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 profile
- mechanical arm
- motion
- control method
- target
- 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 present embodiments relate to robot control field, in particular to a kind of control method of manipulator motion, terminal
And readable storage medium storing program for executing.
Background technique
Aging of population has become the serious problem that the whole world faces, it is anticipated that in the near future with comprehensive
The increase of conjunction property labour demand, labor cost can also steeply rise, it is seen then that intelligent robot is in following production and living
Important role will be played the part of.Human arm can be completed simply to move, but relatively difficult for robot, and one
The flexible arm that task can be completed in unstructured space is just particularly important.In robotic arm control technology, usually need
The motion profile of mechanical arm is planned in advance, and then control mechanical arm and moved according to the motion profile of planning.Machine at present
The planning of the motion profile of tool arm, which is mainly based upon, geometrically to be considered, is solved by carrying out separate division to space to be formed
Scheme.
At least there are the following problems in the related technology for inventor's discovery: in the technology controlled at present manipulator motion, machine
The planning speed of tool arm motion track is very slow, and mechanical brachial plexus initial position is caused to move to the speed of target position very
Slowly.
Summary of the invention
The control method, terminal and readable storage medium for being designed to provide a kind of manipulator motion of embodiment of the present invention
Matter allows mechanical arm quickly to move to target position.
In order to solve the above technical problems, embodiments of the present invention provide a kind of control method of manipulator motion, packet
It includes: obtaining the initial position message of mechanical arm and the target position information of mechanical arm；By initial position message and target position
In information input motion profile prediction model, motion profile prediction model is the fortune that training extremely restrains for predicting mechanical arm in advance
The extreme learning machine model of dynamic rail mark；The mechanical arm for obtaining the output of motion profile prediction model moves to target position from initial position
The motion profile set, and control mechanical arm and moved according to motion profile.
Embodiments of the present invention additionally provide a kind of terminal, comprising: at least one processor；And at least one
The memory of processor communication connection；Wherein, memory is stored with the instruction that can be executed by least one processor, instructs by extremely
A few processor executes, so that at least one processor is able to carry out the control method of above-mentioned manipulator motion.
Embodiments of the present invention additionally provide a kind of computer readable storage medium, are stored with computer program, calculate
Machine program realizes the control method of above-mentioned manipulator motion when being executed by processor.
Embodiment of the present invention in terms of existing technologies, passes through the initial position message and the machine of acquisition mechanical arm
The target position information of tool arm is obtained in the initial position message and target position information input motion trajectory predictions model
The motion profile of motion profile prediction model output controls the mechanical arm and moves according to the motion profile；Motion profile predicts mould
Type uses extreme learning machine model, since the network structure of extreme learning machine model is simple, for the network model of single hidden layer
Structure, so that it is fast using the training speed of the motion profile prediction model of extreme learning machine training acquisition, meanwhile, the movement rail
Mark prediction model can be adapted for different degree-of-freedom manipulators, improve the scope of application；The motion profile predicts mould after having trained
The network structure of type is simple, so that according to the input data, can quickly be exported as a result, making the movement rail to mechanical arm
Mark planning is very fast, improves the speed of manipulator motion to target position, substantially reduces the speed of control manipulator motion.
In addition, mechanical arm, by multiple joint changing path of movement, control mechanical arm is moved according to motion profile, comprising:
According to motion profile generate with motion profile have mapping relations multiple joint angles, wherein multiple joint angles with it is multiple
Joint corresponds, and multiple joint angles are compounded to form motion profile；Corresponding joint is adjusted according to multiple joint angles, so that machine
Tool arm is moved along motion profile.It is raw according to the motion profile of output since the movement of mechanical arm and the joint angles in joint are related
At multiple joint angles with motion profile with mapping relations, machinery can be realized by controlling the transformation of multiple joint angles
The motion profile of arm, control mode is simple, and movement velocity is fast.
In addition, motion profile prediction model is before training, the random node for determining hidden layer in extreme learning machine model
Several and input weight.Before training, the random number of nodes and input weight for determining hidden layer in extreme learning machine model, by
In without determining by training, the speed of subsequent motion trajectory predictions model training can be further improved.
