CN109382828A - A kind of Robot Peg-in-Hole assembly system and method based on learning from instruction - Google Patents
A kind of Robot Peg-in-Hole assembly system and method based on learning from instruction Download PDFInfo
- Publication number
- CN109382828A CN109382828A CN201811275792.8A CN201811275792A CN109382828A CN 109382828 A CN109382828 A CN 109382828A CN 201811275792 A CN201811275792 A CN 201811275792A CN 109382828 A CN109382828 A CN 109382828A
- Authority
- CN
- China
- Prior art keywords
- mechanical arm
- hole
- peg
- torque sensor
- torque
- 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.)
- Granted
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
- B25J9/0081—Programme-controlled manipulators with master teach-in means
-
- B—PERFORMING OPERATIONS; TRANSPORTING
- B23—MACHINE TOOLS; METAL-WORKING NOT OTHERWISE PROVIDED FOR
- B23P—METAL-WORKING NOT OTHERWISE PROVIDED FOR; COMBINED OPERATIONS; UNIVERSAL MACHINE TOOLS
- B23P19/00—Machines for simply fitting together or separating metal parts or objects, or metal and non-metal parts, whether or not involving some deformation; Tools or devices therefor so far as not provided for in other classes
-
- B—PERFORMING OPERATIONS; TRANSPORTING
- B25—HAND TOOLS; PORTABLE POWER-DRIVEN TOOLS; MANIPULATORS
- B25J—MANIPULATORS; CHAMBERS PROVIDED WITH MANIPULATION DEVICES
- B25J19/00—Accessories fitted to manipulators, e.g. for monitoring, for viewing; Safety devices combined with or specially adapted for use in connection with manipulators
Landscapes
- Engineering & Computer Science (AREA)
- Mechanical Engineering (AREA)
- Robotics (AREA)
- Manipulator (AREA)
Abstract
The invention discloses a kind of Robot Peg-in-Hole assembly system and method based on learning from instruction, system includes mechanical arm, six-dimensional force/torque sensor, passive flexibility RCC device and PC host computer, mechanical arm is Multi-shaft mechanical arm, six-dimensional force/torque sensor is mounted on mechanical arm tail end, passive flexibility RCC device is mounted on six-dimensional force/torque sensor, clamping device for clamping component to be assembled is installed, PC host computer and mechanical arm and six-dimensional force/torque sensor can carry out real time communication on passive flexibility RCC device.Teaching recorder artificial first completes fittage data, matches skill models using learning algorithm training cartridge, and then mechanical arm carries pin part and carry out peg-in-hole assembly, the control system that PC host computer is built is based on ROS platform under the control instruction of PC host computer.Present invention combination learning from instruction imitates flexible behavior when people completes fittage, realizes robot autonomous flexible assembly operation, can meet job requirements well.
Description
Technical field
The invention belongs to intelligence manufacture fields, be related to a kind of teaching robot, and in particular to a kind of based on learning from instruction
Robot Peg-in-Hole assembly system and method.
Background technique
Manufacturing technology is the core of economic competition, and manufacture atomization degree is higher and higher.Industrial robot is used extensively
In manufacturing field, for improving production efficiency and product quality.Industrial robot is mainly used for carrying, painting etc. not by about at present
The movement of Shu Zuoye, robot end's tool are unrestricted.For controlled task, such as times that assembly is contacted with workpiece
It is engaged in, has certain tolerance fit required precision in general precision assembly task, fit-up gap is small, is easy to that assembly is caused to block
Extremely.Industrial robot contact stiffness based on position or speed control is big, and larger contact force is also easy to produce in contact process, causes work
Part or injury are unable to complete precision assembly task substantially.When traditional industrial robot completes constrained task, generally adopt
Become with taught point or off-line programing, there are many deficiencies, and if deployment time is long, algorithm, programming are complicated, require operator
It is higher, be only used for structured environment, to adaptive capacity to environment difference etc..Therefore fittage is and artificial still based on craft at present
Operating efficiency is low, at high cost, and manufacturing environment seriously affects worker's health, and product homogeneity is poor, and defect rate is high.Therefore, needle
Contact force real-time measurement and feedback control need to be introduced to realize Automated assembly to controlled assembling work, assembly is reduced and connect
Touch, and estimate that assembled state, adjustment assembly Motion realize flexible assembly operation based on contact force.
