CN114282636A - Robot control model establishing method and robot control method - Google Patents
Robot control model establishing method and robot control method Download PDFInfo
- Publication number
- CN114282636A CN114282636A CN202111364110.2A CN202111364110A CN114282636A CN 114282636 A CN114282636 A CN 114282636A CN 202111364110 A CN202111364110 A CN 202111364110A CN 114282636 A CN114282636 A CN 114282636A
- Authority
- CN
- China
- Prior art keywords
- robot control
- robot
- weight
- optimized
- learning machine
- 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
Links
- 238000000034 method Methods 0.000 title claims abstract description 59
- 239000002245 particle Substances 0.000 claims abstract description 85
- 238000013528 artificial neural network Methods 0.000 claims abstract description 49
- 230000006399 behavior Effects 0.000 claims abstract description 38
- 238000004422 calculation algorithm Methods 0.000 claims abstract description 36
- 238000012549 training Methods 0.000 claims abstract description 24
- 230000009471 action Effects 0.000 claims description 15
- 238000004590 computer program Methods 0.000 claims description 3
- 238000005457 optimization Methods 0.000 description 15
- 230000006870 function Effects 0.000 description 9
- 230000002441 reversible effect Effects 0.000 description 6
- 238000010586 diagram Methods 0.000 description 5
- 239000011159 matrix material Substances 0.000 description 5
- 230000008569 process Effects 0.000 description 5
- 230000008859 change Effects 0.000 description 4
- 238000011217 control strategy Methods 0.000 description 4
- 230000007613 environmental effect Effects 0.000 description 4
- 230000008447 perception Effects 0.000 description 4
- 238000011160 research Methods 0.000 description 4
- 238000004364 calculation method Methods 0.000 description 3
- 230000004927 fusion Effects 0.000 description 3
- 238000012545 processing Methods 0.000 description 3
- 238000011161 development Methods 0.000 description 2
- 238000012423 maintenance Methods 0.000 description 2
- 239000000463 material Substances 0.000 description 2
- 230000007246 mechanism Effects 0.000 description 2
- 230000004075 alteration Effects 0.000 description 1
- 230000003542 behavioural effect Effects 0.000 description 1
- 238000006243 chemical reaction Methods 0.000 description 1
- 238000013135 deep learning Methods 0.000 description 1
- 230000002431 foraging effect Effects 0.000 description 1
- 238000010801 machine learning Methods 0.000 description 1
- 230000005012 migration Effects 0.000 description 1
- 238000013508 migration Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 230000009467 reduction Effects 0.000 description 1
- 238000012216 screening Methods 0.000 description 1
- 238000010845 search algorithm Methods 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
Images
Landscapes
- Manipulator (AREA)
- Feedback Control In General (AREA)
Abstract
The application provides a robot control model establishing method and a robot control method, wherein the method comprises the following steps: acquiring a robot control data sample; acquiring initial weight and initial threshold of the neural network of the extreme learning machine; optimizing the initial weight and the initial threshold value by utilizing a particle swarm PSO algorithm to obtain the optimized weight and the optimized threshold value; and training the extreme learning machine neural network by using the robot control data sample, the optimized weight and the optimized threshold value to obtain a robot control model. The technical problem of low execution efficiency of behavior control of the robot in the prior art is solved.
Description
Technical Field
The application belongs to the field of computer software, and particularly relates to a robot model establishing method, a robot control method, computer equipment and a computer readable storage medium.
Background
The robot technology is developed rapidly, the application scene is more refined, the functional requirements are higher and higher, and the requirements on control precision, real-time feedback and the like are obviously improved. With the gradual maturity of the mode identification technology development and the deep learning algorithm, the robot control comes with new development opportunities, the traditional passive acceptance instruction and the gradual robot control mode are being changed, and the semi-automatic and full-automatic autonomous learning are gradually embedded into the robot control system. Currently, both in the aspect of robot application environment perception and in the aspects of action and behavior diversification and control method refinement, more advanced technical support is provided, for example, the robot can realize more complex and accurate behavior operation through diverse weighted fusion of instruction operation, sensitively perceive the surrounding environment, can realize function application in a complex environment, and the like.
At present, much control research is carried out on robots, and mainly focuses on environmental perception research at the front end and positioning and control algorithm strategy research at the rear end.
It should be noted that the behavior control execution efficiency of the robot in the related art is low.
Disclosure of Invention
In order to overcome the problems in the related art at least to a certain extent, the application provides a robot control model establishing method, a robot control method, a computer device and a storage medium, which can solve the technical problem of low execution efficiency of behavior control of a robot in the prior art.