In addition, the training process of motion profile prediction model, specifically includes: different periods in sample set are corresponding
Initial position message, target position information and a plurality of target movement rail that target position is moved to from initial position of mechanical arm
Mark makees the corresponding expectation target motion profile for moving to target position from initial position of different periods as training set
Collect for verifying, expectation target motion profile is identical as the motion profile of human arm；By the data input limits in training set
In habit machine model, realistic objective motion profile is obtained；It compares between realistic objective motion profile and expectation target motion profile
Error whether within a preset range, if it is not, then iterative cycles adjustment extreme learning machine model in hidden layer parameter, until accidentally
Terminate when difference is within preset range.By constantly comparing the mistake between realistic objective motion profile and expectation target motion profile
The accuracy of the motion profile prediction model finally obtained can be improved whether within preset range in difference.
In addition, the condition of convergence in the training of motion profile prediction model includes: to meet the stabilization of Li Yapu love function
Property and meet default constraint condition, preset constraint condition are as follows:biIndicate the biasing of i-th of node of hidden layer,
βiIndicate the output weight of i-th of concealed nodes, wiIndicate the input weight of i-th of concealed nodes, 0 < i < N, N are whole greater than 1
Number, N indicate the number of nodes of hidden layer.By the constraint condition, the convergence rate of motion profile prediction model can be accelerated, into one
Step improves training speed, and meets the stability of Li Yapu love function, it is ensured that motion profile prediction model stablizes output
Motion profile, improve motion profile prediction model stability.
In addition, whether the error between comparison realistic objective motion profile and expectation target motion profile is in preset range
It is interior, it specifically includes: obtaining the First Speed vector and realistic objective motion profile in each joint in expectation target motion profile
In each joint second speed vector；According to each First Speed vector and each second speed vector, error is calculated, and
Within a preset range whether error in judgement.Due in the movement control technology of mechanical arm, to stability requirement height, and joint
Velocity variations can reflect out the stability of model, carries out error calculation by velocity vector, can relatively accurately reflect
The stability of the model.
In addition, by the initial position message and target position information of the corresponding mechanical arm of different periods in sample set
Before training set, the control method of manipulator motion further include: every target trajectory of acquisition is smoothed
And/or denoising.It is smoothed and/or denoising, the accuracy of model training can be improved.
In addition, by the initial position message and target position information of the corresponding mechanical arm of different periods in sample set
Before training set, the control method of manipulator motion further include: by coordinate where joint each in every target trajectory
System is adjusted to preset coordinate system.Unified coordinate system, convenient for training motion profile prediction model.
Detailed description of the invention
One or more embodiments are illustrated by the picture in corresponding attached drawing, these exemplary theorys
The bright restriction not constituted to embodiment, the element in attached drawing with same reference numbers label are expressed as similar element, remove
Non- to have special statement, composition does not limit the figure in attached drawing.
Fig. 1 is the specific flow chart of the control method for the manipulator motion that first embodiment provides according to the present invention；
Fig. 2 be compare in second embodiment according to the present invention realistic objective motion profile and expectation target motion profile it
Between error whether a kind of specific implementation schematic diagram within a preset range；
Fig. 3 is a kind of concrete structure schematic diagram for terminal that third embodiment provides according to the present invention.
Specific embodiment
In order to make the object, technical scheme and advantages of the embodiment of the invention clearer, below in conjunction with attached drawing to the present invention
Each embodiment be explained in detail.However, it will be understood by those skilled in the art that in each embodiment party of the present invention
In formula, in order to make the reader understand this application better, many technical details are proposed.But even if without these technical details
And various changes and modifications based on the following respective embodiments, the application technical solution claimed also may be implemented.
The division of each embodiment is for convenience, should not to constitute to specific implementation of the invention any below
It limits, each embodiment can be combined with each other mutual reference under the premise of reconcilable.
Inventor has found the planning algorithm planned manipulator motion track in traditional robotic arm control technology
It is mainly based upon and geometrically considers, separate division is carried out to solve the problems, such as by space, this method is for mechanical arm
The limitation of the kinematics such as speed is not contemplated.In the operative scenario for needing smooth motion to work, to moving for mechanical arm
Movement velocity proposes harsher requirement in the process；The efficiency that algorithm samples in discrete space simultaneously is relatively relatively low, cannot
The complexity calculated is reduced, the planning horizon of control motion profile is extended.In addition, current manipulator motion is to target position
Motion profile and track without reference to human motion, but the movement carried out by the way of violating human arm movement, this
Cause mechanical arm can not normal use into human lives, affect the application of mechanical arm.
