CN116512254B - Direction-based intelligent control method and system for mechanical arm, equipment and storage medium - Google Patents
Direction-based intelligent control method and system for mechanical arm, equipment and storage medium Download PDFInfo
- Publication number
- CN116512254B CN116512254B CN202310382403.6A CN202310382403A CN116512254B CN 116512254 B CN116512254 B CN 116512254B CN 202310382403 A CN202310382403 A CN 202310382403A CN 116512254 B CN116512254 B CN 116512254B
- Authority
- CN
- China
- Prior art keywords
- mechanical arm
- vector
- joint angular
- coordinate system
- joint
- 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
- 238000000034 method Methods 0.000 title claims abstract description 60
- 238000003860 storage Methods 0.000 title claims abstract description 12
- 239000013598 vector Substances 0.000 claims abstract description 183
- 230000033001 locomotion Effects 0.000 claims abstract description 70
- 238000012549 training Methods 0.000 claims description 46
- 238000003062 neural network model Methods 0.000 claims description 30
- 230000008569 process Effects 0.000 claims description 22
- 239000011159 matrix material Substances 0.000 claims description 21
- 238000004590 computer program Methods 0.000 claims description 19
- 238000005070 sampling Methods 0.000 claims description 15
- 230000006870 function Effects 0.000 claims description 11
- 238000010606 normalization Methods 0.000 claims description 11
- 238000013528 artificial neural network Methods 0.000 claims description 8
- 238000012545 processing Methods 0.000 claims description 8
- 238000012360 testing method Methods 0.000 claims description 8
- 230000009466 transformation Effects 0.000 claims description 7
- 238000006243 chemical reaction Methods 0.000 claims description 6
- 238000013519 translation Methods 0.000 claims description 6
- 238000010276 construction Methods 0.000 claims description 3
- 238000009434 installation Methods 0.000 claims description 2
- 238000004458 analytical method Methods 0.000 abstract description 4
- 238000010586 diagram Methods 0.000 description 7
- 238000012986 modification Methods 0.000 description 7
- 230000004048 modification Effects 0.000 description 7
- 230000005540 biological transmission Effects 0.000 description 3
- 238000004891 communication Methods 0.000 description 3
- 230000004913 activation Effects 0.000 description 2
- 210000004556 brain Anatomy 0.000 description 2
- 230000008859 change Effects 0.000 description 2
- 238000011217 control strategy Methods 0.000 description 2
- 230000000694 effects Effects 0.000 description 2
- 238000013178 mathematical model Methods 0.000 description 2
- 230000003287 optical effect Effects 0.000 description 2
- 238000011160 research Methods 0.000 description 2
- ORILYTVJVMAKLC-UHFFFAOYSA-N Adamantane Natural products C1C(C2)CC3CC1CC2C3 ORILYTVJVMAKLC-UHFFFAOYSA-N 0.000 description 1
- RYGMFSIKBFXOCR-UHFFFAOYSA-N Copper Chemical compound [Cu] RYGMFSIKBFXOCR-UHFFFAOYSA-N 0.000 description 1
- 230000003044 adaptive effect Effects 0.000 description 1
- 230000004075 alteration Effects 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000004364 calculation method Methods 0.000 description 1
- 238000001514 detection method Methods 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 239000000835 fiber Substances 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- 238000009776 industrial production Methods 0.000 description 1
- 238000004519 manufacturing process Methods 0.000 description 1
- 238000005259 measurement Methods 0.000 description 1
- 210000000337 motor cortex Anatomy 0.000 description 1
- 210000002569 neuron Anatomy 0.000 description 1
- 238000005457 optimization Methods 0.000 description 1
- 210000000707 wrist Anatomy 0.000 description 1
Classifications
-
- B—PERFORMING OPERATIONS; TRANSPORTING
- B25—HAND TOOLS; PORTABLE POWER-DRIVEN TOOLS; MANIPULATORS
- B25J—MANIPULATORS; CHAMBERS PROVIDED WITH MANIPULATION DEVICES
- B25J9/00—Programme-controlled manipulators
- B25J9/16—Programme controls
- B25J9/1656—Programme controls characterised by programming, planning systems for manipulators
- B25J9/1664—Programme controls characterised by programming, planning systems for manipulators characterised by motion, path, trajectory planning
-
- B—PERFORMING OPERATIONS; TRANSPORTING
- B25—HAND TOOLS; PORTABLE POWER-DRIVEN TOOLS; MANIPULATORS
- B25J—MANIPULATORS; CHAMBERS PROVIDED WITH MANIPULATION DEVICES
- B25J9/00—Programme-controlled manipulators
- B25J9/16—Programme controls
- B25J9/1602—Programme controls characterised by the control system, structure, architecture
-
- B—PERFORMING OPERATIONS; TRANSPORTING
- B25—HAND TOOLS; PORTABLE POWER-DRIVEN TOOLS; MANIPULATORS
- B25J—MANIPULATORS; CHAMBERS PROVIDED WITH MANIPULATION DEVICES
- B25J9/00—Programme-controlled manipulators
- B25J9/16—Programme controls
- B25J9/1602—Programme controls characterised by the control system, structure, architecture
- B25J9/161—Hardware, e.g. neural networks, fuzzy logic, interfaces, processor
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02P—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN THE PRODUCTION OR PROCESSING OF GOODS
- Y02P90/00—Enabling technologies with a potential contribution to greenhouse gas [GHG] emissions mitigation
- Y02P90/02—Total factory control, e.g. smart factories, flexible manufacturing systems [FMS] or integrated manufacturing systems [IMS]
Landscapes
- Engineering & Computer Science (AREA)
- Robotics (AREA)
- Mechanical Engineering (AREA)
- Automation & Control Theory (AREA)
- Physics & Mathematics (AREA)
- Artificial Intelligence (AREA)
- Evolutionary Computation (AREA)
- Fuzzy Systems (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Numerical Control (AREA)
Abstract
The invention discloses a direction-based intelligent control method of a mechanical arm, a system, equipment and a storage medium, wherein the direction-based intelligent control method of the mechanical arm constructs a functional relationship among a joint angular position vector, a joint angular velocity vector and a final velocity vector of the mechanical arm, and constructs an action analysis model based on a motion direction on the basis, wherein the input of the model is the configuration and the motion direction of the mechanical arm and is output as a motion parameter of a joint space of the mechanical arm, so that the motion direction of the three-dimensional space of the mechanical arm can be automatically converted into a motion instruction of the joint space, the intelligent control of the mechanical arm is irrelevant to the positions of a starting point and a final point and is only relevant to the motion direction, and the intelligent control method is well applicable to complex non-structural environments.
