CN110524544A - A kind of control method of manipulator motion, terminal and readable storage medium storing program for executing - Google Patents
A kind of control method of manipulator motion, terminal and readable storage medium storing program for executing Download PDFInfo
- Publication number
- CN110524544A CN110524544A CN201910950422.8A CN201910950422A CN110524544A CN 110524544 A CN110524544 A CN 110524544A CN 201910950422 A CN201910950422 A CN 201910950422A CN 110524544 A CN110524544 A CN 110524544A
- Authority
- CN
- China
- Prior art keywords
- motion
- mechanical arm
- target
- position information
- controlling
- 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
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
- B25J9/1602—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
- B25J9/1656—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
Landscapes
- Engineering & Computer Science (AREA)
- Robotics (AREA)
- Mechanical Engineering (AREA)
- Automation & Control Theory (AREA)
- Manipulator (AREA)
Abstract
The present embodiments relate to robot control fields, disclose the control method, terminal and readable storage medium storing program for executing of a kind of manipulator motion.The control method of manipulator motion in the present invention, comprising: obtain the initial position message of mechanical arm and the target position information of mechanical arm；By in initial position message and target position information input motion trajectory predictions model, motion profile prediction model is the extreme learning machine model that training extremely restrains for predicting the motion profile of mechanical arm in advance；The mechanical arm for obtaining the output of motion profile prediction model moves to the motion profile of target position from initial position, and controls mechanical arm and move according to motion profile.Present embodiment allows mechanical arm quickly to move to target position.
Description
Technical Field
The embodiment of the invention relates to the field of robot control, in particular to a method, a terminal and a readable storage medium for controlling the motion of a mechanical arm.
Background
The aging of population becomes a serious problem facing the world, and the labor cost can be predicted to rise sharply in the near future along with the increase of comprehensive labor demand, so that the intelligent robot plays an important role in the future production life. The human arm can complete simple movement, but is difficult for the robot, and a flexible arm capable of completing tasks in an unstructured space is important. In the robot arm control technology, it is generally necessary to plan a motion trajectory of a robot arm in advance, and then control the robot arm to move according to the planned motion trajectory. The current planning of the motion trajectory of a mechanical arm is mainly based on geometric considerations, and a solution is formed by performing discrete segmentation on a space.
The inventors found that at least the following problems exist in the related art: in the current technology for controlling the motion of the mechanical arm, the planning speed of the motion track of the mechanical arm is very slow, so that the speed of the mechanical arm moving from the initial position to the target position is very slow.
Disclosure of Invention
An object of the embodiments of the present invention is to provide a method, a terminal and a readable storage medium for controlling a motion of a robot arm, so that the robot arm can move to a target position quickly.
In order to solve the above technical problem, an embodiment of the present invention provides a method for controlling a motion of a robot arm, including: acquiring initial position information of the mechanical arm and target position information of the mechanical arm; inputting the initial position information and the target position information into a motion trail prediction model, wherein the motion trail prediction model is an extreme learning machine model which is trained in advance to be converged and used for predicting the motion trail of the mechanical arm; and acquiring the motion trail of the mechanical arm from the initial position to the target position output by the motion trail prediction model, and controlling the mechanical arm to move according to the motion trail.
An embodiment of the present invention further provides a terminal, including: at least one processor; and a memory communicatively coupled to the at least one processor; the storage stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor, so that the at least one processor can execute the control method for the movement of the mechanical arm.
Embodiments of the present invention also provide a computer-readable storage medium storing a computer program, which when executed by a processor implements the above-described method for controlling the movement of a robot arm.
Compared with the prior art, the method and the device have the advantages that the initial position information and the target position information of the mechanical arm are obtained, the initial position information and the target position information are input into the motion trail prediction model, the motion trail output by the motion trail prediction model is obtained, and the mechanical arm is controlled to move according to the motion trail; the motion trail prediction model adopts an extreme learning machine model, and because the network structure of the extreme learning machine model is simple and is a network model structure with a single hidden layer, the training speed of the motion trail prediction model obtained by training the extreme learning machine is high, and meanwhile, the motion trail prediction model can be suitable for mechanical arms with different degrees of freedom, so that the application range is improved; after training, the network structure of the motion trail prediction model is simple, so that an output result can be quickly obtained according to input data, the motion trail of the mechanical arm is very quickly planned, the speed of the mechanical arm moving to a target position is increased, and the speed of controlling the motion of the mechanical arm is greatly shortened.
In addition, the mechanical arm changes the movement track through a plurality of joints, and control mechanical arm moves according to the movement track, including: generating a plurality of joint angles having a mapping relation with the motion track according to the motion track, wherein the joint angles correspond to the joints one by one, and the joint angles are compounded to form the motion track; and adjusting the corresponding joints according to the angles of the joints so as to enable the mechanical arm to move along the motion track. Because the motion of the mechanical arm is related to the joint angles of the joints, a plurality of joint angles which have a mapping relation with the motion track are generated according to the output motion track, the motion track of the mechanical arm can be realized by controlling the transformation of the plurality of joint angles, the control mode is simple, and the motion speed is high.