The first embodiment of the present invention is related to a kind of control methods of manipulator motion.This method, which can be applied, to be had
In the robot of mechanical arm, also it can be applied in the terminal of control manipulator motion.The control method of the manipulator motion
Detailed process is as shown in Figure 1.
Step 101: obtaining the initial position message of mechanical arm and the target position information of mechanical arm.
Specifically, mechanical arm is at least the mechanical arm of 2DOF, in present embodiment by taking seven freedom mechanical arm as an example
It is illustrated.Seven freedom arm includes 7 joints, and the movement in each joint will affect the position of mechanical arm tail end mechanism,
The position in i.e. 7 joints can uniquely determine out the position of mechanical arm tail end mechanism, similarly, know the position of mechanical arm tail end mechanism
It sets, the joint angles in each joint of mechanical arm can be solved using inverse kinematics.Thus in present embodiment, at the beginning of mechanical arm
Beginning location information can refer to the initial coordinate information of the joint angles that each joint is initial in mechanical arm or each joint, also
It can refer to the coordinate information of the initial position of the tail house of mechanical arm；Similarly, the target position information of mechanical arm can be
The target joint angle in each joint of the mechanical arm or the coordinates of targets information in each joint, can also be the end of mechanical arm
The coordinate information of the target position of terminal structure.
Mechanical arm can know the initial position message of itself, and the target position information of robotic arm can be manually entered,
Target position information can also be determined by the camera and/or sensing device of robot.
As long as can use direct kinematics it should be noted that knowing the joint angles in each joint of mechanical arm and solve
The position of the mechanical arm tail end mechanism；Similarly, if knowing the position of mechanical arm tail end mechanism, it can use inverse kinematics solution
The joint angles in each joint of the mechanical arm out.
Step 102: by initial position message and target position information input motion trajectory predictions model, motion profile is pre-
Surveying model is the extreme learning machine model that training extremely restrains for predicting the motion profile of mechanical arm in advance.
In one concrete implementation, the motion profile prediction model is pre- advanced using the network structure of extreme learning machine model
Row training.Before training, the number of nodes and input weight of hidden layer in extreme learning machine model are determined at random.
Specifically, which is usually single hidden layer network structure, that is, includes input layer, single hiding
Layer and output layer, since the extreme learning machine model only has one layer of hidden layer, training speed is fast.It is random to determine extreme learning machine
Number of nodes and input weight in model in hidden layer.The principle of the extreme learning machine is described below:
For example, if input isGiven excitation function is g, then objective function is indicated as shown in formula (1):
Wherein, N indicates that the interstitial content of hidden layer, bi indicate the biasing of i-th of node of hidden layer, wiIndicate i-th it is hidden
Hide the input weight of node, βiIndicate the output weight of i-th of concealed nodes.
The target of the extreme learning machine model learning is the error minimum so that output, can be as shown in formula (2):
min||HβT- o | | formula (2)；
Wherein, o is the target value of output layer output, o=(o1,…oN)T, H is the output of hidden layer, and H (w, b) is indicated such as
Shown in formula (3),
Excitation function meets formula (4):
The training process of the motion profile prediction model can be with are as follows: by the corresponding mechanical arm of different periods in sample set
Initial position message, target position information and a plurality of target trajectory conduct that target position is moved to from initial position
Training set moves to the expectation target motion profile of target position as verifying from initial position for different periods are corresponding
Collection, expectation target motion profile are identical as the motion profile of human arm；By the data input limits learning machine mould in training set
In type, realistic objective motion profile is obtained；The error compared between realistic objective motion profile and expectation target motion profile is
It is no within a preset range, if it is not, then in iterative cycles adjustment extreme learning machine model hidden layer parameter, until error is pre-
Terminate when within the scope of if.
It specifically, can be taking human as the sample set acquired for training extreme learning machine, without carrying out machine in discrete space
The acquisition of tool arm motion profile directly can acquire the mechanical arm in different periods and move to respective correspondence from different initial positions
Target position to a target trajectory.For example, Xi=X1,X2…Xn,Ti=T1,T2…Tn, i=1,2 ... Ntra, wherein
NtraIndicate that the total number of initial position and corresponding target position, i indicate i-th initial position to corresponding target position
The sample of target trajectory, TiFor the matrix that multiple acquisition moment are formed, XnFor the position in each joint in a target trajectory
Confidence breath composition matrix, n be the initial position to corresponding target position target trajectory sample number.