Description
Technical Field
The present invention relates to the field of mechanical arm control technologies, and in particular, to a direction-based mechanical arm intelligent control method and system, an electronic device, and a computer readable storage medium.
Background
A robot arm is a type of automated mechanical device widely used in the robot field, which implements complex operation tasks by simulating human arm, wrist and hand functions. For a long time, the problem of motion control of the mechanical arm is a difficult problem of limiting the development of the mechanical arm, so that the mechanical arm is mainly applied to the field of structural industrial production, and the operation effect in a complex unstructured environment is not ideal. Currently, the mechanical arm mainly utilizes a control strategy based on position to complete a target capturing task, and accurate measurement of target position information and attitude information is required. Neurophysiologists found through research that the movement of motor cortex is related to the movement direction, the brain contains nerve cell structures in the coding direction, the field vector in the coding direction can be used for controlling the arm movement, and when the direction-based control is adopted, the activation of the brain is irrelevant to the precise positions of the starting point and the tail end point of the movement and is only related to the direction. Therefore, the control of the mechanical arm by adopting a control strategy based on the direction is a new research direction, but the movement direction cannot be analyzed and converted into the joint space at present so as to realize the intelligent movement control of the mechanical arm.
Disclosure of Invention
The invention provides a direction-based intelligent control method and system for a mechanical arm, electronic equipment and a computer-readable storage medium, which are used for solving the technical problem that the prior art cannot realize intelligent control of the mechanical arm based on a movement direction.
According to one aspect of the invention, there is provided a direction-based intelligent control method for a robot arm, including the following:
constructing a mechanical arm motion coordinate system;
acquiring a joint angular position vector, a joint angular velocity vector and a tail end velocity vector of the mechanical arm as training data;
normalizing the tail end velocity vector and the joint angular velocity vector;
constructing a neural network model, taking a joint angular position vector and a normalized tail end speed vector as input, taking a normalized joint angular speed vector as output, and training the neural network model until the model converges;
and controlling the mechanical arm to move by using the trained neural network model.
Further, the process of constructing the mechanical arm motion coordinate system specifically comprises the following steps:
defining the working space of the mechanical arm as W, the configuration space as zeta, and the mechanical arm comprising n driving degrees of freedom, wherein the mechanical arm configuration is represented by a joint angular position vector q= [ q ] of the mechanical arm 1 ,q 2 ,L,...,q n ]Q εζ, L is the connecting rod position vector of the mechanical arm, q n Represents the nth joint angle;
defining a base coordinate system b, an end coordinate system e and a connecting rod coordinate system L of the mechanical arm in a working space W i I=1, 2, n, the transformation relationship between the two coordinate systems is denoted as T, then the base coordinate system b and the i-th link joint coordinate system L i The conversion relation between is thatThe conversion relationship between the terminal coordinate system e and the base coordinate system b is expressed as +.>
Further, when the mechanical arm system has a kinematic model, the process of acquiring the joint angular position vector, the joint angular velocity vector, and the tip speed vector of the mechanical arm as training data includes the following:
constructing a kinematic model from a base coordinate system to an end coordinate system Represents the coordinate transformation relation from the base coordinate system to the terminal coordinate system when the joint angular position vector is q,/>Representing the rotation relation from the end coordinate system to the base coordinate system,/->A representation of the origin of the terminal coordinate system in the base coordinate system;
calculating a Jacobian matrix according to the constructed kinematic model, and constructing a functional relation among a joint angular position vector, a joint angular velocity vector and a terminal velocity vector:χ represents the terminal velocity vector, represented by the translational velocity v and rotational velocity w in the terminal three-dimensional space, J (q) represents the Jacobian matrix, q & Representing the angular velocity vector of the joint and determining q by the inverse Jacobian matrix method & =J -1 (q)χ;
Randomly generating a group of joint angle position vectors q in the effective joint angle range of the mechanical arm, and estimating the pose of the tail end by adopting a constructed kinematic model
By means of terminal positionEstimating Jacobian matrix J (q) and inverse Jacobian matrix J -1 (q);
According to the speed constraint of the joint space of the mechanical arm, randomly sampling N groups of joint angular speed vectors q in the joint space & Calculating to obtain a tail end speed vector χ through a Jacobian matrix J (q), and obtaining N sampling point data corresponding to the set of joint angular position vectors q, wherein the data of each sampling point is expressed as (q, q) & ,χ);
Repeating the above stepsProcedure, obtaining sampling point data of M joint angular position vectors to construct training data setRandomly selecting 2K/3 data as a training set and the rest K/3 data as a test set.
Further, when the mechanical arm system does not have a kinematic model, the process of acquiring the joint angular position vector, the joint angular velocity vector, and the tip speed vector of the mechanical arm as training data includes the following:
the installation auxiliary system is used for acquiring the position p, the posture R, the translation speed v and the rotation speed w of the tail end of the mechanical arm in real time, wherein the translation speed v and the rotation speed w of the tail end of the mechanical arm form a tail end speed vector
Driving the mechanical arm to move in the whole space in the working space, and obtaining a joint angular position vector q and a joint angular velocity vector q of the mechanical arm by utilizing body feedback & Thereby collecting K groups of dataRandomly selecting 2K/3 data as a training set and the rest K/3 data as a test set.