In addition, the motion trail prediction model randomly determines the node number and the input weight of a hidden layer in the extreme learning machine model before training. Before training, the node number and the input weight of a hidden layer in the extreme learning machine model are randomly determined, and the training speed of a subsequent motion trail prediction model can be further increased due to the fact that the node number and the input weight are not required to be determined through training.
In addition, the training process of the motion trail prediction model specifically includes: taking initial position information and target position information of mechanical arms corresponding to different periods of time in a sample set and a plurality of target motion tracks moving from the initial positions to the target positions as training sets, taking expected target motion tracks moving from the initial positions to the target positions corresponding to different periods of time as verification sets, wherein the expected target motion tracks are the same as the motion tracks of human arms; inputting the data in the training set into an extreme learning machine model to obtain an actual target motion track; and comparing whether the error between the actual target motion track and the expected target motion track is within a preset range, if not, repeatedly and circularly adjusting the parameters of the hidden layer in the extreme learning machine model until the error is within the preset range. By continuously comparing whether the error between the actual target motion track and the expected target motion track is within a preset range, the accuracy of the finally obtained motion track prediction model can be improved.
In addition, the convergence condition in the training of the motion trajectory prediction model includes: the stability of the Lyapunov function is met, and preset constraint conditions are met, wherein the preset constraint conditions are as follows:biindicating the bias of the ith node of the hidden layer, betaiRepresenting the output weight, w, of the ith hidden nodeiInput weight, 0, representing the ith hidden node<i<And N is an integer larger than 1, and the N represents the number of nodes of the hidden layer. By the constraint condition, the convergence rate of the motion trail prediction model can be increased, the training speed is further increased, the stability of the Lyapunov function is met, the motion trail stably output by the motion trail prediction model can be ensured, and the stability of the motion trail prediction model is improved.
In addition, comparing whether the error between the actual target motion trajectory and the expected target motion trajectory is within a preset range specifically includes: acquiring a first velocity vector of each joint in an expected target motion track and a second velocity vector of each joint in an actual target motion track; and calculating an error according to each first speed vector and each second speed vector, and judging whether the error is within a preset range. Because the motion control technology of the mechanical arm has high requirement on stability, the stability of the model can be reflected by the speed change of the joint, and the stability of the model can be more accurately reflected by carrying out error calculation through the speed vector.
In addition, before the initial position information and the target position information of the mechanical arm corresponding to different periods in the sample set are used as a training set, the method for controlling the movement of the mechanical arm further comprises the following steps: and carrying out smoothing processing and/or denoising processing on the collected motion trail of each target. And the accuracy of model training can be improved by carrying out smoothing processing and/or denoising processing.
In addition, before the initial position information and the target position information of the mechanical arm corresponding to different periods in the sample set are used as a training set, the method for controlling the movement of the mechanical arm further comprises the following steps: and adjusting the coordinate system of each joint in each target motion track to be a preset coordinate system. And the coordinate system is unified, so that the motion trail prediction model is convenient to train.
Drawings
One or more embodiments are illustrated by way of example in the accompanying drawings, which correspond to the figures in which like reference numerals refer to similar elements and which are not to scale unless otherwise specified.
Fig. 1 is a detailed flowchart of a method for controlling the movement of a robot arm according to a first embodiment of the present invention;
fig. 2 is a schematic diagram of a specific implementation of comparing whether an error between an actual target motion trajectory and a desired target motion trajectory is within a preset range according to a second embodiment of the present invention;
fig. 3 is a schematic structural diagram of a terminal according to a third embodiment of the present invention.
Detailed Description
In order to make the objects, technical solutions and advantages of the embodiments of the present invention more apparent, embodiments of the present invention will be described in detail below with reference to the accompanying drawings. However, it will be appreciated by those of ordinary skill in the art that numerous technical details are set forth in order to provide a better understanding of the present application in various embodiments of the present invention. However, the technical solution claimed in the present application can be implemented without these technical details and various changes and modifications based on the following embodiments.
The following embodiments are divided for convenience of description, and should not constitute any limitation to the specific implementation manner of the present invention, and the embodiments may be mutually incorporated and referred to without contradiction.
The inventor finds that in the traditional robot arm control technology, a planning algorithm for planning the motion trail of the robot arm is mainly based on geometric consideration and solves the problem by performing discrete segmentation on space, and the method does not consider the kinematic limitations such as the speed of the robot arm. In a working scene needing stable movement work, a more rigorous requirement is provided for the movement speed of the mechanical arm in the movement process; meanwhile, the sampling efficiency of the algorithm in a discrete space is relatively low, the calculation complexity cannot be reduced, and the planning period of the control motion track is prolonged. In addition, the motion trail of the current mechanical arm moving to the target position does not refer to the trail of human motion, but adopts the motion mode violating the motion of human arms, so that the mechanical arm cannot be normally applied to human life, and the application of the mechanical arm is influenced.
A first embodiment of the present invention relates to a method for controlling the movement of a robot arm. The method can be applied to a robot with a mechanical arm and can also be applied to a terminal for controlling the movement of the mechanical arm. The specific flow of the control method of the mechanical arm movement is shown in fig. 1.