In training process, multiple initial positions and corresponding target position are inputted in the extreme learning machine model, it is defeated
The realistic objective motion profile is compared with expectation target motion profile, judges the two by realistic objective motion profile out
Between error whether within preset range.Expectation target motion profile is identical as the motion profile of human arm, for example, mechanical
The initial position of arm tail house is point A, and corresponding target position is point B, if acquiring 20 targets fortune from point A to point B
There is the motion profile for not meeting human arm kinematic principle in 20 target trajectories in dynamic rail mark, or existing can be with barrier
Hinder the target trajectory under object collision situation, if from point A to point B, the motion profile phase of target trajectory 1 and human arm
Together, then it regard target trajectory 1 as expectation target motion profile.
It should be noted that in the training process, the condition of convergence of the motion profile prediction model includes: to meet Li Yapu
The default constraint condition of stability and satisfaction of love function, presets constraint condition are as follows:biIt indicates to hide
Layer biasing, βiIndicate the weight of output layer, wiIndicate the weight of input layer, 0 < i < N, N are the integer greater than 1, and N indicates hidden layer
Number of nodes.
Specifically, it since manipulator motion needs stability, in the training process, is verified by Li Yapu love function
Whether trained limit learning model output is stable, while needing to meet default constraint condition, and constraint condition is used for assistance-limit
Learning machine fast convergence.
Output weight is arranged to β=- α w to start to solve in the present embodiment, and wherein α is a positive parameter, this
Initial value meets constraint condition simultaneously, thus it is possible to vary α determines obstructed initial value.The location information in each joint can be in sample set
It is normalized, is normalized to one compared with minizone, while input weight is also initialized in a minizone, so that
The matrix H of hidden layer is nonsingular matrix, improves the study precision of motion profile prediction model.
Step 103: the mechanical arm for obtaining the output of motion profile prediction model moves to the fortune of target position from initial position
Dynamic rail mark, and control mechanical arm and moved according to motion profile.
Specifically, the motion profile of motion profile prediction model output can be the mechanical arm in each discrete time
Location information, the location information can be the joint angles in each joint of the mechanical arm, be also possible to each joint of the mechanical arm pre-
If the coordinate in coordinate system.If location information is the joint angles in each joint of the mechanical arm, when can be directly discrete according to this
Between upper mechanical arm location information adjust the mechanical arm each joint joint angles.
In one concrete implementation, if the location information is coordinate of each joint of the mechanical arm in preset coordinate system；Machine
By multiple joint changing path of movement, control mechanical arm moves tool arm according to motion profile, and specific process can be with are as follows: according to
Motion profile generates the multiple joint angles for having mapping relations with motion profile, wherein multiple joint angles and multiple joints
It corresponds, multiple joint angles are compounded to form motion profile；Corresponding joint is adjusted according to multiple joint angles, so that mechanical arm
It is moved along motion profile.
Specifically, reversely inverse solution can be carried out according to the location information of mechanical arm, can determine that the mechanical arm is each
The joint angles in joint, thus can determine according to motion profile the joint angle in each joint of the mechanical arm at each moment
Degree.Multiple joint angles are compounded to form motion profile, adjust corresponding joint by multiple joint angles, so that mechanical arm edge moves
Track movement.
Embodiment of the present invention in terms of existing technologies, passes through the initial position message and the machine of acquisition mechanical arm
The target position information of tool arm is obtained in the initial position message and target position information input motion trajectory predictions model
The motion profile of motion profile prediction model output controls the mechanical arm and moves according to the motion profile；Motion profile predicts mould
Type uses extreme learning machine model, since the network structure of extreme learning machine model is simple, for the network model of single hidden layer
Structure, so that it is fast using the training speed of the motion profile prediction model of extreme learning machine training acquisition, meanwhile, the movement rail
Mark prediction model can be adapted for different degree-of-freedom manipulators, improve the scope of application；The motion profile predicts mould after having trained
The network structure of type is simple, so that according to the input data, can quickly be exported as a result, making the movement rail to mechanical arm
Mark planning is very fast, improves the speed of manipulator motion to target position, substantially reduces the speed of control manipulator motion.