In order to achieve the purpose, the following technical scheme is adopted in the application:
according to an aspect of the present application, there is provided a method of establishing a robot control model, including: acquiring a robot control data sample; acquiring initial weight and initial threshold of the neural network of the extreme learning machine; optimizing the initial weight and the initial threshold value by utilizing a particle swarm PSO algorithm to obtain the optimized weight and the optimized threshold value; and training the extreme learning machine neural network by using the robot control data sample, the optimized weight and the optimized threshold value to obtain a robot control model.
Optionally, the training the extreme learning machine neural network by using the robot control data sample, the optimized weight, and the optimized threshold to obtain a robot control model, including: based on the optimized weight and the optimized threshold, solving an optimal weight and an optimal threshold by using an extreme learning machine algorithm; and training the neural network of the extreme learning machine by using the robot control data sample, the optimal weight and the optimal threshold value to obtain the robot control model.
Optionally, the training the extreme learning machine neural network by using the robot control data sample, the optimal weight, and the optimal threshold to obtain the robot control model includes: inputting the robot control data sample, the optimal weight and the optimal threshold value into the neural network of the extreme learning machine to obtain a prediction accuracy rate; and updating the parameters of the neural network of the extreme learning machine based on the prediction accuracy rate to obtain the robot control model.
Optionally, the optimizing the initial weight and the initial threshold by using a particle swarm PSO algorithm to obtain an optimized weight and an optimized threshold includes: establishing a particle swarm according to the initial weight and the initial threshold; calculating the fitness value of each particle in the particle swarm; updating the parameters of each particle by using the fitness value of each particle; and acquiring the optimized weight and the optimized threshold value based on the updated parameters of the particles.
Optionally, the updating the parameter of each particle by using the fitness value of each particle includes: acquiring the update execution times of the parameters for updating each particle by using the fitness value of each particle; and finishing updating the parameters of the particles when the updating execution times reach a time threshold value.
Optionally, the parameter of each particle includes at least one of: an extremum value; the speed of movement; the position of movement.
Optionally, the obtaining the initial weight and the initial threshold of the extreme learning machine neural network includes: and randomly acquiring initial weight and initial threshold of the neural network of the extreme learning machine.
According to another aspect of the present application, there is also provided a control method of a robot, including: acquiring a robot control model based on any one of the establishment methods of the robot control model in the first aspect; and performing control on the target robot based on the robot control model.
Optionally, the performing control on the target robot based on the robot control model includes: acquiring target weights corresponding to various actions of the robot based on the robot control model; weighting and summing all the behavior actions of the robot based on all the target weights to obtain behavior results; and executing control on the target robot according to the behavior result.
According to another aspect of the present application, there is also provided an apparatus for establishing a control model of a robot, including: the system comprises a sample acquisition unit, a data acquisition unit and a data processing unit, wherein the sample acquisition unit is used for acquiring a robot control data sample; the initial weight and initial threshold value acquisition unit is used for acquiring the initial weight and the initial threshold value of the neural network of the extreme learning machine; the optimization unit is used for optimizing the initial weight and the initial threshold value by utilizing a particle swarm PSO algorithm to obtain the optimized weight and the optimized threshold value; and the training unit is used for training the neural network of the extreme learning machine by using the robot control data sample, the optimized weight and the optimized threshold value to obtain a robot control model.
Optionally, the training unit includes: the optimal weight and optimal threshold solving module is used for solving the optimal weight and the optimal threshold by utilizing an extreme learning machine algorithm based on the optimized weight and the optimized threshold; and the training module is used for training the neural network of the extreme learning machine by using the robot control data sample, the optimal weight and the optimal threshold value to obtain the robot control model.
Optionally, the training module includes: the prediction accuracy rate obtaining sub-module is used for inputting the robot control data sample, the optimal weight and the optimal threshold value into the extreme learning machine neural network to obtain the prediction accuracy rate; and the updating submodule is used for updating the parameters of the neural network of the extreme learning machine based on the prediction accuracy rate so as to obtain the robot control model.
Optionally, the optimizing unit includes: the particle swarm establishing module is used for establishing a particle swarm according to the initial weight and the initial threshold; the calculation module is used for calculating the fitness value of each particle in the particle swarm; a particle parameter updating module for updating the parameter of each particle by using the fitness value of each particle; and the optimization module is used for acquiring the optimized weight and the optimized threshold value based on the updated parameters of the particles.
Optionally, the particle parameter updating module includes: an update execution time obtaining module, configured to obtain the update execution time for updating the parameter of each particle by using the fitness value of each particle; and the updating control module is used for finishing updating the parameters of the particles under the condition that the updating execution times reaches a time threshold.
Optionally, the parameter of each particle includes at least one of: an extremum value; the speed of movement; the position of movement.