Step 101: acquiring initial position information of the mechanical arm and target position information of the mechanical arm.
Specifically, the robot arm has at least 2 degrees of freedom, and a seven-degree-of-freedom robot arm will be described as an example in the present embodiment. The seven-degree-of-freedom arm comprises 7 joints, the position of the tail end mechanism of the mechanical arm can be influenced by the motion of each joint, namely the position of the tail end mechanism of the mechanical arm can be uniquely determined by the position of the 7 joints, and similarly, the position of the tail end mechanism of the mechanical arm can be obtained, and the joint angle of each joint of the mechanical arm can be solved by utilizing reverse kinematics. Therefore, in the present embodiment, the initial position information of the robot arm may refer to an initial joint angle of each joint in the robot arm, or initial coordinate information of each joint, or may refer to coordinate information of an initial position of the end mechanism of the robot arm; similarly, the target position information of the robot arm may be a target joint angle of each joint of the robot arm, or target coordinate information of each joint, or may also be coordinate information of a target position of an end mechanism of the robot arm.
The mechanical arm can acquire the initial position information of the mechanical arm, the target position information of the mechanical arm can be manually input, and the target position information can also be determined by a camera and/or a sensing device of the robot.
It should be noted that, as long as the joint angle of each joint of the mechanical arm is known, the position of the end mechanism of the mechanical arm can be solved by using forward kinematics; similarly, if the position of the end mechanism of the mechanical arm is known, the joint angle of each joint of the mechanical arm can be solved by utilizing inverse kinematics.
Step 102: inputting the initial position information and the target position information into a motion trail prediction model, wherein the motion trail prediction model is an extreme learning machine model which is trained in advance to be converged for predicting the motion trail of the mechanical arm.
In a specific implementation, the motion trajectory prediction model is trained in advance by using a network structure of an extreme learning machine model. Before training, the node number of the hidden layer in the extreme learning machine model and the input weight are randomly determined.
Specifically, the extreme learning machine model is generally of a single-hidden-layer network structure, namely, the extreme learning machine model comprises an input layer, a single hidden layer and an output layer. And randomly determining the number of nodes in the hidden layer and the input weight in the extreme learning machine model. The principle of the extreme learning machine is described below:
for example, if the input isGiven an excitation function of g, the objective function is expressed as shown in equation (1):
where N represents the number of nodes of the hidden layer, bi represents the bias of the ith node of the hidden layer, wiRepresenting the input weight, β, of the i-th hidden nodeiRepresenting the output weight of the i-th hidden node.
The learning goal of the extreme learning machine model is to minimize the error of the output, which can be expressed as formula (2):
min||HβT-o | | formula (2);
where o is the target value of the output layer, and o ═ o1,…oN)TH is the output of the hidden layer, H (w, b), expressed as shown in equation (3),
the excitation function satisfies formula (4):
the training process of the motion trajectory prediction model may be: taking initial position information and target position information of mechanical arms corresponding to different periods of time in a sample set and a plurality of target motion tracks moving from the initial positions to the target positions as training sets, taking expected target motion tracks moving from the initial positions to the target positions corresponding to different periods of time as verification sets, wherein the expected target motion tracks are the same as the motion tracks of human arms; inputting the data in the training set into an extreme learning machine model to obtain an actual target motion track; and comparing whether the error between the actual target motion track and the expected target motion track is within a preset range, if not, repeatedly and circularly adjusting the parameters of the hidden layer in the extreme learning machine model until the error is within the preset range.
In particular, the training limit study can be artificially acquiredThe sample set of the learning machine does not need to collect the motion trail of the mechanical arm in a discrete space, and can directly collect the motion trail of the paired targets moving from different initial positions to the corresponding target positions in different time periods. For example, Xi＝X1,X2…Xn,Ti＝T1,T2…Tn,i＝1,2…NtraIn which N istraRepresenting the total number of initial positions and corresponding target positions, i representing a sample of the target motion trajectory from the ith initial position to the corresponding target position, TiMatrix formed for a plurality of acquisition instants, XnForming a matrix for the position information of each joint in a target motion track, wherein n is the number of samples of the target motion track from the initial position to the corresponding target position.
In the training process, a plurality of initial positions and corresponding target positions are input into the extreme learning machine model, an actual target motion track is output, the actual target motion track is compared with an expected target motion track, and whether the error between the actual target motion track and the expected target motion track is within a preset range or not is judged. The expected target motion trajectory is the same as the motion trajectory of the human arm, for example, the initial position of the mechanical arm end mechanism is point a, the corresponding target position is point B, if 20 item target motion trajectories from point a to point B are collected, a motion trajectory which does not conform to the kinematics principle of the human arm exists in the 20 item target motion trajectories, or a target motion trajectory which can collide with an obstacle exists, and if the target motion trajectory 1 is the same as the motion trajectory of the human arm from point a to point B, the target motion trajectory 1 is taken as the expected target motion trajectory.
It should be noted that, in the training process, the convergence condition of the motion trajectory prediction model includes: the stability of the Lyapunov function is met, and preset constraint conditions are met, wherein the preset constraint conditions are as follows:biindicating a hidden layer bias, betaiWeight, w, of the output layeriWeight of input layer, 0<i<N, N is an integer greater than 1 and N representsThe number of nodes of the hidden layer.