Further, the normalization processing of the terminal velocity vector and the joint angular velocity vector specifically includes:
the end velocity vector χ i Normalized toObtaining normalization factor k i Then normalize the factor k i And joint angular velocity vector->Multiplication to obtain->Thereby obtaining the normalized training data set
Further, the loss function expression of the neural network model is:
wherein L (q, q ') represents the sum of losses of all samples, q' i Represents the joint angular velocity, q, of the neural network model output i Representing the sampled data true values.
Further, the following is included in the model training process:
joint angular velocity vector output by neural networkAnd formula->Calculating an estimated terminal velocity vector +.>Judging the estimated terminal velocity vector by cosine similarity>And the actual terminal velocity vector +.>The closer the cosine value between the two is to 1, the closer the vector direction determined by the neural network is to the actual direction.
In addition, the invention also provides a direction-based intelligent control system of the mechanical arm, which comprises the following components:
the coordinate system construction module is used for constructing a mechanical arm movement coordinate system;
the data acquisition module is used for acquiring a joint angular position vector, a joint angular velocity vector and a tail end velocity vector of the mechanical arm as training data;
the normalization module is used for carrying out normalization processing on the tail end speed vector and the joint angular speed vector;
the model training module is used for constructing a neural network model, taking the joint angular position vector and the normalized tail end speed vector as input, taking the normalized joint angular speed vector as output, and training the neural network model until the model converges;
and the intelligent control module is used for controlling the movement of the mechanical arm by using the trained neural network model.
In addition, the invention also provides an electronic device comprising a processor and a memory, wherein the memory stores a computer program, and the processor is used for executing the steps of the method by calling the computer program stored in the memory.
In addition, the invention also provides a computer readable storage medium for storing a computer program for intelligent control of a mechanical arm based on direction, the computer program executing the steps of the method when running on a computer.
The invention has the following effects:
according to the intelligent control method for the mechanical arm based on the direction, after a mechanical arm motion coordinate system is constructed, a series of joint angular position vectors, joint angular velocity vectors and tail end velocity vectors of the mechanical arm in the motion process are obtained to serve as training data, after the velocity vectors are normalized, the joint angular position vectors and the normalized tail end velocity vectors are taken as input, the normalized joint angular velocity vectors are taken as output, the model is trained, and finally the mechanical arm motion is controlled by using a trained neural network model. According to the intelligent control method for the mechanical arm based on the direction, the functional relation among the joint angular position vector, the joint angular velocity vector and the final velocity vector of the mechanical arm is constructed, the motion analysis model based on the motion direction is constructed on the basis, the input of the model is the mechanical arm configuration and the motion direction, and the output is the motion parameter of the joint space of the mechanical arm, so that the motion direction of the three-dimensional space of the mechanical arm can be automatically converted into the motion instruction of the joint space, the intelligent control of the mechanical arm is independent of the positions of the starting point and the final point and is only related to the motion direction, and the intelligent control method is well applicable to complex non-structural environments.
In addition, the intelligent control system of the direction-based mechanical arm has the advantages.
In addition to the objects, features and advantages described above, the present invention has other objects, features and advantages. The present invention will be described in further detail with reference to the drawings.
Drawings
The accompanying drawings, which are included to provide a further understanding of the invention and are incorporated in and constitute a part of this specification, illustrate embodiments of the invention and together with the description serve to explain the invention. In the drawings:
fig. 1 is a schematic flow chart of a direction-based intelligent control method for a mechanical arm according to a preferred embodiment of the present invention.
Fig. 2 is a schematic block diagram of a direction-based intelligent control system for a robot arm according to another embodiment of the present invention.
Detailed Description
Embodiments of the invention are described in detail below with reference to the attached drawing figures, but the invention can be practiced in a number of different ways, as defined and covered below.
As shown in fig. 1, a preferred embodiment of the present invention provides a direction-based intelligent control method for a mechanical arm, which includes the following steps:
step S1: constructing a mechanical arm motion coordinate system;
step S2: acquiring a joint angular position vector, a joint angular velocity vector and a tail end velocity vector of the mechanical arm as training data;
step S3: normalizing the tail end velocity vector and the joint angular velocity vector;
step S4: constructing a neural network model, taking a joint angular position vector and a normalized tail end speed vector as input, taking a normalized joint angular speed vector as output, and training the neural network model until the model converges;
step S5: and controlling the mechanical arm to move by using the trained neural network model.
It can be understood that, in the direction-based intelligent control method for the mechanical arm of the present embodiment, after the mechanical arm motion coordinate system is constructed, a series of joint angular position vectors, joint angular velocity vectors and terminal velocity vectors of the mechanical arm in the motion process are obtained as training data, after the velocity vectors are normalized, the joint angular position vectors and the normalized terminal velocity vectors are taken as inputs, the normalized joint angular velocity vectors are taken as outputs, so as to train a model, and finally, the motion of the mechanical arm is controlled by using the trained neural network model. According to the intelligent control method for the mechanical arm based on the direction, the functional relation among the joint angular position vector, the joint angular velocity vector and the final velocity vector of the mechanical arm is constructed, the motion analysis model based on the motion direction is constructed on the basis, the input of the model is the mechanical arm configuration and the motion direction, and the output is the motion parameter of the joint space of the mechanical arm, so that the motion direction of the three-dimensional space of the mechanical arm can be automatically converted into the motion instruction of the joint space, the intelligent control of the mechanical arm is independent of the positions of the starting point and the final point and is only related to the motion direction, and the intelligent control method is well applicable to complex non-structural environments.