Summary of the invention
In view of above-mentioned deficiencies of the prior art, the purpose of the present invention is to provide a kind of putting together machines based on learning from instruction
The assembly learning platform based on force feedback has been built by people and its control system, the robot, extracts people in conjunction with learning from instruction and completes
Flexible behavior when fittage is realized robot autonomous flexible assembly operation, is known without complicated program and control algorithm design
Know, allows robot to imitate people and independently complete fittage.
In order to solve the above-mentioned technical problem, the technical solution adopted by the present invention is that:
A kind of Robot Peg-in-Hole assembly system based on learning from instruction, it is characterised in that: including mechanical arm, six-dimensional force/power
Square sensor, passive flexibility RCC (Remote Center Compliance) device and PC host computer, the six-dimensional force/power
Square sensor is used to carry out the measurement of contact force, and the mechanical arm is the Multi-shaft mechanical arm with motion control function, and six-dimensional force/
Torque sensor is mounted on mechanical arm tail end, and passive flexibility RCC device is mounted on six-dimensional force/torque sensor, PC host computer
Be connected with mechanical arm and six-dimensional force/torque sensor communication, in operational process, mechanical arm under the control instruction of PC host computer,
It carries passive flexibility RCC device clamping pin part and carries out peg-in-hole assembly, the control system that the PC host computer is built is based on ROS
(Robot operating system) platform.
As an improvement, six-dimensional force/the torque sensor in force feedback real-time monitoring passively flexibility RCC device by clamping
Stress between workpiece and axis hole adjusts workpiece posture by mechanical arm in time.
A kind of Robot Peg-in-Hole assembly method based on learning from instruction, the equipment packet which uses
Include mechanical arm, six-dimensional force/torque sensor, passive flexibility RCC device and PC host computer, which is characterized in that including following step
It is rapid:
Step 1: artificial teaching, mechanical arm drives pin part mobile toward shaft hole part, and control pin and hole plane contact are simultaneously protected
Constant contact force is held, mechanical arm drives pin part mobile search axis hole in the plane of hole, when six-dimensional force/torque sensor detection
Contact force mutation, that is, navigate to axis hole, and artificial traction teaching later carries out pin part and peg-in-hole assembly, record cooperation assembling process
In torque and angular speed value, according to the method described above carry out M time repeat teaching, and acquire torque and angular speed formation data
Collection;
Step 2: model learning, will be repeated several times the data set that teaching generates and shows by using gauss hybrid models coding
Data are taught, and obtain the relational model of power and angular speed, gauss hybrid models training is carried out based on expectation maximization (EM) algorithm,
The Function Mapping relationship of torque and angular speed can be obtained;
Step 3: peg-in-hole assembly, considers that there is deviation in the direction of axis hole and attitude angle, mechanical arm drives pin to axle hole
Part surface location is moved, and then carries out axis hole search positioning on shaft hole part surface, after positioning shaft hole, by six-dimensional force/
Torque sensor feedback control carries out pose adjustment, according to the Function Mapping relationship of torque in model learning and angular speed, according to
The output position coordinate of the magnitude of angular velocity of the Calculating Torque during Rotary mechanical arm tail end acquired in real time, binding site controller passes through inverse movement
It learns and calculates six shaft mechanical shoulder joint angles, to control mechanical arm tail end movement, complete peg-in-hole assembly operation.
As an improvement, mechanical arm drives pin part during mobile search axis hole in the plane of hole in step 1, sell in hole
Plane carries out the search of axis hole position with spiral of Archimedes motion profile, when six-dimensional force/torque sensor detects the direction z
Power jump when, show that axis hole position has been found, stopping search hole.
As an improvement, the mechanical arm has joint moment estimation function, may be implemented to be driven in the reverse direction, according to joint moment
It can estimate external environment power, and be mapped in robot motion, realize that people's driving machinery arm completes free movement.