Optionally, the initial weight and initial threshold obtaining unit includes: and the random acquisition module is used for randomly acquiring the initial weight and the initial threshold of the neural network of the extreme learning machine.
According to another aspect of the present application, there is also provided a control apparatus of a robot, including: a robot control model acquisition unit configured to acquire a robot control model based on any one of the robot control model establishment apparatuses according to the third aspect; and the robot control unit is used for controlling the target robot based on the robot control model.
Optionally, the robot control unit includes: the target weight acquisition module is used for acquiring target weights corresponding to various actions of the robot based on the robot control model; the weighted summation module is used for carrying out weighted summation on each behavior action of the robot based on each target weight to obtain a behavior result; and the robot control module is used for controlling the target robot according to the behavior result.
According to another aspect of the present application, there is also provided a computer device comprising a memory and a processor, the memory having stored thereon computer instructions which, when executed by the processor, cause the method of any one of the first aspect to be performed.
According to another aspect of the present application, there is also provided a non-transitory computer readable storage medium having stored thereon a computer program which, when executed by a processor, causes the method of any of the first aspects to be performed.
The application provides a method for establishing a robot control model, which comprises the following steps: acquiring a robot control data sample; acquiring initial weight and initial threshold of the neural network of the extreme learning machine; optimizing the initial weight and the initial threshold value by utilizing a particle swarm PSO algorithm to obtain the optimized weight and the optimized threshold value; and training the extreme learning machine neural network by using the robot control data sample, the optimized weight and the optimized threshold value to obtain a robot control model. The technical problem of low execution efficiency of behavior control of the robot in the prior art is solved.
It is to be understood that both the foregoing general description and the following detailed description are exemplary and explanatory only and are not restrictive of the application.
Drawings
In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings used in the description of the embodiments or the prior art will be briefly described below, it is obvious that the drawings in the following description are only some embodiments of the present application, and for those skilled in the art, other drawings can be obtained according to the drawings without creative efforts.
FIG. 1 is a flow diagram illustrating a method for building a robot control model in accordance with an exemplary embodiment;
FIG. 2 is a schematic diagram illustrating robot control system functional modules in accordance with an exemplary embodiment;
FIG. 3 is a schematic diagram illustrating a robot reactive control architecture according to an exemplary embodiment;
FIG. 4 is a schematic diagram illustrating a robot control flow according to an exemplary embodiment;
fig. 5 is a schematic diagram illustrating an apparatus for building a robot control model according to an exemplary embodiment.
Detailed Description
In order to make the technical solutions better understood by those skilled in the art, the technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the drawings in the embodiments of the present application, and it is obvious that the described embodiments are only partial embodiments of the present application, but not all embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present application.
It should be noted that the terms "first," "second," and the like in the description and claims of this application and in the drawings described above are used for distinguishing between similar elements and not necessarily for describing a particular sequential or chronological order. It is to be understood that the data so used is interchangeable under appropriate circumstances such that the embodiments of the application described herein are capable of operation in sequences other than those illustrated or described herein. Furthermore, the terms "comprises," "comprising," and "having," and any variations thereof, are intended to cover a non-exclusive inclusion, such that a process, method, system, article, or apparatus that comprises a list of steps or elements is not necessarily limited to those steps or elements expressly listed, but may include other steps or elements not expressly listed or inherent to such process, method, article, or apparatus.
As shown in fig. 1, the present application provides a method for establishing a robot control model, the method comprising:
step S11, a robot control data sample is obtained.
Specifically, a server or other hardware devices with a data processing function may be used as an execution subject of the method of the present invention, the present solution may acquire a robot control data sample, where the robot control data sample may include a large number of control instructions (behavior instructions, such as target coordinates and action behaviors) for the robot, obstacle information (such as obstacle coordinates), and the like, and an execution result (for example, accuracy of obstacle avoidance and accuracy of approaching to a target, and the like).
In step S13, the initial weight and initial threshold of the extreme learning machine neural network are obtained.
Specifically, an Extreme Learning Machine (ELM) is a Machine Learning system or method constructed based on a feed Forward Neural Network (FNN), and is suitable for supervised Learning and unsupervised Learning problems.
Alternatively, the initial weights and initial thresholds for the extreme learning machine neural network may be randomly obtained.
Optionally, the extreme learning machine neural network takes the coordinates of the obstacle, the coordinates of the target, the action behavior and the like as inputs, and takes the accuracy of obstacle avoidance and the accuracy of approaching the target as outputs.
And step S15, optimizing the initial weight and the initial threshold value by utilizing a particle swarm PSO algorithm to obtain the optimized weight and the optimized threshold value.