Specifically, as the mechanical arm movement needs stability, in the training process, whether the output of the trained extreme learning model is stable or not is verified through the Lyapunov function, and meanwhile, preset constraint conditions need to be met, and the constraint conditions are used for assisting the extreme learning machine to quickly converge.
In this embodiment the output weight is set to β - α w to start the solution, where α is a positive parameter, and this initial value satisfies the constraint condition at the same time, α can be changed to determine the initial value of the failure. The position information of each joint in the sample set can be normalized to a small interval, and meanwhile, the input weight is initialized in a small interval, so that the matrix H of the hidden layer is a nonsingular matrix, and the learning precision of the motion trail prediction model is improved.
Step 103: and acquiring the motion trail of the mechanical arm from the initial position to the target position output by the motion trail prediction model, and controlling the mechanical arm to move according to the motion trail.
Specifically, the motion trajectory output by the motion trajectory prediction model may be position information of the mechanical arm at each discrete time, where the position information may be a joint angle of each joint of the mechanical arm, or a coordinate of each joint of the mechanical arm in a preset coordinate system. If the position information is the joint angle of each joint of the robot arm, the joint angle of each joint of the robot arm may be adjusted directly according to the discrete-time position information of the upper robot arm.
In a specific implementation, if the position information is coordinates of each joint of the mechanical arm in a preset coordinate system; the mechanical arm changes the motion track through a plurality of joints, and the mechanical arm is controlled to move according to the motion track, and the specific process can be as follows: generating a plurality of joint angles having a mapping relation with the motion track according to the motion track, wherein the joint angles correspond to the joints one by one, and the joint angles are compounded to form the motion track; and adjusting the corresponding joints according to the angles of the joints so as to enable the mechanical arm to move along the motion track.
Specifically, the reverse inverse solution can be performed according to the position information of the mechanical arm, and the joint angle of each joint of the mechanical arm can be determined, so that the joint angle of each joint of the mechanical arm at each moment can be determined according to the motion trail. The plurality of joint angles are compounded to form a motion track, and the corresponding joints are adjusted through the plurality of joint angles so that the mechanical arm moves along the motion track.
Compared with the prior art, the method and the device have the advantages that the initial position information and the target position information of the mechanical arm are obtained, the initial position information and the target position information are input into the motion trail prediction model, the motion trail output by the motion trail prediction model is obtained, and the mechanical arm is controlled to move according to the motion trail; the motion trail prediction model adopts an extreme learning machine model, and because the network structure of the extreme learning machine model is simple and is a network model structure with a single hidden layer, the training speed of the motion trail prediction model obtained by training the extreme learning machine is high, and meanwhile, the motion trail prediction model can be suitable for mechanical arms with different degrees of freedom, so that the application range is improved; after training, the network structure of the motion trail prediction model is simple, so that an output result can be quickly obtained according to input data, the motion trail of the mechanical arm is very quickly planned, the speed of the mechanical arm moving to a target position is increased, and the speed of controlling the motion of the mechanical arm is greatly shortened.
A second embodiment of the present invention relates to a method for controlling the movement of a robot arm. The method for controlling the movement of the robot arm in the second embodiment includes: acquiring initial position information of the mechanical arm and target position information of the mechanical arm; inputting the initial position information and the target position information into a motion trail prediction model, wherein the motion trail prediction model is an extreme learning machine model which is trained in advance to be converged and used for predicting the motion trail of the mechanical arm; and acquiring the motion trail of the mechanical arm from the initial position to the target position output by the motion trail prediction model, and controlling the mechanical arm to move according to the motion trail.
The second embodiment is a further improvement of the training motion trajectory prediction model in the first embodiment, and the main improvement is that: in the second embodiment of the present invention, whether the error between the actual target motion trajectory and the expected target motion trajectory is within the preset range is determined by calculating the error using the first velocity vector of each joint in the expected target motion trajectory and the second velocity vector of each joint in the actual target motion trajectory, and determining whether the error is within the preset range. Fig. 2 is a schematic diagram of a specific implementation of comparing whether an error between an actual target motion trajectory and a desired target motion trajectory is within a preset range.
Step 201: and acquiring a first velocity vector of each joint in the expected target motion trail and a second velocity vector of each joint in the actual target motion trail.
Specifically, the velocity vector of each joint can be calculated using a velocity vector formula, as shown in formula (5),
where x represents the pose of the current joint and t represents the current time.
From this equation (5), a first velocity vector for each joint in the desired target motion profile and a second velocity vector for each joint in the actual motion profile of the target can be determined.
Step 202: and calculating an error according to each first speed vector and each second speed vector, and judging whether the error is within a preset range.
Because the motion of the mechanical arm requires stability, and the trend of the velocity change of each joint of the mechanical arm is very sensitive to the stability and precision of the model, the accuracy of the baseline learning machine model in the training process can be judged by using the velocity vector, and the error calculation formula can be as shown in formula (6):
wherein the first velocity vectorAnd a second velocity vectorr and q are two parameters, which in this embodiment may be set to 0.6 and 0.4, respectively, epsilon is a very small positive parameter, N represents the number of samples,and judging whether the error is in a preset range or not for calculating the obtained error, wherein the preset range can be set according to actual needs.