As an improvement, being directed to data set ξ=[F, the X] of teachingT, wherein F={ Mx,My, X={ ωx,ωy, using Gauss
Mixed model encodes training data, and obtains the relational model of power and position, wherein some data point ξ ∈ RD×NProbability are as follows:
Wherein πk∈ [0,1] is prior probability, andK is the number of Gaussian Profile, and R is real number field, and N is number
The sum at strong point, D are the dimension of data, μk∈RD,∑k∈RD×DRespectively indicate the mean value and covariance square of k-th of Gaussian Profile
Battle array gives input variable ξX, export ξFConditional probability distribution are as follows:
Wherein Indicate the mean value and variance of k-th of Gaussian Profile in posterior probability:
AndK-th of Gaussian Profile mean value of respectively X and F,Covariance matrix between X and F:
ξFProbability in k-th of Gaussian Profile is
Gaussian Mixture returns as given ξFLower ξXConditional probability distribution expectation
Therefore, gauss hybrid models/Gaussian Mixture is returned by parameterIt is determined, the determination one of parameter value
As use expectation-maximization algorithm, hyper parameter K be Gaussian Profile number, determined, obtained optimal by bayesian information criterion
Model parameter, final torque and angular speedBetween mapping relations, when can be completed fittage
Adjustable strategies.
As an improvement, when carrying out axis hole search, being controlled using admittance controller as position in the step 1 and step 3
Device processed realizes power control tracking, while keeping constant with hole plane vertical direction contact force, reduces the contact force in the direction x and y,
Realize flexible contact.
As an improvement, the mechanical arm carries out the fine tuning of posture by torque controller, torque controller is using above-mentioned
The torque learnt and angular speed between mapping relations:According to real-time collected torque,
Target end magnitude of angular velocity is calculated, end spin matrix R (t) is by R (t+ Δ t)=(Δ tS (ωa)+I3×3) R (t) calculating, ωx,
ωy,ωzIt is mechanical arm tail end axis in x, y, the angular speed in the direction z, t indicates the time, and Δ t indicates the control period of control system,
I3×3Indicate three rank unit matrixs, wherein backslash symmetrical matrix:
After calculating spin matrix according to the control period, the position output coordinate x of binding site controller passes through inverse movement
It learns and calculates six shaft mechanical shoulder joint angles, to control robot motion, realize the flexible axis hole operation that personalizes.
As an improvement, the range of the teaching number M is 3-9 times.
The beneficial effects of the present invention are:
1, learning from instruction is applied in assembly, it is only necessary to which artificial teaching can carry out peg-in-hole assembly to workpiece, avoid
Complicated staking-out work, it is low using threshold;
2, deployment quickly, does not need high-precision installation calibrating, improves the efficiency of assembly;
3, algorithm versatility is good, is based on ROS system, transplants convenient for algorithm, is applicable to different robot and sensor
System building.
Detailed description of the invention
Fig. 1 is Robot Peg-in-Hole assembly system schematic diagram of the present invention;
Fig. 2 is Robot Peg-in-Hole assembly system block diagram of the present invention;
Fig. 3 is learning from instruction frame diagram;
Fig. 4 is torque of the present invention and angular speed relation schematic diagram;
Fig. 5 is that Robot Peg-in-Hole assembly system assembly of the present invention executes control program schematic diagram.
Wherein, 1- mechanical arm, 2- six-dimensional force/torque sensor, the passive flexibility RCC device of 3-, 4-PC host computer.