It can be understood that in the step S1, the process of constructing the motion coordinate system of the mechanical arm is specifically:
defining the working space of the robot arm as W, the working space at the plane can be expressed as w=τ 2 The working space of the three-dimensional space can be expressed as w=τ 3 τ represents the motion dimension of the robotic arm. The configuration space of the mechanical arm is defined as zeta, and the hook type of the mechanical arm is expressed by q epsilon zeta. The mechanical arm comprises n driving degrees of freedom, and the mechanical arm configuration can be formed by a joint angle position vector q= [ q ] of the mechanical arm 1 ,q 2 ,L,...,q n ]Q εζ, L is the connecting rod position vector of the mechanical arm, q n Representing the nth jointAnd (5) corners. Then the base coordinate system of the mechanical arm in the working space W is defined as b, the tail end coordinate system is defined as e, and the connecting rod coordinate system is defined as L i I=1, 2, n, the transformation relationship between the two coordinate systems is denoted as T, then the base coordinate system b and the i-th link joint coordinate system L i The conversion relation between them can be expressed asThe conversion relationship between the terminal coordinate system e and the base coordinate system b is expressed as +.>In addition, the coordinate transformation relation between adjacent links (i.e., any of the i-1 link and the i link)>Is determined by the joint angle connecting the two links.
It will be appreciated that, in the step S2, when the mechanical arm system has a kinematic model, the training data may be directly obtained through a mathematical model, and specifically, the process of obtaining the joint angular position vector, the joint angular velocity vector, and the tip speed vector of the mechanical arm as the training data includes the following:
firstly, constructing a kinematic model from a base coordinate system to an end coordinate system by adopting a DH parameter method or other existing methodsWherein (1)>Represents the coordinate transformation relation from the base coordinate system to the terminal coordinate system when the joint angular position vector is q,/>Representing the rotation relation from the end coordinate system to the base coordinate system,/->Origin at the terminal coordinate systemRepresentation under the base coordinate system.
And then calculating a Jacobian matrix according to the constructed kinematic model, and constructing a functional relation among a joint angular position vector, a joint angular velocity vector and a terminal velocity vector:χ represents the tip speed vector, which may represent the direction of motion of the robot arm, by the translational velocity v and rotational velocity w of the tip in three dimensions, J (q) represents the jacobian matrix, which is related only to the robot arm configuration, i.e., to the joint angular position vector q, q & Representing the angular velocity vector of the joint and determining q by the inverse Jacobian matrix method & =J -1 (q) χ. The process of calculating the jacobian matrix based on the known kinematic model belongs to the prior art, and therefore will not be described herein.
Then, randomly generating a group of joint angle position vectors q in the effective joint angle range of the mechanical arm, and adopting a kinematic model constructed previouslyEstimating the pose of the tip->Wherein, collision detection is needed after a group of joint angular position vectors q are randomly generated, so as to ensure that the sampled joint angular position vectors are reasonable and no collision exists between connecting rods.
Then pass through the tail end poseEstimating Jacobian matrix J (q) and inverse Jacobian matrix J -1 (q). It can be appreciated that the terminal pose +.>It is described that the tip position and direction change with the change of the joint angle, and therefore, useThe equation derives time to determine the translational velocity v and rotational velocity w of the tip and the joint angular velocity q & Is x=j (q) q & Thereby estimating the jacobian matrix J (q). Then inverting J (q) to obtain an inverse Jacobian matrix J -1 (q)。
Then, according to the speed constraint of the joint space of the mechanical arm, randomly sampling N groups of joint angular speed vectors q in the joint space & The sampling speed should meet the constraint requirement of the maximum joint angular speed of the mechanical arm, and the sampling points exceeding the constraint requirement are directly abandoned until the N groups of joint angular speeds are sampled. Then through Jacobian matrix J (q) and formula χ=J (q) q & Calculating to obtain the tail end velocity vector χ, thereby obtaining N sampling point data corresponding to the set of joint angular position vectors q, wherein the data of each sampling point is expressed as (q, q & ,χ)。
Repeating the above process to obtain sampling point data of M joint angular position vectors to construct a training data setNamely, M mechanical arm configurations are provided, each mechanical arm configuration is provided with N sampling points, 2K/3 data are randomly selected as a training set, and the remaining K/3 data are selected as a test set. Of course, the proportion of the training set and the testing set may be set according to actual needs, and is not specifically limited herein.
In addition, when the mechanical arm system does not have a kinematic model, training data cannot be directly obtained through the mathematical model, and at this time, the process of obtaining the joint angular position vector, the joint angular velocity vector and the terminal velocity vector of the mechanical arm as the training data includes the following steps:
an auxiliary system, such as a three-dimensional motion capture system, is firstly installed for acquiring the position p, the posture R, the translation speed v and the rotation speed w of the tail end of the mechanical arm in real time, wherein the translation speed v and the rotation speed w of the tail end of the mechanical arm form a tail end speed vectorIn addition, v and w may also be fitted from the data of position p and pose R.
Then, driving the mechanical arm to move in the working space in a full space, and obtaining a joint angular position vector q and a joint angular velocity vector q of the mechanical arm at each movement position by using body feedback & Thereby, K groups of data can be acquiredRandomly selecting 2K/3 data as a training set and the rest K/3 data as a test set.
It may be understood that in the step S3, the normalization process is specifically performed on the terminal velocity vector and the joint angular velocity vector:
for the data of training set and test set, the end velocity vector χ is calculated i Normalized toObtaining normalization factor k i Then normalize the factor k i And joint angular velocity vector->Multiplication to obtain->Thereby obtaining the normalized training data setWherein (1)>The direction of movement of the robotic arm may be indicated.
It can be appreciated that in the step S4, the neural network model adopts a fully-connected neural network, which includes C hidden layers, and the i-th hidden layer includes a number of nodes T i Then the number of nodes of the model is sharedThe activation function of the neural network model can be a tanh function or a relu function, and the optimization method is an adaptive moment estimation Adam or a random gradient descent SGD. The input data of the neural network model during training is +.>i=1, 2..2.k/3, output data is +.>i=1, 2K/3, i.e. the input of the model is the mechanical arm configuration and the motion direction, and the output is the motion parameter of the mechanical arm joint space, so that the motion direction of the mechanical arm three-dimensional space can be automatically converted into the motion instruction of the joint space.