It is worth mentioning that after the motion tracks of the targets are collected, the data in the sample set can be preprocessed before the initial position information and the target position information of the mechanical arms corresponding to different periods in the sample set are used as training sets, so that the precision of subsequent training is improved.
The pre-processing may include: and carrying out smoothing processing and/or denoising processing on the collected motion trail of each target. Each target motion track can be filtered and trimmed, and the standardized residual error is utilized for denoising.
For convenience of calculation, the coordinate system where each joint in each target motion track is located can be adjusted to be a preset coordinate system. The step of unifying the coordinate system may also precede the smoothing process and/or the denoising process.
In the method for controlling the motion of the mechanical arm provided in the embodiment, the requirement on the stability is high in the motion control of the mechanical arm, the stability of the model can be reflected by the speed change of the joint, and the stability of the model can be reflected more accurately by performing error calculation through the speed vector.
The steps of the above methods are divided for clarity, and the implementation may be combined into one step or split some steps, and the steps are divided into multiple steps, so long as the same logical relationship is included, which are all within the protection scope of the present patent; it is within the scope of the patent to add insignificant modifications to the algorithms or processes or to introduce insignificant design changes to the core design without changing the algorithms or processes.
A third embodiment of the present invention relates to a terminal, a specific structure of which is shown in fig. 3, and the terminal includes: at least one processor 301; and a memory 302 communicatively coupled to the at least one processor 301; the memory 302 stores instructions executable by the at least one processor 301, and the instructions are executed by the at least one processor 301, so that the at least one processor 301 can execute the control method of the robot arm motion in the first embodiment or the second embodiment.
The memory 302 and the processor 301 are connected by a bus, which may include any number of interconnected buses and bridges that link one or more of the various circuits of the processor 301 and the memory 302. The bus may also link various other circuits such as peripherals, voltage regulators, power management circuits, and the like, which are well known in the art, and therefore, will not be described any further herein. A bus interface provides an interface between the bus and the transceiver. The transceiver may be one element or a plurality of elements, such as a plurality of receivers and transmitters, providing a means for communicating with various other apparatus over a transmission medium. The data processed by the processor 301 is transmitted over a wireless medium through an antenna, which further receives the data and transmits the data to the processor 301.
The processor 301 is responsible for managing the bus and general processing and may also provide various functions including timing, peripheral interfaces, voltage regulation, power management, and other control functions. And memory 302 may be used to store data used by processor 301 in performing operations.
A fourth embodiment of the present invention relates to a computer-readable storage medium storing a computer program which, when executed by a processor, implements a method for controlling the movement of a robot arm according to the first or second embodiment.
Those skilled in the art can understand that all or part of the steps in the method of the foregoing embodiments may be implemented by a program to instruct related hardware, where the program is stored in a storage medium and includes several instructions to enable a device (which may be a single chip, a chip, etc.) or a processor (processor) to execute all or part of the steps of the method described in the embodiments of the present application. And the aforementioned storage medium includes: a U-disk, a removable hard disk, a Read-only Memory (ROM), a Random Access Memory (RAM), a magnetic disk or an optical disk, and other various media capable of storing program codes.
It will be understood by those of ordinary skill in the art that the foregoing embodiments are specific examples for carrying out the invention, and that various changes in form and details may be made therein without departing from the spirit and scope of the invention in practice.
Claims (10)
1. A method for controlling the motion of a mechanical arm is characterized by comprising the following steps:
acquiring initial position information of a mechanical arm and target position information of the mechanical arm;
inputting the initial position information and the target position information into a motion trail prediction model, wherein the motion trail prediction model is an extreme learning machine model which is trained in advance to be converged for predicting the motion trail of the mechanical arm;
and acquiring the motion trail of the mechanical arm from the initial position to the target position output by the motion trail prediction model, and controlling the mechanical arm to move according to the motion trail.
2. The method for controlling the motion of a robot arm according to claim 1, wherein the robot arm changes a motion trajectory by a plurality of joints, and the controlling the robot arm to move according to the motion trajectory comprises:
generating a plurality of joint angles having a mapping relation with the motion trail according to the motion trail, wherein the joint angles correspond to the joints one by one, and the joint angles are compounded to form the motion trail;
and adjusting the corresponding joints according to the plurality of joint angles so as to enable the mechanical arm to move along the motion track.
3. The method for controlling the movement of a robot arm according to claim 1, wherein the movement trajectory prediction model randomly determines the number of nodes of a hidden layer and the input weight in the extreme learning machine model before training.
4. The method for controlling the motion of the mechanical arm according to any one of claims 1 to 3, wherein the training process of the motion trajectory prediction model specifically comprises:
taking initial position information and target position information of the mechanical arm corresponding to different periods in a sample set and a plurality of target motion tracks moving from the initial position to the target position as training sets, and taking expected target motion tracks moving from the initial position to the target position corresponding to the different periods as verification sets, wherein the expected target motion tracks are the same as the motion tracks of human arms;
inputting the data in the training set into an extreme learning machine model to obtain an actual target motion track;
and comparing whether the error between the actual target motion track and the expected target motion track is within a preset range, if not, repeatedly and circularly adjusting the parameters of a hidden layer in the extreme learning machine model until the error is within the preset range.