Specific embodiment
(2) as shown in Fig. 2, control system block diagram includes mechanical arm 1, six-dimensional force/torque sensor 2, PC host computer 4, institute
The measurement that six-dimensional force/torque sensor 2 is used to carry out contact force is stated, the mechanical arm 1 is the multiaxis with motion control function
Mechanical arm 1, six-dimensional force/torque sensor 2 are mounted on 1 end of mechanical arm, and passive flexibility RCC device 3 is mounted on six-dimensional force/torque
On sensor 2, PC host computer 4 is connected with mechanical arm 1 and the communication of six-dimensional force/torque sensor 2, and in operational process, mechanical arm 1 exists
Under the control instruction of PC host computer 4, carries passive 3 clamping pin part of flexibility RCC device and carry out peg-in-hole assembly, the PC host computer
4 control systems built are based on ROS platform.The control system is based on ROS platform, including control algolithm node, based on 6 DOF
The feedback node and mechanical arm 1 of power/torque sensor 2 move real-time control node.Six-dimensional force/torque sensor 2 and mechanical arm
1 is connected to same local area network by cable and PC.Modbus communication protocol is used between mechanical arm 1 and PC, using unified in ROS
Communication interface realize 1 state of mechanical arm read and motion control.Six-dimensional force/torque sensor 2 uses Ethernet agreement will
Data are sent to the end PC, and each sensor states are received in control algolithm node, and calculate control instruction of lower period.
The mechanical arm 1 is UR3 mechanical arm, has driver and function packet based on ROS, is installed and base can be realized
This motion control and encoder data are read, to complete building for control system.Six-dimensional force/the torque sensor 2 has
Driver based on ROS, Real-time Feedback contact force data.In the present embodiment, select UR3 robotic arm flat as robot
Platform, six-dimensional force/torque sensor 2 select HEX-70-XE-200N six-dimensional force/torque sensor of Hungarian OPTOFORCE.
The control algolithm be designed based on the active force control algorithm of power/position mixing control using Simulink, and by its
It is converted into C++ code, which is based on using six-dimensional force/torque sensor 2 and 1 joint encoders data of mechanical arm as input quantity
Six-dimensional force/torque sensor 2 is fed back, and is realized the control of 1 position and speed of mechanical arm, is completed the operation of assembly.
(3) as shown in figure 3, the learning from instruction is by acquiring data when people's execution task, and machine learning is combined
Method carries out data processing and model training, obtains technical ability when people completes certain task, the technical ability is finally given to machine
People realizes the task execution of robot humanoid.Learning from instruction can be divided into three parts, comprising artificial teaching, model learning,
It is autonomous to execute.
The artificial teaching process is divided into the following steps, and: A sells to hole component and is moved, the direction of setting hole and appearance
State angular deviation, with hole plane contact and the contact force that keeps constant;B, in moving process, axis is in hole plane with Archimedes
Spiral motion track carries out the search of hole site, when six-dimensional force/torque sensor 2 detects the jump of the power in the direction z, table
Bright hole site has been found, and hole is searched in stopping;C presses the I/O button of end effector, allows mechanical arm 1 to execute and freely drives mould
Formula realizes that people's bolt that cooperates with mechanical arm 1 enters hole.And record the torque M={ M in cooperation assembling processx,MyAnd angular velocity omega
={ ωx,ωyValue, repeatedly, Fig. 4 (a) show the torque of single teaching and the phasor of angular speed.It is used in this example
Mechanical arm 1 has joint moment estimation function, may be implemented to be driven in the reverse direction, can estimate external environment power according to joint moment,
And it is mapped in robot motion, it can be achieved that people's driving machinery arm completes free movement.
For data set ξ=[F, the X] of teachingT, wherein F={ Mx,My, X={ ωx,ωy, using gauss hybrid models
(GMM) training data is encoded, and obtains the relational model of power and position.Wherein some data point ξ ∈ RD×NProbability are as follows:
Wherein πk∈ [0,1] is prior probability, andK is the number of Gaussian Profile, and R is real number field, and N is number
The sum at strong point, D are the dimension of data.μk∈RD,∑k∈RD×DRespectively indicate the mean value and covariance square of k-th of Gaussian Profile
Battle array.Given input variable ξX, export ξFConditional probability distribution are as follows:
Wherein Indicate the mean value and variance of k-th of Gaussian Profile in posterior probability:
AndK-th of Gaussian Profile mean value of respectively X and F,For covariance matrix:
ξFProbability in k-th of Gaussian Profile is
It is given ξ that Gaussian Mixture, which returns (GMR),FLower ξXConditional probability distribution expectation
Therefore, GMM/GMR is by parameterIt is determined, the determination of parameter value generally uses expectation maximization to calculate
Method (Expectation Maximization, EM).Hyper parameter K is the number of Gaussian Profile, by bayesian information criterion
(Bayesian Information Criterion, BIC) is determined.Optimal model parameter is obtained, torque and angle are finally obtained
SpeedBetween mapping relations, adjustable strategies when fittage can be completed, as shown in Fig. 4 (b)
For the torque and angular speed distribution relation of x-axis direction.