The particle swarm optimization is a global random search algorithm based on swarm intelligence and provided by simulating migration and clustering behaviors in the foraging process of a bird swarm. The PSO is initialized to a population of random particles (random solution). The optimal solution is then found by iteration. In each iteration, the particle updates itself by tracking two "extrema". The first is the optimal solution found by the particle itself, which is called the individual extremum. The other extreme is the best solution currently found for the entire population, which is the global extreme.
And step S17, training the extreme learning machine neural network by using the robot control data sample, the optimized weight and the optimized threshold value to obtain a robot control model.
In the embodiment of the disclosure, the robot obtains environmental characteristics through a sensor, builds an environmental model, defines a target task, generates a behavior control instruction, and finally executes the behavior instruction through an action executor. Alternatively, in this embodiment, the robot function module may be simply divided into 6 parts according to the workflow, as shown in fig. 2. Sensor perception data fusion and motion planning and control are key contents of robot control research and generally need to be achieved through algorithms, the former need to perform feature screening, dimension reduction, fusion and the like on multi-sensor perception environment data through the algorithms, and the latter need to achieve path selection and accurate control through the algorithms.
In this embodiment of the disclosure, optionally, the robot control data sample, the optimized weight, and the optimized threshold are used to train the extreme learning machine neural network to obtain a robot control model, which may be implemented in the following manner: based on the optimized weight and the optimized threshold, solving an optimal weight and an optimal threshold by using an extreme learning machine algorithm; and training the neural network of the extreme learning machine by using the robot control data sample, the optimal weight and the optimal threshold value to obtain the robot control model.
In this embodiment, the robot makes behavioral feedback based on the perceived environmental characteristics, as shown in fig. 3. In the accurate control of the robot, the control strategy is a control strategy of superposition and summation of a plurality of behaviors, such as obstacle avoidance, attitude maintenance, target tendency and the like. The command executed by the execution mechanism is obtained after weighting through various behaviors, the weight is closely related to the environment of the robot and can change along with the change of the environment, and the weight values of various behaviors need to be accurately calculated to finish the efficient control of the robot, so that the robot can be ensured to smoothly run in a complex and changeable environment. The present embodiment employs an extreme learning machine to realize the calculation of the weight.
Preferably, the weight and the threshold optimal solution can be solved by means of a reversible matrix based on the extreme learning machine and according to the minimum norm two-times solution theorem, and finally, a stable extreme learning machine robot control model is obtained.
In this embodiment of the present disclosure, optionally, the robot control data sample, the optimal weight, and the optimal threshold are used to train the extreme learning machine neural network to obtain the robot control model, which may be implemented in the following manner: inputting the robot control data sample, the optimal weight and the optimal threshold value into the neural network of the extreme learning machine to obtain a prediction accuracy rate; and updating the parameters of the neural network of the extreme learning machine based on the prediction accuracy rate to obtain the robot control model.
In the embodiment of the present disclosure, optionally, the initial weight and the initial threshold are optimized by using a particle swarm PSO algorithm to obtain an optimized weight and an optimized threshold, which may be implemented by the following steps: establishing a particle swarm according to the initial weight and the initial threshold; calculating the fitness value of each particle in the particle swarm; updating the parameters of each particle by using the fitness value of each particle; and acquiring the optimized weight and the optimized threshold value based on the updated parameters of the particles.
Since the algorithm time is increased in the case where the weight and the threshold of the extreme learning are randomly set, it is considered to use the particle swarm algorithm for the optimization of the initial values of the weight and the threshold to improve the speed of the extreme learning machine by the optimization.
In the embodiment, the various behavior control of the robot is realized by adopting an extreme learning machine algorithm, and the initial value optimization is performed on the weight and the threshold of the extreme learning machine by combining a Particle Swarm Optimization (PSO) algorithm, so that the control efficiency of the extreme learning machine is effectively improved.
In this embodiment of the present disclosure, optionally, updating the parameter of each particle by using the fitness value of each particle may be implemented by: acquiring the update execution times of the parameters for updating each particle by using the fitness value of each particle; and finishing updating the parameters of the particles when the updating execution times reach a time threshold value.
In this embodiment of the present disclosure, optionally, the parameter of each particle includes at least one of: an extremum value; the speed of movement; the position of movement.
According to the method, the establishment of the robot control model is realized by adopting a PSO (Power System optimization) optimized extreme learning machine algorithm, the single-hidden-layer neural network robot control model is established, and the weight and the threshold of the extreme learning machine are optimized by adopting the PSO algorithm. The robot control trend based on the PSO optimization extreme learning machine is high in target accuracy, and the PSO extreme learning machine algorithm optimizes the weight and the initial threshold value, so that better convergence time can be obtained, and the convergence speed is high. In addition, preferably, according to the minimum norm two-times solution theorem, the weight and the threshold value optimal solution are solved by means of the reversible matrix, and a stable extreme learning robot control model can be obtained.
