x, for an intermediate series of states exec1 、F exec1 、x exec2 、F exec2 … position constraint and force constraint in the completed state x fini 、F fini These constraints are used to determine the states in which the robot is in during operation and the transition conditions between the states;
step S2, iterative training is carried out by utilizing a particle swarm optimization algorithm, and optimal control parameters are generated, wherein the step comprises the following steps:
initializing a particle swarm, and generating impedance control parameters from the particle positions; wherein the impedance control parameters alpha, beta, gamma are determined according to the state of each particle and the task parameters α 、γ β Desired trajectory parameter x d 、F d ；
Generating impedance control rates corresponding to the particles according to the impedance control parameters and the expected track parameters
Wherein F is ff Is self-adaptive feedforward moment; f (F) d The feedforward moment corresponding to the ideal track; k is impedance control stiffness; d is a damping matrix; e is the deviation; alpha, gamma α Learning factors and forgetting factors which are torque feedforward; beta, gamma β Learning factors and forgetting factors for stiffness; t is the step length of the controller; kappa is the adaptive tracking error; j is a robot jacobian matrix;
performing robot impedance control according to the generated control parameters to complete the manipulation task; the robot body is controlled in real time by utilizing the generated impedance control rate, and the real-time loop is controlled according to the control parameters to complete the control task of the robot;
examining the manipulation results corresponding to the particles to obtain the fitness of the particles, and updating the historical best position d pbest And group history best location d gbest And updating the speed and position of the particles;
the speed and the position of the particles are updated through the following formula, so that updated robot impedance control parameters are obtained:
v i ＝v i +c 1 ·R 1 ·(d pbest -d i )+c 2 ·R 1 ·(d gbest -d i )
d i ＝d i +v i
in the formula, v i For particle velocity, d i For particle position, i.e. a multidimensional vector characterizing impedance control parameters, R 1 、R 2 A random number between 0 and 1, c representing the randomness of the movement of the particles 1 、c 2 The acceleration constant is used for respectively representing the influence of individual cognition and group cognition on particles;
if the global optimal position meets the minimum limit, the training layer iterates, the optimal control parameters are output to the task layer, if the termination condition is not met, the impedance parameters are regenerated from the updated particle state, and a new round of training is performed.
2. The robot manipulating method based on the particle swarm optimization algorithm according to claim 1, wherein in said step S2, said initializing the particle swarm comprises: setting particles as n-dimensional vectors related to the number of the impedance control parameters, setting population scale to be about 25, setting the maximum speed to be within 20% of the parameter variation range, and randomly generating initial positions d of the particles init And velocity v init 。
3. The robot manipulating method based on particle swarm optimization according to claim 1, wherein in said step S2, the fitness of each particle in the initial position is first calculated using the robot manipulating time as the fitness of the particle, the position with the highest fitness is selected as the historic best position, and in each iteration, the fitness of the current particle is compared with the fitness of the historic best position to update the individual historic best position d pbest With particle population history best location d gbest 。
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 CN109711527A (en) | 2019-05-03 |
| CN109711527B true 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) |
Families Citing this family (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| EP3771522A1 (en) * | 2019-07-30 | 2021-02-03 | Siemens Aktiengesellschaft | Method and manipulation system for manipulating an object by a robot with vector fields |
| CN113985864B (en) * | 2021-08-17 | 2024-06-14 | 北京空间飞行器总体设计部 | Dynamic walking four-foot detection robot and control method |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| 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 |
| 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 |
Family Cites Families (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| EP1974869A1 (en) * | 2007-03-26 | 2008-10-01 | 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 |
| CN103869704B (en) * | 2014-04-08 | 2016-05-25 | 哈尔滨工业大学 | Based on the robot for space star arm control method for coordinating of expansion Jacobian matrix |
| CN108068113B (en) * | 2017-11-13 | 2020-06-12 | 苏州大学 | 7-DOF humanoid arm flying object operation minimum acceleration trajectory optimization |
-
2018
- 2018-12-25 CN CN201811594798.1A patent/CN109711527B/en active Active
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| 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 |
| 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 |
Also Published As
| Publication number | Publication date |
|---|---|
| CN109711527A (en) | 2019-05-03 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110238839B (en) | Multi-shaft-hole assembly control method for optimizing non-model robot by utilizing environment prediction | |
| CN111496792B (en) | Method and system for tracking and controlling input saturation fixed time trajectory of mechanical arm | |
| CN105956297B (en) | Comprehensive evaluation and optimization method for redundant robot motion flexibility performance | |
| CN105772917B (en) | A kind of three joint spot welding robot's Trajectory Tracking Control methods | |
| CN109676610B (en) | Circuit breaker assembly robot and method for realizing work track optimization | |
| CN109901403A (en) | A kind of face autonomous underwater robot neural network S control method | |
| CN108115681A (en) | Learning by imitation method, apparatus, robot and the storage medium of robot | |
| CN106406098B (en) | A kind of man-machine interaction control method of robot system under circumstances not known | |
| Kartoun et al. | A human-robot collaborative reinforcement learning algorithm | |
| Lu et al. | Design and optimization of interval type-2 fuzzy logic controller for delta parallel robot trajectory control | |
| CN103984230B (en) | A kind of space manipulator pedestal zero disturbs optimal control method | |
| CN109711527B (en) | Robot control method based on particle swarm optimization algorithm | |
| CN110450156A (en) | The optimum design method of multi-freedom Mechanism adaptive fuzzy controller | |
| CN108527372A (en) | A kind of joint of robot self-adaptation control method of variation rigidity series elastic driver | |
| CN117103282B (en) | Double-arm robot cooperative motion control method based on MATD3 algorithm | |
| WO2018227820A1 (en) | Method and device for controlling manipulator movement, storage medium, and terminal device | |
| CN111702767A (en) | Manipulator impedance control method based on inversion fuzzy self-adaptation | |
| CN112936286B (en) | Self-adaptive consistency tracking control method and system for multi-flexible mechanical arm system | |
| CN115366099A (en) | Mechanical arm depth certainty strategy gradient training method based on forward kinematics | |
| CN113219825B (en) | Single-leg track tracking control method and system for four-leg robot | |
| Hu et al. | Robot time optimal trajectory planning based on improved simplified particle swarm optimization algorithm | |
| Wu et al. | Optimal trajectory planning of a flexible dual-arm space robot with vibration reduction | |
| Wang et al. | Path planning optimization for teaching and playback welding robot | |
| Jawad et al. | Design of adaptive controller for robot arm manipulator based on ANN with optimized PID by IWO algorithm | |
| CN110703792B (en) | Underwater robot attitude control method based on reinforcement learning |
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 |