The loss function expression of the neural network model is as follows:
wherein L (q, q ') represents the sum of losses of all samples, q' i Represents the joint angular velocity, q, of the neural network model output i The data true value representing the sample, i represents the sample number. In the training process, when the loss sum L (q, q') of all samples is smaller than a preset threshold value, the judgment model is converged.
Optionally, in step S4, the model training process further includes the following:
joint angular velocity vector output by neural networkAnd formula->Calculating an estimated terminal velocity vector +.>Then through cosine similarityJudging the estimated terminal velocity vector +.>And the actual terminal velocity vector +.>The specific calculation formula is as follows: />And theta represents an included angle between the estimated vector direction and the actual vector direction output by the neural network, and the closer the included angle theta is to 0, the closer a cosine value between the two directions is to 1, the closer the estimated vector direction and the actual vector direction determined by the neural network are, and the higher the model precision is. It can be understood that the accuracy of the model is evaluated by calculating the cosine similarity between the estimated vector direction and the actual vector direction output by the model, which is beneficial to improving the control accuracy of the mechanical arm.
It can be understood that in the step S5, when the intelligent control of the mechanical arm is required after the model training is completed, only the joint angular position vector and the final speed vector of the mechanical arm, that is, the configuration and the movement direction of the mechanical arm are input, the joint angular speed vector of the mechanical arm, that is, the movement parameters of the joint space of the mechanical arm, can be automatically output, so that the movement direction of the three-dimensional space of the mechanical arm is automatically converted into the movement instruction of the joint space, the intelligent control of the mechanical arm is independent of the positions of the starting point and the final point, and is only related to the movement direction, thereby being well applicable to complex non-structural environments.
In addition, as shown in fig. 2, another embodiment of the present invention further provides a direction-based intelligent control system for a mechanical arm, preferably adopting the method as described above, including:
the coordinate system construction module is used for constructing a mechanical arm movement coordinate system;
the data acquisition module is used for acquiring a joint angular position vector, a joint angular velocity vector and a tail end velocity vector of the mechanical arm as training data;
the normalization module is used for carrying out normalization processing on the tail end speed vector and the joint angular speed vector;
the model training module is used for constructing a neural network model, taking the joint angular position vector and the normalized tail end speed vector as input, taking the normalized joint angular speed vector as output, and training the neural network model until the model converges;
and the intelligent control module is used for controlling the movement of the mechanical arm by using the trained neural network model.
It can be understood that, in the direction-based intelligent control system for a mechanical arm of this embodiment, after the mechanical arm motion coordinate system is constructed, a series of joint angular position vectors, joint angular velocity vectors and terminal velocity vectors of the mechanical arm in the motion process are obtained as training data, after the velocity vectors are normalized, the joint angular position vectors and the normalized terminal velocity vectors are taken as inputs, the normalized joint angular velocity vectors are taken as outputs, so as to train a model, and finally, the mechanical arm motion is controlled by using the trained neural network model. According to the intelligent control system for the mechanical arm based on the direction, the functional relation among the joint angular position vector, the joint angular velocity vector and the final velocity vector of the mechanical arm is constructed, the motion analysis model based on the motion direction is constructed on the basis, the input of the model is the mechanical arm configuration and the motion direction, and the output is the motion parameter of the joint space of the mechanical arm, so that the motion direction of the three-dimensional space of the mechanical arm can be automatically converted into the motion instruction of the joint space, the intelligent control of the mechanical arm is independent of the positions of the starting point and the final point and is only related to the motion direction, and the intelligent control system can be well applied to complex non-structural environments.
In addition, another embodiment of the present invention also provides an electronic device, including a processor and a memory, where the memory stores a computer program, and the processor is configured to execute the steps of the method described above by calling the computer program stored in the memory.
In addition, another embodiment of the present invention also provides a computer readable storage medium for storing a computer program for performing intelligent control of a manipulator based on a direction, wherein the computer program when run on a computer performs the steps of the method as described above.
Forms of general computer-readable storage media include: a floppy disk (floppy disk), a flexible disk (flexible disk), hard disk, magnetic tape, any other magnetic medium, a CD-ROM, any other optical medium, punch cards, paper tape, any other physical medium with patterns of holes, a Random Access Memory (RAM), a programmable read-only memory (PROM), an erasable programmable read-only memory (EPROM), a FLASH erasable programmable read-only memory (FLASH-EPROM), any other memory chip or cartridge, or any other medium from which a computer can read. The instructions may further be transmitted or received over a transmission medium. The term transmission medium may include any tangible or intangible medium that may be used to store, encode, or carry instructions for execution by a machine, and includes digital or analog communications signals or their communications with intangible medium that facilitate communication of such instructions. Transmission media includes coaxial cables, copper wire and fiber optics, including the wires that comprise a bus for transmitting a computer data signal.
The above description is only of the preferred embodiments of the present invention and is not intended to limit the present invention, but various modifications and variations can be made to the present invention by those skilled in the art. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention should be included in the protection scope of the present invention.
It will be appreciated by those skilled in the art that embodiments of the present application may be provided as a method, system, or computer program product. Accordingly, the present application may take the form of an entirely hardware embodiment, an entirely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present application may take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, and the like) having computer-usable program code embodied therein. The solutions in the embodiments of the present application may be implemented in various computer languages, for example, object-oriented programming language Java, and an transliterated scripting language JavaScript, etc.
The present application is described with reference to flowchart illustrations and/or block diagrams of methods, apparatus (systems) and computer program products according to embodiments of the application. It will be understood that each flow and/or block of the flowchart illustrations and/or block diagrams, and combinations of flows and/or blocks in the flowchart illustrations and/or block diagrams, can be implemented by computer program instructions. These computer program instructions may be provided to a processor of a general purpose computer, special purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, create means for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.