The present disclosure also provides a method for establishing a robot control model, as shown in fig. 4, the method includes:
1. acquiring a robot control data sample set;
2. establishing a single hidden layer neural network, and selecting a Gaussian function as a conversion function;
3. randomly setting weight and threshold of an extreme learning neural network;
4. constructing a particle swarm by using randomly set weights and threshold values, and carrying out initial value encoding;
5. the prediction accuracy rate obtained by training the extreme learning machine is used as a fitness function;
6. calculating the fitness value of each particle;
7. updating the extreme value of the particle, and updating the speed and the position of the particle;
8. judging whether the iteration times reach the maximum iteration times or not;
9. under the condition that the iteration times reach the maximum iteration times, outputting the weight and the threshold value after particle swarm optimization; under the condition that the iteration times do not reach the maximum iteration times, re-executing the step 6-8;
10. solving the optimal weight and threshold value by adopting an extreme learning machine and a reversible matrix;
11. whether the prediction accuracy rate meets the requirement or not;
12. and under the condition that the accuracy reaches the requirement, obtaining a stable robot control model, and under the condition that the accuracy does not reach the requirement, re-executing the steps 5-11 until the accuracy reaches the requirement.
In this embodiment, behavior control of the robot is realized by using a PSO-optimized extreme learning machine algorithm, a single hidden layer neural network robot control model is established, the PSO algorithm is used to optimize the weight and the threshold of the extreme learning machine, the optimal solution of the weight and the threshold is solved by using a reversible matrix according to the minimum norm-two-times solution theorem, and finally, a stable extreme learning machine robot control model can be obtained. And then can get higher robot through the control model of this robot and tend to the target accuracy, and improve the performance to robot control effectively.
The present disclosure also provides a control method of a robot, the method including:
step S501, a robot control model is obtained based on the method for establishing a robot control model according to any of the above embodiments.
In this step, the method for obtaining the robot control model is described in detail in the above embodiments, and is not described herein again.
Step S502, the target robot is controlled based on the robot control model.
Optionally, the target robot is controlled based on the robot control model, and the control may be implemented by: acquiring target weights corresponding to various actions of the robot based on the robot control model; weighting and summing all the behavior actions of the robot based on all the target weights to obtain behavior results; and executing control on the target robot according to the behavior result.
In the embodiment, in the precise control of the robot, the control strategy is a control strategy of superposition and summation of a plurality of behaviors, such as obstacle avoidance, attitude maintenance, target tendency and the like. The command executed by the execution mechanism is obtained after weighting through various behaviors, the weight is closely related to the environment of the robot and can change along with the change of the environment, and the weight values of various behaviors need to be accurately calculated to finish the efficient control of the robot, so that the robot can be ensured to smoothly run in a complex and changeable environment. In this embodiment, the weight of the corresponding behavior is obtained by solving with an extreme learning machine, and finally, the action result is obtained by weighted summation.
The embodiment of the disclosure adopts the extreme learning machine algorithm to realize the various behavior control of the robot, and combines the particle swarm algorithm to perform initial value optimization on the weight and the threshold value of the extreme learning machine, thereby effectively improving the control efficiency of the extreme learning machine and realizing the high-efficiency and accurate control of the robot.
As shown in fig. 5, the present disclosure also provides an apparatus for building a robot control model, which is configured to perform any one of the methods for building a robot control model described above. The device includes:
a sample obtaining unit 60 for obtaining a robot control data sample;
an initial weight and initial threshold value obtaining unit 62 for obtaining an initial weight and an initial threshold value of the extreme learning machine neural network;
an optimizing unit 64, configured to optimize the initial weight and the initial threshold by using a particle swarm PSO algorithm, so as to obtain an optimized weight and an optimized threshold;
and the training unit 66 is configured to train the extreme learning machine neural network by using the robot control data sample, the optimized weight, and the optimized threshold value, so as to obtain a robot control model.
Optionally, the training unit includes: the optimal weight and optimal threshold solving module is used for solving the optimal weight and the optimal threshold by utilizing an extreme learning machine algorithm based on the optimized weight and the optimized threshold; and the training module is used for training the neural network of the extreme learning machine by using the robot control data sample, the optimal weight and the optimal threshold value to obtain the robot control model.
Optionally, the training module includes: the prediction accuracy rate obtaining sub-module is used for inputting the robot control data sample, the optimal weight and the optimal threshold value into the extreme learning machine neural network to obtain the prediction accuracy rate; and the updating submodule is used for updating the parameters of the neural network of the extreme learning machine based on the prediction accuracy rate so as to obtain the robot control model.