5. The method for controlling the motion of a robot arm according to any one of claims 1 to 4, wherein the convergence condition in the training of the motion trajectory prediction model includes:
the stability of the Lyapunov function is met, and preset constraint conditions are met, wherein the preset constraint conditions are as follows:biindicating the bias of the ith node of the hidden layer, betaiRepresenting the output weight, w, of the ith hidden nodeiInput weight, 0, representing the ith hidden node<i<And N is an integer larger than 1, and the N represents the number of nodes of the hidden layer.
6. The method for controlling motion of a mechanical arm according to claim 4, wherein the comparing whether the error between the actual target motion trajectory and the expected target motion trajectory is within a preset range specifically comprises:
acquiring a first velocity vector of each joint in an expected target motion track and a second velocity vector of each joint in the actual target motion track;
and calculating the error according to each first speed vector and each second speed vector, and judging whether the error is in a preset range.
7. The method of controlling robot arm movement according to claim 4, wherein before the step of using the initial position information and the target position information of the robot arm corresponding to different periods in the sample set as a training set, the method further comprises:
and carrying out smoothing processing and/or denoising processing on the collected motion trail of each target.
8. The method for controlling robot arm movement according to claim 6 or 7, wherein before the step of using the initial position information and the target position information of the robot arm corresponding to different periods in the sample set as a training set, the method for controlling robot arm movement further comprises:
and adjusting the coordinate system of each joint in each target motion track to be a preset coordinate system.
9. A terminal, comprising:
at least one processor; and the number of the first and second groups,
a memory communicatively coupled to the at least one processor; wherein,
the memory stores instructions executable by the at least one processor to enable the at least one processor to perform a method of controlling movement of a robotic arm as claimed in any one of claims 1 to 8.
10. A computer-readable storage medium storing a computer program, wherein the computer program, when executed by a processor, implements the method of controlling the motion of a robotic arm of any one of claims 1 to 8.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910950422.8A CN110524544A (en) | 2019-10-08 | 2019-10-08 | A kind of control method of manipulator motion, terminal and readable storage medium storing program for executing |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910950422.8A CN110524544A (en) | 2019-10-08 | 2019-10-08 | A kind of control method of manipulator motion, terminal and readable storage medium storing program for executing |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN110524544A true CN110524544A (en) | 2019-12-03 |
Family
ID=68671388
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910950422.8A Pending CN110524544A (en) | 2019-10-08 | 2019-10-08 | A kind of control method of manipulator motion, terminal and readable storage medium storing program for executing |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110524544A (en) |
Cited By (16)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110967042A (en) * | 2019-12-23 | 2020-04-07 | 襄阳华中科技大学先进制造工程研究院 | Industrial robot positioning precision calibration method, device and system |
| CN110986953A (en) * | 2019-12-13 | 2020-04-10 | 深圳前海达闼云端智能科技有限公司 | Path planning method, robot and computer readable storage medium |
| CN111037589A (en) * | 2019-12-11 | 2020-04-21 | 西安工程大学 | A robot end structure and its control method |
| CN112077841A (en) * | 2020-08-10 | 2020-12-15 | 北京大学 | Multi-joint linkage method and system for manipulator precision of elevator robot arm |
| CN112870595A (en) * | 2020-12-30 | 2021-06-01 | 国电南瑞科技股份有限公司 | Control method, device and system for elevating fire-fighting robot |
| CN113070878A (en) * | 2021-03-26 | 2021-07-06 | 中国科学院深圳先进技术研究院 | Robot control method based on impulse neural network, robot and storage medium |
| CN113276120A (en) * | 2021-05-25 | 2021-08-20 | 中国煤炭科工集团太原研究院有限公司 | Control method and device for mechanical arm movement and computer equipment |
| CN113625699A (en) * | 2020-04-24 | 2021-11-09 | 库卡机器人制造（上海）有限公司 | Motion control method, control device and computer readable medium |
| CN113715029A (en) * | 2021-09-26 | 2021-11-30 | 中山大学 | Autonomous learning method of mechanical arm control method |
| CN114029961A (en) * | 2021-12-16 | 2022-02-11 | 滁州学院 | Robot control method and system for high-precision transmission of mechanical arm |
| CN114378830A (en) * | 2022-02-18 | 2022-04-22 | 深圳市大族机器人有限公司 | Robot wrist joint singularity avoidance method and system |