These computer program instructions may also be stored in a computer-readable memory that can direct a computer or other programmable data processing apparatus to function in a particular manner, such that the instructions stored in the computer-readable memory produce an article of manufacture including instruction means which implement the function specified in the flowchart flow or flows and/or block diagram block or blocks.
These computer program instructions may also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer implemented process such that the instructions which execute on the computer or other programmable apparatus provide steps for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.
While preferred embodiments of the present application have been described, additional variations and modifications in those embodiments may occur to those skilled in the art once they learn of the basic inventive concepts. It is therefore intended that the following claims be interpreted as including the preferred embodiments and all such alterations and modifications as fall within the scope of the application.
It will be apparent to those skilled in the art that various modifications and variations can be made in the present application without departing from the spirit or scope of the application. Thus, if such modifications and variations of the present application fall within the scope of the claims and the equivalents thereof, the present application is intended to cover such modifications and variations.
Claims (9)
1. The intelligent control method for the mechanical arm based on the direction is characterized by comprising the following steps of:
constructing a mechanical arm motion coordinate system;
acquiring a joint angular position vector, a joint angular velocity vector and a tail end velocity vector of the mechanical arm as training data;
normalizing the tail end velocity vector and the joint angular velocity vector;
constructing a neural network model, taking a joint angular position vector and a normalized tail end speed vector as input, taking a normalized joint angular speed vector as output, and training the neural network model until the model converges;
controlling the mechanical arm to move by using the trained neural network model;
the process for constructing the mechanical arm motion coordinate system comprises the following steps:
defining the working space of the mechanical arm as W, the configuration space as zeta, and the mechanical arm comprising n driving degrees of freedom, wherein the mechanical arm configuration is represented by a joint angular position vector q= [ q ] of the mechanical arm 1 ,q 2 ,L,...,q n ]Q εζ, L is the connecting rod position vector of the mechanical arm, q n Represents the nth joint angle;
defining a base coordinate system b, an end coordinate system e and a connecting rod coordinate system L of the mechanical arm in a working space W i I=1, 2, n, the transformation relationship between the two coordinate systems is denoted as T, then the base coordinate system b and the i-th link joint coordinate system L i The conversion relation between is thatThe conversion relationship between the end coordinate system e and the base coordinate system b is expressed as T e b 。
2. The intelligent control method for a direction-based robot arm according to claim 1, wherein when the robot arm system has a kinematic model, the process of acquiring the joint angular position vector, the joint angular velocity vector, and the tip velocity vector of the robot arm as training data comprises the following:
constructing a kinematic model from a base coordinate system to an end coordinate system Represents the coordinate transformation relation from the base coordinate system to the terminal coordinate system when the joint angular position vector is q,/>Representing the rotation relation from the end coordinate system to the base coordinate system,/->A representation of the origin of the terminal coordinate system in the base coordinate system;
calculating a Jacobian matrix according to the constructed kinematic model, and constructing a functional relation among a joint angular position vector, a joint angular velocity vector and a terminal velocity vector:χ represents the terminal velocity vector, represented by the translational velocity v and rotational velocity w in the terminal three-dimensional space, J (q) represents the Jacobian matrix, q & Representing the angular velocity vector of the joint and determining q by the inverse Jacobian matrix method & =J -1 (q)χ;
Randomly generating a group of joint angle position vectors q in the effective joint angle range of the mechanical arm, and estimating the pose of the tail end by adopting a constructed kinematic model
By means of terminal positionEstimating Jacobian matrix J (q) and inverse Jacobian matrix J -1 (q);
According to the speed constraint of the joint space of the mechanical arm, randomly sampling N groups of joint angular speed vectors q in the joint space & Calculating to obtain a tail end speed vector χ through a Jacobian matrix J (q), and obtaining N sampling point data corresponding to the set of joint angular position vectors q, wherein the data of each sampling point is expressed as (q, q) & ,χ);
Repeating the above process to obtain sampling point data of M joint angular position vectors to construct a training data seti=1, 2,..k, k=m×n }, 2K/3 data were randomly selected as training set and the remaining K/3 data were selected as test set.
3. The intelligent control method for a direction-based robot arm according to claim 1, wherein when the robot arm system does not have a kinematic model, the process of acquiring the joint angular position vector, the joint angular velocity vector, and the tip velocity vector of the robot arm as training data includes:
the installation auxiliary system is used for acquiring the position p, the posture R, the translation speed v and the rotation speed w of the tail end of the mechanical arm in real time, wherein the translation speed v and the rotation speed w of the tail end of the mechanical arm form a tail end speed vector
Driving the mechanical arm to move in the whole space in the working space, and obtaining a joint angular position vector q and a joint angular velocity vector q of the mechanical arm by utilizing body feedback & Thereby collecting K groups of dataRandomly selecting 2K/3 data as a training set and the rest K/3 data as a test set.
4. The intelligent control method for a direction-based mechanical arm according to claim 2 or 3, wherein the normalizing process for the terminal velocity vector and the joint angular velocity vector specifically comprises:
the end velocity vector χ i Normalized toObtaining normalization factor k i Then normalize the factor k i And joint angular velocity vector->Multiplication to obtain->Thereby obtaining the normalized training data set
5. The intelligent control method of a direction-based mechanical arm according to claim 4, wherein the loss function expression of the neural network model is:
wherein L (q, q ') represents the sum of losses of all samples, q' i Represents the joint angular velocity, q, of the neural network model output i Representing the sampled data true values.
6. The intelligent control method for a direction-based mechanical arm according to claim 5, further comprising the following steps in the model training process:
joint angular velocity vector output by neural networkAnd formula->Calculating an estimated terminal velocity vector +.>Judging the estimated terminal velocity vector by cosine similarity>And the actual terminal velocity vector +.>The closer the cosine value between the two is to 1, the closer the vector direction determined by the neural network is to the actual direction.