Second embodiment of the present invention is related to a kind of control method of manipulator motion.Machinery in second embodiment
The control method of arm movement includes: the initial position message for obtaining mechanical arm and the target position information of mechanical arm；It will be initial
In location information and target position information input motion trajectory predictions model, motion profile prediction model is that training extremely restrains in advance
For predicting the extreme learning machine model of the motion profile of mechanical arm；The mechanical arm of motion profile prediction model output is obtained from first
Beginning position moves to the motion profile of target position, and controls mechanical arm and move according to motion profile.
Second embodiment is mainly changed to the further improvement of training motion profile prediction model in first embodiment
It is into place: in second embodiment of the invention, compares between realistic objective motion profile and expectation target motion profile
Error whether within a preset range, be the practical mesh of First Speed vector sum by joint each in expectation target motion profile
The second speed vector in each joint in motion profile is marked, calculates error, and judge the error whether within a preset range.It compares
Error between realistic objective motion profile and expectation target motion profile whether a kind of specific implementation within a preset range
Schematic diagram is as shown in Figure 2.
Step 201: obtaining First Speed vector and the realistic objective movement in each joint in expectation target motion profile
The second speed vector in each joint in track.
Specifically, it can use the velocity vector that velocity vector formula calculates each joint, as shown in formula (5),
Wherein, x indicates that the pose of current joint, t indicate the current time.
According to the formula (5), the First Speed vector in each joint in expectation target motion profile can be determined, and
The second speed vector in each joint in target actual motion track.
Step 202: according to each First Speed vector and each second speed vector, calculating error, and judge the mistake
Within a preset range whether difference.
Due to the movement needs stability of mechanical arm, and the trend of the velocity variations in each joint of mechanical arm is to the steady of model
Fixed and precision is very sensitive, thus the accuracy of baseline learning machine model in the training process can use velocity vector into
Row judgement, error calculation formula can be as shown in formula (6):
Wherein, First Speed vectorWith second speed vectorR and q is two parameters, can be distinguished in present embodiment
It is set as 0.6 and 0.4, ε is the positive parameter an of very little, and N indicates sample number,For the error being calculated, judge that the error is
It is no within a preset range, preset range can be configured according to actual needs.
It is noted that respectively being corresponded to after acquiring each target trajectory by different periods in sample set
Mechanical arm initial position message and target position information as training set before, the data in sample set can be carried out pre-
Processing, to improve the precision of subsequent training.
Pretreatment may include: to be smoothed and/or denoising to every target trajectory of acquisition.It can be right
Every target trajectory is filtered trimming, carries out denoising using standardized residual.
For ease of calculation, coordinate system where joint each in every target trajectory can also be adjusted to preset coordinate
System.The step of unified coordinate system, can also be before smoothing processing and/or denoising.
The control method of the manipulator motion provided in present embodiment, due in the motion control to mechanical arm, to steady
Qualitative requirement is high, and the velocity variations in joint can reflect out the stability of model, carries out error calculation by velocity vector, can
Relatively accurately to reflect the stability of the model.
The step of various methods divide above, be intended merely to describe it is clear, when realization can be merged into a step or
Certain steps are split, multiple steps are decomposed into, as long as including identical logical relation, all in the protection scope of this patent
It is interior；To adding inessential modification in algorithm or in process or introducing inessential design, but its algorithm is not changed
Core design with process is all in the protection scope of the patent.
Third embodiment of the invention is related to a kind of terminal, and the specific structure is shown in FIG. 3 for the terminal, comprising: at least one
A processor 301；And the memory 302 with the communication connection of at least one processor 301；Wherein, memory 302 is stored with
The instruction that can be executed by least one processor 301, instruction are executed by least one processor 301, so that at least one is handled
Device 301 is able to carry out the control method of the manipulator motion in first embodiment or second embodiment.
Wherein, memory 302 is connected with processor 301 using bus mode, and bus may include any number of interconnection
Bus and bridge, bus the various circuits of one or more processors 301 and memory 302 are linked together.Bus may be used also
To link together various other circuits of such as peripheral equipment, voltage-stablizer and management circuit or the like, these are all
It is known in the art, therefore, it will not be further described herein.Bus interface provides between bus and transceiver
Interface.Transceiver can be an element, be also possible to multiple element, such as multiple receivers and transmitter, provide for
The unit communicated on transmission medium with various other devices.The data handled through processor 301 pass through antenna on the radio medium
It is transmitted, further, antenna also receives data and transfers data to processor 301.