Optionally, the optimizing unit includes: the particle swarm establishing module is used for establishing a particle swarm according to the initial weight and the initial threshold; the calculation module is used for calculating the fitness value of each particle in the particle swarm; a particle parameter updating module for updating the parameter of each particle by using the fitness value of each particle; and the optimization module is used for acquiring the optimized weight and the optimized threshold value based on the updated parameters of the particles.
Optionally, the particle parameter updating module includes: an update execution time obtaining module, configured to obtain the update execution time for updating the parameter of each particle by using the fitness value of each particle; and the updating control module is used for finishing updating the parameters of the particles under the condition that the updating execution times reaches a time threshold.
Optionally, the parameter of each particle includes at least one of: an extremum value; the speed of movement; the position of movement.
Optionally, the initial weight and initial threshold obtaining unit includes: and the random acquisition module is used for randomly acquiring the initial weight and the initial threshold of the neural network of the extreme learning machine.
According to the method, the establishment of the robot control model is realized by adopting a PSO (Power System optimization) optimized extreme learning machine algorithm, the single-hidden-layer neural network robot control model is established, and the weight and the threshold of the extreme learning machine are optimized by adopting the PSO algorithm. The robot control trend based on the PSO optimization extreme learning machine is high in target accuracy, and the PSO extreme learning machine algorithm optimizes the weight and the initial threshold value, so that better convergence time can be obtained, and the convergence speed is high. In addition, preferably, according to the minimum norm two-times solution theorem, the weight and the threshold value optimal solution are solved by means of the reversible matrix, and a stable extreme learning robot control model can be obtained.
The present disclosure also provides a control apparatus for a robot, which is configured to perform any one of the above-described methods for controlling a robot. The device includes:
a robot control model obtaining unit 70, configured to obtain a robot control model by the robot control model establishing apparatus according to any one of the above embodiments;
and a robot control unit 72 for performing control of the target robot based on the robot control model.
Optionally, the robot control unit includes: the target weight acquisition module is used for acquiring target weights corresponding to various actions of the robot based on the robot control model; the weighted summation module is used for carrying out weighted summation on each behavior action of the robot based on each target weight to obtain a behavior result; and the robot control module is used for controlling the target robot according to the behavior result.
The embodiment of the disclosure adopts the extreme learning machine algorithm to realize the various behavior control of the robot, and combines the particle swarm algorithm to perform initial value optimization on the weight and the threshold value of the extreme learning machine, thereby effectively improving the control efficiency of the extreme learning machine and realizing the high-efficiency and accurate control of the robot.
The present disclosure also provides a computer device comprising a memory and a processor, the memory having stored thereon computer instructions that, when executed by the processor, cause any of the methods described above to be performed.
The present disclosure also provides a non-transitory computer readable storage medium having stored thereon a computer program which, when executed by a processor, causes the method of any of the above to be performed.
It is understood that the same or similar parts in the above embodiments may be mutually referred to, and the same or similar parts in other embodiments may be referred to for the content which is not described in detail in some embodiments.
It should be noted that, in the description of the present application, the terms "first", "second", etc. are used for descriptive purposes only and are not to be construed as indicating or implying relative importance. In addition, in the description of the present application, the meaning of "plurality" means at least two unless otherwise specified.
It will be understood that when an element is referred to as being "secured to" or "disposed on" another element, it can be directly on the other element or intervening elements may also be present; when an element is referred to as being "connected" to another element, it can be directly connected to the other element or intervening elements may also be present, and further, as used herein, connected may include wirelessly connected; the term "and/or" is used to include any and all combinations of one or more of the associated listed items.
Any process or method descriptions in flow charts or otherwise described herein may be understood as: represents modules, segments or portions of code which include one or more executable instructions for implementing specific logical functions or steps of a process, and the scope of the preferred embodiments of the present application includes other implementations in which functions may be executed out of order from that shown or discussed, including substantially concurrently or in reverse order, depending on the functionality involved, as would be understood by those reasonably skilled in the art of the embodiments of the present application.
It should be understood that portions of the present application may be implemented in hardware, software, firmware, or a combination thereof. In the above embodiments, the various steps or methods may be implemented in software or firmware stored in memory and executed by a suitable instruction execution system. For example, if implemented in hardware, as in another embodiment, any one or combination of the following techniques, which are known in the art, may be used: a discrete logic circuit having a logic gate circuit for implementing a logic function on a data signal, an application specific integrated circuit having an appropriate combinational logic gate circuit, a Programmable Gate Array (PGA), a Field Programmable Gate Array (FPGA), or the like.
It will be understood by those skilled in the art that all or part of the steps carried by the method for implementing the above embodiments may be implemented by hardware related to instructions of a program, and the program may be stored in a computer readable storage medium, and when executed, the program includes one or a combination of the steps of the method embodiments.