7. A direction-based intelligent control system for a robotic arm, employing the method of any of claims 1-6, comprising:
the coordinate system construction module is used for constructing a mechanical arm movement coordinate system;
the data acquisition module is used for acquiring a joint angular position vector, a joint angular velocity vector and a tail end velocity vector of the mechanical arm as training data;
the normalization module is used for carrying out normalization processing on the tail end speed vector and the joint angular speed vector;
the model training module is used for constructing a neural network model, taking the joint angular position vector and the normalized tail end speed vector as input, taking the normalized joint angular speed vector as output, and training the neural network model until the model converges;
and the intelligent control module is used for controlling the movement of the mechanical arm by using the trained neural network model.
8. An electronic device comprising a processor and a memory, said memory having stored therein a computer program for executing the steps of the method according to any of claims 1-6 by invoking said computer program stored in said memory.
9. A computer-readable storage medium storing a computer program for intelligent control of a robotic arm based on direction, characterized in that the computer program when run on a computer performs the steps of the method according to any one of claims 1-6.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202310382403.6A CN116512254B (en) | 2023-04-11 | 2023-04-11 | Direction-based intelligent control method and system for mechanical arm, equipment and storage medium |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202310382403.6A CN116512254B (en) | 2023-04-11 | 2023-04-11 | Direction-based intelligent control method and system for mechanical arm, equipment and storage medium |
Publications (2)
Publication Number | Publication Date |
---|---|
CN116512254A CN116512254A (en) | 2023-08-01 |
CN116512254B true CN116512254B (en) | 2024-01-23 |
Family
ID=87391318
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN202310382403.6A Active CN116512254B (en) | 2023-04-11 | 2023-04-11 | Direction-based intelligent control method and system for mechanical arm, equipment and storage medium |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN116512254B (en) |
Citations (21)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JPH0699372A (en) * | 1992-09-21 | 1994-04-12 | Toyoda Mach Works Ltd | Robot control device |
CN106113034A (en) * | 2016-06-12 | 2016-11-16 | 哈尔滨工程大学 | A kind of sixdegree-of-freedom simulation considers the method for planning track of force constraint |
CN106346478A (en) * | 2016-11-09 | 2017-01-25 | 广州视源电子科技股份有限公司 | control method and device of mechanical arm |
CN107378944A (en) * | 2017-06-20 | 2017-11-24 | 东南大学 | A kind of multi-dimensional surface electromyographic signal prosthetic hand control method based on PCA |
CN109129487A (en) * | 2018-09-26 | 2019-01-04 | 华南理工大学 | Repetitive motion planning method for redundant manipulator based on Taylor's type discrete periodic rhythm and pace of moving things neural network under periodic noise |
CN109623814A (en) * | 2018-12-14 | 2019-04-16 | 中南大学 | A kind of Mechanical arm control method |
KR20190073907A (en) * | 2017-12-19 | 2019-06-27 | 고동휘 | Face Verification with Deep Convolutional Neural Network |
CN110000787A (en) * | 2019-04-12 | 2019-07-12 | 清华大学深圳研究生院 | A kind of control method of super redundant mechanical arm |
CN110488611A (en) * | 2019-09-02 | 2019-11-22 | 山东建筑大学 | A kind of biomimetic robot fish movement control method, controller and bionic machine fish |
AU2019101186A4 (en) * | 2019-10-02 | 2020-01-23 | Guo, Zhongliang MR | A Method of Video Recognition Network of Face Tampering Based on Deep Learning |
CN111240195A (en) * | 2020-01-08 | 2020-06-05 | 北京迈格威科技有限公司 | Automatic control model training and target object recycling method and device based on machine vision |
CN111292230A (en) * | 2020-02-18 | 2020-06-16 | 上海交通大学 | Method, system, medium, and apparatus for spiral transform data augmentation in deep learning |
CN112765879A (en) * | 2021-01-12 | 2021-05-07 | 中国人民解放军国防科技大学 | Mechanical arm inverse kinematics solving method based on neural network and configuration coding |
CN113400307A (en) * | 2021-06-16 | 2021-09-17 | 清华大学 | Control method of space robot mechanical arm |
CN113500599A (en) * | 2021-07-16 | 2021-10-15 | 中山大学 | Track tracking method and system of rope-driven mechanical arm based on neural network |
CN113743287A (en) * | 2021-08-31 | 2021-12-03 | 之江实验室 | Robot self-adaptive grabbing control method and system based on impulse neural network |
CN113771027A (en) * | 2021-08-17 | 2021-12-10 | 浙江工业大学 | Two-arm cooperative grabbing method based on deep learning |
CN114603564A (en) * | 2022-04-28 | 2022-06-10 | 中国电力科学研究院有限公司 | Mechanical arm navigation obstacle avoidance method and system, computer equipment and storage medium |
WO2022156475A1 (en) * | 2021-01-20 | 2022-07-28 | 华为技术有限公司 | Neural network model training method and apparatus, and data processing method and apparatus |
CN114888801A (en) * | 2022-05-16 | 2022-08-12 | 南京邮电大学 | Mechanical arm control method and system based on offline strategy reinforcement learning |
WO2023046074A1 (en) * | 2021-09-26 | 2023-03-30 | 广州市微眸医疗器械有限公司 | Autonomous learning method for mechanical arm control method |
Family Cites Families (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US11927965B2 (en) * | 2016-02-29 | 2024-03-12 | AI Incorporated | Obstacle recognition method for autonomous robots |
EP3742347B1 (en) * | 2016-03-03 | 2022-11-02 | Google LLC | Deep machine learning methods and apparatus for robotic grasping |
US11610109B2 (en) * | 2018-09-26 | 2023-03-21 | Microsoft Technology Licensing, Llc | Language agnostic machine learning model for title standardization |
-
2023
- 2023-04-11 CN CN202310382403.6A patent/CN116512254B/en active Active