| WO2022088593A1 (en) * | 2020-10-26 | 2022-05-05 | 北京理工大学 | Robotic arm control method and device, and human-machine cooperation model training method |
| CN114687538A (en) * | 2020-12-29 | 2022-07-01 | 广东博智林机器人有限公司 | A working method, device, equipment and medium for floor paint coating equipment |
| CN116000911A (en) * | 2021-10-22 | 2023-04-25 | 瑞龙诺赋(上海)医疗科技有限公司 | Mechanical arm control method and device and mechanical arm |
| CN116864440A (en) * | 2023-09-04 | 2023-10-10 | 泓浒(苏州)半导体科技有限公司 | Automated handling system, method, apparatus and medium for semiconductor workpieces |
| CN118809612A (en) * | 2024-08-19 | 2024-10-22 | 北京积加科技有限公司 | Method, device and computer-readable medium for moving opening and closing components based on a mechanical arm |
Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103737592A (en) * | 2013-12-27 | 2014-04-23 | 柳州职业技术学院 | Manipulator precise control system and method |
| KR20170108526A (en) * | 2016-03-18 | 2017-09-27 | 한국과학기술원 | Method for Generating Robot Task Motion Based on Imitation Learning and Motion Composition and Apparatus Therefor |
| CN107457780A (en) * | 2017-06-13 | 2017-12-12 | 广州视源电子科技股份有限公司 | Method and device for controlling mechanical arm movement, storage medium and terminal equipment |
| CN107457781A (en) * | 2017-07-12 | 2017-12-12 | 国机智能技术研究院有限公司 | A kind of method and system of control machine people motion |
| CN108115681A (en) * | 2017-11-14 | 2018-06-05 | 深圳先进技术研究院 | Learning by imitation method, apparatus, robot and the storage medium of robot |
| CN109702740A (en) * | 2018-12-14 | 2019-05-03 | 中国科学院深圳先进技术研究院 | Robot compliance control method, device, equipment and storage medium |
| CN109960880A (en) * | 2019-03-26 | 2019-07-02 | 上海交通大学 | An obstacle avoidance path planning method for industrial robots based on machine learning |
| CN110125930A (en) * | 2019-04-18 | 2019-08-16 | 华中科技大学 | It is a kind of that control method is grabbed based on the mechanical arm of machine vision and deep learning |
-
2019
- 2019-10-08 CN CN201910950422.8A patent/CN110524544A/en active Pending
Patent Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103737592A (en) * | 2013-12-27 | 2014-04-23 | 柳州职业技术学院 | Manipulator precise control system and method |
| KR20170108526A (en) * | 2016-03-18 | 2017-09-27 | 한국과학기술원 | Method for Generating Robot Task Motion Based on Imitation Learning and Motion Composition and Apparatus Therefor |
| CN107457780A (en) * | 2017-06-13 | 2017-12-12 | 广州视源电子科技股份有限公司 | Method and device for controlling mechanical arm movement, storage medium and terminal equipment |
| CN107457781A (en) * | 2017-07-12 | 2017-12-12 | 国机智能技术研究院有限公司 | A kind of method and system of control machine people motion |
| CN108115681A (en) * | 2017-11-14 | 2018-06-05 | 深圳先进技术研究院 | Learning by imitation method, apparatus, robot and the storage medium of robot |
| CN109702740A (en) * | 2018-12-14 | 2019-05-03 | 中国科学院深圳先进技术研究院 | Robot compliance control method, device, equipment and storage medium |
| CN109960880A (en) * | 2019-03-26 | 2019-07-02 | 上海交通大学 | An obstacle avoidance path planning method for industrial robots based on machine learning |
| CN110125930A (en) * | 2019-04-18 | 2019-08-16 | 华中科技大学 | It is a kind of that control method is grabbed based on the mechanical arm of machine vision and deep learning |
Non-Patent Citations (1)
| Title |
|---|
| 胡建兵: "学习人类策略在机械臂的应用研究", 《中国优秀硕士学位论文全文数据库 信息科技辑》 * |
Cited By (22)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111037589A (en) * | 2019-12-11 | 2020-04-21 | 西安工程大学 | A robot end structure and its control method |
| CN110986953A (en) * | 2019-12-13 | 2020-04-10 | 深圳前海达闼云端智能科技有限公司 | Path planning method, robot and computer readable storage medium |
| CN110967042A (en) * | 2019-12-23 | 2020-04-07 | 襄阳华中科技大学先进制造工程研究院 | Industrial robot positioning precision calibration method, device and system |
| CN113625699A (en) * | 2020-04-24 | 2021-11-09 | 库卡机器人制造（上海）有限公司 | Motion control method, control device and computer readable medium |
| CN112077841A (en) * | 2020-08-10 | 2020-12-15 | 北京大学 | Multi-joint linkage method and system for manipulator precision of elevator robot arm |
| WO2022088593A1 (en) * | 2020-10-26 | 2022-05-05 | 北京理工大学 | Robotic arm control method and device, and human-machine cooperation model training method |
| CN114687538B (en) * | 2020-12-29 | 2023-08-15 | 广东博智林机器人有限公司 | Working method, device, equipment and medium of floor paint coating equipment |
| CN114687538A (en) * | 2020-12-29 | 2022-07-01 | 广东博智林机器人有限公司 | A working method, device, equipment and medium for floor paint coating equipment |
| CN112870595A (en) * | 2020-12-30 | 2021-06-01 | 国电南瑞科技股份有限公司 | Control method, device and system for elevating fire-fighting robot |
| CN112870595B (en) * | 2020-12-30 | 2022-07-08 | 国电南瑞科技股份有限公司 | Control method, device and system for elevating fire-fighting robot |
| CN113070878A (en) * | 2021-03-26 | 2021-07-06 | 中国科学院深圳先进技术研究院 | Robot control method based on impulse neural network, robot and storage medium |