Processor 301 is responsible for management bus and common processing, can also provide various functions, including timing, periphery connects
Mouthful, voltage adjusting, power management and other control functions.And memory 302 can be used for storage processor 301 and execute
Used data when operation.
Four embodiment of the invention is related to a kind of computer readable storage medium, is stored with computer program, computer
The control method of manipulator motion in first embodiment or second embodiment is realized when program is executed by processor.
It will be appreciated by those skilled in the art that implementing the method for the above embodiments is that can pass through
Program is completed to instruct relevant hardware, which is stored in a storage medium, including some instructions are used so that one
A equipment (can be single-chip microcontroller, chip etc.) or processor (processor) execute each embodiment the method for the application
All or part of the steps.And storage medium above-mentioned includes: USB flash disk, mobile hard disk, read-only memory (ROM, Read-Only
Memory), random access memory (RAM, Random Access Memory), magnetic or disk etc. are various can store journey
The medium of sequence code.
It will be understood by those skilled in the art that the respective embodiments described above are to realize specific embodiments of the present invention,
And in practical applications, can to it, various changes can be made in the form and details, without departing from the spirit and scope of the present invention.
Claims (10)
1. a kind of control method of manipulator motion characterized by comprising
Obtain the initial position message of mechanical arm and the target position information of the mechanical arm；
By in the initial position message and the target position information input motion trajectory predictions model, the motion profile is pre-
Surveying 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 motion profile prediction model output moves to the movement rail of target position from initial position
Mark, and control the mechanical arm and moved according to the motion profile.
2. the control method of manipulator motion according to claim 1, which is characterized in that the mechanical arm passes through multiple passes
Changing path of movement is saved, the control mechanical arm is moved according to the motion profile, comprising:
Multiple joint angles that there are mapping relations with the motion profile are generated according to the motion profile, wherein described more
A joint angles and the multiple joint correspond, and the multiple joint angles are compounded to form the motion profile；
Corresponding joint is adjusted according to the multiple joint angles, so that the mechanical arm is moved along the motion profile.
3. the control method of manipulator motion according to claim 1, which is characterized in that the motion profile prediction model
Before training, the number of nodes and input weight of hidden layer in extreme learning machine model are determined at random.
4. the control method of manipulator motion according to any one of claim 1 to 3, which is characterized in that the movement
The training process of trajectory predictions model, specifically includes:
By the initial position message of the corresponding mechanical arm of different periods in sample set, target position information and from first
Beginning position moves to a plurality of target trajectory of target position as training set, and the different periods are corresponding from first
Beginning position moves to the expectation target motion profile of target position as verifying collection, the expectation target motion profile and mankind's hand
The motion profile of arm is identical；
By in the data input limits learning machine model in training set, realistic objective motion profile is obtained；
Whether within a preset range to compare error between the realistic objective motion profile and the expectation target motion profile,
If it is not, then iterative cycles adjust the parameter of hidden layer in the extreme learning machine model, until the error is in preset range
Within when terminate.
5. the control method of manipulator motion according to any one of claim 1 to 4, which is characterized in that the movement
The condition of convergence in the training of trajectory predictions model includes:
Meet the stability of Li Yapu love function and meet default constraint condition, the default constraint condition are as follows:biIndicate the biasing of i-th of node of hidden layer, βiIndicate the output weight of i-th of concealed nodes, wiIt indicates
The input weight of i-th of concealed nodes, 0 < i < N, N are the integer greater than 1, and N indicates the number of nodes of hidden layer.
6. the control method of manipulator motion according to claim 4, which is characterized in that described to compare the realistic objective
Within a preset range whether the error between motion profile and expectation target motion profile, specifically include:
It obtains every in the First Speed vector and the realistic objective motion profile in each joint in expectation target motion profile
The second speed vector in a joint；
According to each First Speed vector and each second speed vector, the error is calculated, and whether judge the error
Within a preset range.