Patent Citations (21)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JPH0699372A (en) * | 1992-09-21 | 1994-04-12 | Toyoda Mach Works Ltd | Robot control device |
CN106113034A (en) * | 2016-06-12 | 2016-11-16 | 哈尔滨工程大学 | A kind of sixdegree-of-freedom simulation considers the method for planning track of force constraint |
CN106346478A (en) * | 2016-11-09 | 2017-01-25 | 广州视源电子科技股份有限公司 | control method and device of mechanical arm |
CN107378944A (en) * | 2017-06-20 | 2017-11-24 | 东南大学 | A kind of multi-dimensional surface electromyographic signal prosthetic hand control method based on PCA |
KR20190073907A (en) * | 2017-12-19 | 2019-06-27 | 고동휘 | Face Verification with Deep Convolutional Neural Network |
CN109129487A (en) * | 2018-09-26 | 2019-01-04 | 华南理工大学 | Repetitive motion planning method for redundant manipulator based on Taylor's type discrete periodic rhythm and pace of moving things neural network under periodic noise |
CN109623814A (en) * | 2018-12-14 | 2019-04-16 | 中南大学 | A kind of Mechanical arm control method |
CN110000787A (en) * | 2019-04-12 | 2019-07-12 | 清华大学深圳研究生院 | A kind of control method of super redundant mechanical arm |
CN110488611A (en) * | 2019-09-02 | 2019-11-22 | 山东建筑大学 | A kind of biomimetic robot fish movement control method, controller and bionic machine fish |
AU2019101186A4 (en) * | 2019-10-02 | 2020-01-23 | Guo, Zhongliang MR | A Method of Video Recognition Network of Face Tampering Based on Deep Learning |
CN111240195A (en) * | 2020-01-08 | 2020-06-05 | 北京迈格威科技有限公司 | Automatic control model training and target object recycling method and device based on machine vision |
CN111292230A (en) * | 2020-02-18 | 2020-06-16 | 上海交通大学 | Method, system, medium, and apparatus for spiral transform data augmentation in deep learning |
CN112765879A (en) * | 2021-01-12 | 2021-05-07 | 中国人民解放军国防科技大学 | Mechanical arm inverse kinematics solving method based on neural network and configuration coding |
WO2022156475A1 (en) * | 2021-01-20 | 2022-07-28 | 华为技术有限公司 | Neural network model training method and apparatus, and data processing method and apparatus |
CN113400307A (en) * | 2021-06-16 | 2021-09-17 | 清华大学 | Control method of space robot mechanical arm |
CN113500599A (en) * | 2021-07-16 | 2021-10-15 | 中山大学 | Track tracking method and system of rope-driven mechanical arm based on neural network |
CN113771027A (en) * | 2021-08-17 | 2021-12-10 | 浙江工业大学 | Two-arm cooperative grabbing method based on deep learning |
CN113743287A (en) * | 2021-08-31 | 2021-12-03 | 之江实验室 | Robot self-adaptive grabbing control method and system based on impulse neural network |
WO2023046074A1 (en) * | 2021-09-26 | 2023-03-30 | 广州市微眸医疗器械有限公司 | Autonomous learning method for mechanical arm control method |
CN114603564A (en) * | 2022-04-28 | 2022-06-10 | 中国电力科学研究院有限公司 | Mechanical arm navigation obstacle avoidance method and system, computer equipment and storage medium |
CN114888801A (en) * | 2022-05-16 | 2022-08-12 | 南京邮电大学 | Mechanical arm control method and system based on offline strategy reinforcement learning |
Non-Patent Citations (2)
Title |
---|
基于轮廓误差补偿的机械臂自适应非奇异终端滑膜控制;朱大昌等;《机床与液压》;第172-180页 * |
快速换装智能机械手控制系统的设计与实现;姚燕等;《实验室研究与探索》;第41卷(第11期);第100-103页 * |
Also Published As
Publication number | Publication date |
---|---|
CN116512254A (en) | 2023-08-01 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN109719730B (en) | Digital twin robot for flexible assembly process of circuit breaker | |
CN110000785B (en) | Agricultural scene calibration-free robot motion vision cooperative servo control method and equipment | |
CN109960880B (en) | Industrial robot obstacle avoidance path planning method based on machine learning | |
US20210299860A1 (en) | Method and system for robot action imitation learning in three-dimensional space | |
CN113826051A (en) | Generating digital twins of interactions between solid system parts | |
CN109397285B (en) | Assembly method, assembly device and assembly equipment | |
CN112605973A (en) | Robot motor skill learning method and system | |
CN111309035B (en) | Multi-robot cooperative movement and dynamic obstacle avoidance method, device, equipment and medium | |
CN112959326B (en) | Method and device for solving positive kinematics of robot, readable storage medium and robot | |
CN111152227A (en) | Mechanical arm control method based on guided DQN control | |
CN110605711A (en) | Method, device and system for controlling cooperative robot to grab object | |
Skoglund et al. | Programming by demonstration of pick-and-place tasks for industrial manipulators using task primitives | |
CN115847422A (en) | Gesture recognition method, device and system for teleoperation | |
CN114943182B (en) | Robot cable shape control method and equipment based on graph neural network | |
Gutzeit et al. | The besman learning platform for automated robot skill learning | |
CN109764876B (en) | Multi-mode fusion positioning method of unmanned platform | |
CN116512254B (en) | Direction-based intelligent control method and system for mechanical arm, equipment and storage medium | |
CN116337322A (en) | Array type force sensor calibration method and measuring method and device | |
CN113561172B (en) | Dexterous hand control method and device based on binocular vision acquisition | |
Raufelder | Modeling Analysis of Attitude Perception of Engineering Manipulator Supporting Wireless Communication and Internet of Things | |
CN113608622A (en) | Human body posture real-time prediction method, system, medium and equipment | |
Gomes et al. | Deep Reinforcement learning applied to a robotic pick-and-place application | |
CN116383667B (en) | Model training and motion instruction prediction method, device and system | |
CN115256372B (en) | Mechanical arm control method, device, control equipment and storage medium | |
CN116542310B (en) | Model training and motion instruction prediction method, device and system for robot |
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 |