| CN113276120A (en) * | 2021-05-25 | 2021-08-20 | 中国煤炭科工集团太原研究院有限公司 | Control method and device for mechanical arm movement and computer equipment |
| CN113715029A (en) * | 2021-09-26 | 2021-11-30 | 中山大学 | Autonomous learning method of mechanical arm control method |
| 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 |
| CN116000911A (en) * | 2021-10-22 | 2023-04-25 | 瑞龙诺赋(上海)医疗科技有限公司 | Mechanical arm control method and device and mechanical arm |
| CN114029961A (en) * | 2021-12-16 | 2022-02-11 | 滁州学院 | Robot control method and system for high-precision transmission of mechanical arm |
| CN114378830A (en) * | 2022-02-18 | 2022-04-22 | 深圳市大族机器人有限公司 | Robot wrist joint singularity avoidance method and system |
| CN114378830B (en) * | 2022-02-18 | 2024-02-20 | 深圳市大族机器人有限公司 | Robot wrist joint singular avoidance method and system |
| CN116864440A (en) * | 2023-09-04 | 2023-10-10 | 泓浒(苏州)半导体科技有限公司 | Automated handling system, method, apparatus and medium for semiconductor workpieces |
| CN116864440B (en) * | 2023-09-04 | 2023-11-14 | 泓浒(苏州)半导体科技有限公司 | Automated handling system, method, apparatus and medium for semiconductor workpieces |
| CN118809612A (en) * | 2024-08-19 | 2024-10-22 | 北京积加科技有限公司 | Method, device and computer-readable medium for moving opening and closing components based on a mechanical arm |
| CN118809612B (en) * | 2024-08-19 | 2025-03-11 | 北京积加科技有限公司 | Mechanical arm-based opening and closing assembly moving method, equipment and computer readable medium |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110524544A (en) | A kind of control method of manipulator motion, terminal and readable storage medium storing program for executing | |
| CN111536964B (en) | Robot positioning method and device, and storage medium | |
| CN111830822B (en) | System, method and computer readable medium for configuring interaction with environment | |
| CN115781685B (en) | A high-precision robotic arm control method and system based on reinforcement learning | |
| CN115256395B (en) | Safety control method of robot with model uncertainty based on control obstacle function | |
| CN110688920A (en) | An unmanned driving control method, device and server | |
| CN113219506A (en) | Positioning method for multimode fusion seamless switching | |
| CN116661469B (en) | Robot track error correction method and system | |
| CN112857370A (en) | Robot map-free navigation method based on time sequence information modeling | |
| CN117733870A (en) | Mechanical arm target capturing method, device and medium based on track prediction control | |
| CN110940341A (en) | Path planning method, robot and computer readable storage medium | |
| CN104898659B (en) | A kind of man-machine coordination control method based on model prediction | |
| CN118884976B (en) | Method for tracking moving target by unmanned aerial vehicle by utilizing visual information | |
| CN114137950A (en) | Method and equipment for carrying out social perception model predictive control on robot equipment | |
| CN111596691B (en) | A decision modeling and collaborative control method and system for a human-in-the-loop multi-robot system | |
| Li et al. | Model predictive control for constrained robot manipulator visual servoing tuned by reinforcement learning | |
| CN111221318A (en) | Multi-robot state estimation method based on model predictive control algorithm | |
| CN116483108A (en) | State correction method and device for quadruped robot, electronic equipment and storage medium | |
| CN116972853A (en) | Planning method for local obstacle avoidance path of mobile robot in unknown environment | |
| Xu et al. | UAV local path planning based on improved proximal policy optimization algorithm | |
| Zhang et al. | AirPilot: Interpretable PPO-based DRL Auto-Tuned Nonlinear PID Drone Controller for Robust Autonomous Flights | |
| CN118700133B (en) | Path planning method for redundantly driven robotic arm based on DSAW offline reinforcement learning algorithm | |
| CN114700960B (en) | Order-optimized remote control behavior contour planning method | |
| Spanogianopoulos | Particle swarm optimization and applications in robotics: A survey | |
| Sengupta et al. | UniLCD: Unified Local-Cloud Decision-Making via Reinforcement Learning |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| TA01 | Transfer of patent application right | ||
| TA01 | Transfer of patent application right |
Effective date of registration: 20210205 Address after: 200245 2nd floor, building 2, no.1508, Kunyang Road, Minhang District, Shanghai Applicant after: Dalu Robot Co.,Ltd. Address before: 518000 Room 201, building A, No. 1, Qian Wan Road, Qianhai Shenzhen Hong Kong cooperation zone, Shenzhen, Guangdong (Shenzhen Qianhai business secretary Co., Ltd.) Applicant before: CLOUDMINDS (SHENZHEN) ROBOTICS SYSTEMS Co.,Ltd. |
|
| RJ01 | Rejection of invention patent application after publication | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20191203 |