In addition, functional units in the embodiments of the present application may be integrated into one processing module, or each unit may exist alone physically, or two or more units are integrated into one module. The integrated module can be realized in a hardware mode, and can also be realized in a software functional module mode. The integrated module may also be stored in a computer-readable storage medium if it is implemented in the form of a software functional module and sold or used as a separate product.
The storage medium mentioned above may be a read-only memory, a magnetic or optical disk, etc.
In the description herein, reference to the description of the term "one embodiment," "some embodiments," "an example," "a specific example," 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 application. In this specification, the schematic representations of the terms used above do not necessarily refer to the same embodiment or example. 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 application have been shown and described above, it is understood that the above embodiments are exemplary and should not be construed as limiting the present application, and that variations, modifications, substitutions and alterations may be made to the above embodiments by those of ordinary skill in the art within the scope of the present application.
Claims (11)
1. A method for establishing a robot control model is characterized by comprising the following steps:
acquiring a robot control data sample;
acquiring initial weight and initial threshold of the neural network of the extreme learning machine;
optimizing the initial weight and the initial threshold value by utilizing a particle swarm PSO algorithm to obtain the optimized weight and the optimized threshold value;
and training the extreme learning machine neural network by using the robot control data sample, the optimized weight and the optimized threshold value to obtain a robot control model.
2. The method of claim 1, wherein the training the extreme learning machine neural network using the robot control data samples, the optimized weights, and the optimized thresholds to obtain a robot control model comprises:
based on the optimized weight and the optimized threshold, solving an optimal weight and an optimal threshold by using an extreme learning machine algorithm;
and training the neural network of the extreme learning machine by using the robot control data sample, the optimal weight and the optimal threshold value to obtain the robot control model.
3. The method of claim 2, wherein the training the extreme learning machine neural network using the robot control data samples, the optimal weights, and the optimal thresholds to obtain the robot control model comprises:
inputting the robot control data sample, the optimal weight and the optimal threshold value into the neural network of the extreme learning machine to obtain a prediction accuracy rate;
and updating the parameters of the neural network of the extreme learning machine based on the prediction accuracy rate to obtain the robot control model.
4. The method of claim 1, wherein the optimizing the initial weights and initial thresholds using a particle swarm PSO algorithm to obtain optimized weights and optimized thresholds comprises:
establishing a particle swarm according to the initial weight and the initial threshold;
calculating the fitness value of each particle in the particle swarm;
updating the parameters of each particle by using the fitness value of each particle;
and acquiring the optimized weight and the optimized threshold value based on the updated parameters of the particles.
5. The method of claim 4, wherein the updating the parameter of each particle with the fitness value of each particle comprises:
acquiring the update execution times of the parameters for updating each particle by using the fitness value of each particle;
and finishing updating the parameters of the particles when the updating execution times reach a time threshold value.
6. The method according to any one of claims 4 to 5, wherein the parameter of each particle comprises at least one of:
an extremum value; the speed of movement; the position of movement.
7. The method of claim 1, wherein obtaining initial weights and initial thresholds for the extreme learning machine neural network comprises:
and randomly acquiring initial weight and initial threshold of the neural network of the extreme learning machine.
8. A method for controlling a robot, comprising:
acquiring a robot control model based on the method for establishing a robot control model according to any one of claims 1 to 7;
and performing control on the target robot based on the robot control model.
9. The method of claim 8, wherein said performing control of a target robot based on said robot control model comprises:
acquiring target weights corresponding to various actions of the robot based on the robot control model;
weighting and summing all the behavior actions of the robot based on all the target weights to obtain behavior results;
and executing control on the target robot according to the behavior result.
10. A computer device comprising a memory and a processor, the memory having stored thereon computer instructions that, when executed by the processor, cause the method of any of claims 1-9 to be performed.