7. the control method of manipulator motion according to claim 4, which is characterized in that it is described will in sample set it is different
Before the initial position message and target position information of period corresponding mechanical arm are as training set, the mechanical arm
The control method of movement further include:
Every target trajectory of acquisition is smoothed and/or denoising.
8. the control method of manipulator motion according to claim 6 or 7, which is characterized in that it is described will be in sample set
Before the initial position message and target position information of the corresponding mechanical arm of different periods are as training set, the machine
The control method of tool arm movement further include:
Coordinate system where joint each in every target trajectory is adjusted to preset coordinate system.
9. a kind of terminal characterized by comprising
At least one processor；And
The memory being connect at least one described processor communication；Wherein,
The memory is stored with the instruction that can be executed by least one described processor, and described instruction is by described at least one
It manages device to execute, so that at least one described processor is able to carry out the control of manipulator motion a method as claimed in any one of claims 1-8
Method processed.
10. a kind of computer readable storage medium, is stored with computer program, which is characterized in that the computer program is located
Reason device realizes the control method of manipulator motion described in any item of the claim 1 to 8 when executing.
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 (13)
| 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 | Robot tail end structure and control method thereof |
| 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 |
| 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 | Working method, device, equipment and medium of floor paint coating equipment |
| 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 |
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 |
| 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 |
| 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, storage medium and the terminal device of control machinery arm 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, apparatus, equipment and storage medium |
| CN109960880A (en) * | 2019-03-26 | 2019-07-02 | 上海交通大学 | A kind of industrial robot obstacle-avoiding route planning method 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, storage medium and the terminal device of control machinery arm motion |
| 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, apparatus, equipment and storage medium |
| CN109960880A (en) * | 2019-03-26 | 2019-07-02 | 上海交通大学 | A kind of industrial robot obstacle-avoiding route planning method 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 (18)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111037589A (en) * | 2019-12-11 | 2020-04-21 | 西安工程大学 | Robot tail end structure and control method thereof |
| 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 |
| 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 | Working method, device, equipment and medium of floor paint coating equipment |
| 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 |
| 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 |
| 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 |
| 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 |
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 | |
| CN109960880B (en) | Industrial robot obstacle avoidance path planning method based on machine learning | |
| CN107990899A (en) | A kind of localization method and system based on SLAM | |
| US10347001B2 (en) | Localizing and mapping platform | |
| Zhao et al. | A path planning method based on multi-objective cauchy mutation cat swarm optimization algorithm for navigation system of intelligent patrol car | |
| CN108638062A (en) | Robot localization method, apparatus, positioning device and storage medium | |
| CN112669345B (en) | Cloud deployment-oriented multi-target track tracking method and system | |
| CN110174112B (en) | Path optimization method for automatic mapping task of mobile robot | |
| US20220277523A1 (en) | Vr scene and interaction method thereof, and terminal device | |
| CN113219506A (en) | Positioning method for multimode fusion seamless switching | |
| CN115860107B (en) | Multi-machine searching method and system based on multi-agent deep reinforcement learning | |
| CN112344945A (en) | Indoor distribution robot path planning method and system and indoor distribution robot | |
| CN109655058A (en) | A kind of inertia/Visual intelligent Combinated navigation method | |
| WO2024037299A1 (en) | Localization method and apparatus, and robot and storage medium | |
| KR20240052808A (en) | Multi-robot coordination using graph neural networks | |
| Foix et al. | 3D Sensor planning framework for leaf probing | |
| CN112580582A (en) | Action learning method, action learning device, action learning medium and electronic equipment | |
| CN113296504A (en) | Mobile robot mapping and path planning method based on RGBD depth camera | |
| CN112070835B (en) | Mechanical arm pose prediction method and device, storage medium and electronic equipment | |
| CN108712725A (en) | A kind of SLAM methods based on rodent models Yu WIFI fingerprints | |
| CN108717302B (en) | Method and device for robot to follow person, storage medium and robot | |
| CN115862074A (en) | Human body direction determining method, human body direction determining device, screen control method, human body direction determining device and related equipment | |
| CN112991527B (en) | Target object avoiding method and device, storage medium and electronic device | |
| CN116009542A (en) | Dynamic multi-agent coverage path planning method, device, equipment and storage medium | |
| CN115446867A (en) | Industrial mechanical arm control method and system based on digital twinning technology |
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 |