11. A non-transitory computer readable storage medium having stored thereon a computer program which, when executed by a processor, causes the method of any one of claims 1 to 9 to be performed.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111364110.2A CN114282636A (en) | 2021-11-17 | 2021-11-17 | Robot control model establishing method and robot control method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111364110.2A CN114282636A (en) | 2021-11-17 | 2021-11-17 | Robot control model establishing method and robot control method |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN114282636A true CN114282636A (en) | 2022-04-05 |
Family
ID=80869323
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202111364110.2A Pending CN114282636A (en) | 2021-11-17 | 2021-11-17 | Robot control model establishing method and robot control method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN114282636A (en) |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107414830A (en) * | 2017-07-31 | 2017-12-01 | 中南大学 | A kind of carrying machine human arm manipulation multi-level mapping intelligent control method and system |
| CN108120451A (en) * | 2017-12-21 | 2018-06-05 | 苏州大学 | Based on silicon micro accerometer temperature-compensation method, the system for improving PSO optimization neural networks |
| CN111931413A (en) * | 2020-06-22 | 2020-11-13 | 中国人民解放军陆军军事交通学院 | Method for predicting performance of high-altitude diesel engine of optimized extreme learning machine based on chaotic particle swarm with extreme disturbance |
| CN112733423A (en) * | 2020-12-03 | 2021-04-30 | 重庆邮智机器人研究院有限公司 | Industrial robot inverse kinematics solving method based on PSO-RBFNN |
| CN112733301A (en) * | 2021-01-21 | 2021-04-30 | 佛山科学技术学院 | Six-dimensional torque sensor gravity compensation method and system based on neural network |
-
2021
- 2021-11-17 CN CN202111364110.2A patent/CN114282636A/en active Pending
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107414830A (en) * | 2017-07-31 | 2017-12-01 | 中南大学 | A kind of carrying machine human arm manipulation multi-level mapping intelligent control method and system |
| CN108120451A (en) * | 2017-12-21 | 2018-06-05 | 苏州大学 | Based on silicon micro accerometer temperature-compensation method, the system for improving PSO optimization neural networks |
| CN111931413A (en) * | 2020-06-22 | 2020-11-13 | 中国人民解放军陆军军事交通学院 | Method for predicting performance of high-altitude diesel engine of optimized extreme learning machine based on chaotic particle swarm with extreme disturbance |
| CN112733423A (en) * | 2020-12-03 | 2021-04-30 | 重庆邮智机器人研究院有限公司 | Industrial robot inverse kinematics solving method based on PSO-RBFNN |
| CN112733301A (en) * | 2021-01-21 | 2021-04-30 | 佛山科学技术学院 | Six-dimensional torque sensor gravity compensation method and system based on neural network |
Non-Patent Citations (2)
| Title |
|---|
| YUN BAI ET AL.: "Research of Pose Control Algorithm of Coal Mine Rescue Snake Robot", 《MATHEMATICAL PROBLEMS IN ENGINEERING》, 22 February 2018 (2018-02-22), pages 1 - 9 * |
| 李文华等: "基于耦合算法的类人机器人模仿学习控制方法", 《中国机械工程》, vol. 28, no. 14, 31 July 2017 (2017-07-31), pages 1735 - 1741 * |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Jiang et al. | Path planning for intelligent robots based on deep Q-learning with experience replay and heuristic knowledge | |
| CN111612126B (en) | Method and apparatus for reinforcement learning | |
| Duan et al. | Benchmarking deep reinforcement learning for continuous control | |
| US9082079B1 (en) | Proportional-integral-derivative controller effecting expansion kernels comprising a plurality of spiking neurons associated with a plurality of receptive fields | |
| CN112596515B (en) | Multi-logistics robot movement control method and device | |
| CN111144580B (en) | Hierarchical reinforcement learning training method and device based on imitation learning | |
| US20240308068A1 (en) | Data-efficient hierarchical reinforcement learning | |
| CN112362066A (en) | Path planning method based on improved deep reinforcement learning | |
| CN114162146B (en) | Driving strategy model training method and automatic driving control method | |
| WO2020065001A1 (en) | Learning motor primitives and training a machine learning system using a linear-feedback-stabilized policy | |
| JP7448683B2 (en) | Learning options for action selection using meta-gradient in multi-task reinforcement learning | |
| Passalis et al. | Deep reinforcement learning for controlling frontal person close-up shooting | |
| Mohanty et al. | Application of deep Q-learning for wheel mobile robot navigation | |
| CN114239974B (en) | Multi-agent position prediction method and device, electronic equipment and storage medium | |
| US20220414460A1 (en) | Predictive Modeling of Aircraft Dynamics | |
| Rottmann et al. | Adaptive autonomous control using online value iteration with gaussian processes | |
| Chen et al. | Deep reinforcement learning-based robot exploration for constructing map of unknown environment | |
| EP2898996A1 (en) | Method of controlling a robotic system and a robotic system controller for implementing this method | |
| CN114282636A (en) | Robot control model establishing method and robot control method | |
| US20240143975A1 (en) | Neural network feature extractor for actor-critic reinforcement learning models | |
| CN113910221B (en) | Mechanical arm autonomous motion planning method, device, equipment and storage medium | |
| Paudel | Learning for robot decision making under distribution shift: A survey | |
| CN115562258A (en) | Robot social self-adaptive path planning method and system based on neural network | |
| Lee et al. | Combining GRN modeling and demonstration-based programming for robot control | |
| Liu et al. | Robotic cognitive behavior control based on biology-inspired episodic memory |
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 |