It is that the bolt based on force feedback enters hole control program as shown in Figure 5.When executing fittage, control shaft is needed
Position and posture.Force controller is adjusted for shaft position, and torque controller is used for axis pose adjustment.It is automatic to execute assembly times
The whole process of business includes following four step:
1. pin is moved to hole component surface location;
2. parts surface scans for hole site in hole;
3. being based on control program shown in fig. 5, power control tracking is realized as positioner using admittance controller, is realized
Flexible contact.Specific controller architecture are as follows:
Wherein, Md,Bd,KdFor virtual mass, damping, stiffness parameters, x is the target position three-dimensional coordinate of controller output,
Fd=[0,0, -20N]TFor the reference contact force of setting, Fa=[Fax,Fay,Faz]TFor the contact force feedback in the actual direction xyz
Value.While the admittance controller being arranged can realize the tracking direction z 20N contact force herein, reduce the contact force in the direction x and y extremely
0。
4. being based on control program shown in fig. 5, the fine tuning of posture is carried out by torque controller.Torque controller uses
Mapping relations between the torque above-mentioned learnt and angular speed:According to real-time collected power
Square calculates target end magnitude of angular velocity, herein z directional angular velocity wzIt is set as 0.End spin matrix is by R (t+ Δ t)=(Δ
tS(ωa)+I3×3) R (t) calculates, wherein backslash symmetrical matrix:
It is Δ t=0.008s that the period is controlled in this example.After calculating spin matrix, the position of binding site controller is exported
Coordinate x, to control robot motion, realizes the flexible shaft that personalizes by six shaft mechanical arm of the computation of inverse- kinematics, 1 joint angles
Hole assembling work.
Claims (9)
1. a kind of Robot Peg-in-Hole assembly system based on learning from instruction, it is characterised in that: including mechanical arm, six-dimensional force/torque
Sensor, passive flexibility RCC device and PC host computer, the six-dimensional force/torque sensor are used to carry out the measurement of contact force,
The mechanical arm is the Multi-shaft mechanical arm with motion control function, and six-dimensional force/torque sensor is mounted on mechanical arm tail end, quilt
Dynamic flexibility RCC device is mounted on six-dimensional force/torque sensor, is equipped on passive flexibility RCC device to be assembled for clamping
The clamping device of component, PC host computer and mechanical arm and six-dimensional force/torque sensor can carry out real time communication, in operational process,
Mechanical arm carries passive flexibility RCC device clamping pin part and carries out peg-in-hole assembly, the PC under the control instruction of PC host computer
The control system that host computer is built is based on ROS platform.
2. Robot Peg-in-Hole assembly system as described in claim 1, it is characterised in that: the six-dimensional force/torque sensor is logical
Stress of the force feedback real-time monitoring passively on flexibility RCC device between clamping workpiece and axis hole is crossed, control mechanical arm is passed through
Adjustment workpiece posture in real time.
3. a kind of Robot Peg-in-Hole assembly method based on learning from instruction, the equipment which uses include
Mechanical arm, six-dimensional force/torque sensor, passive flexibility RCC device and PC host computer, which comprises the following steps:
Step 1: artificial teaching, mechanical arm drives pin part toward shaft hole part movement, control pin and hole plane contact and keeps permanent
Fixed contact force, mechanical arm drive pin part mobile search axis hole in the plane of hole, when six-dimensional force/torque sensor detection contact
Power mutation, that is, navigate to axis hole, and artificial traction teaching later carries out pin part and peg-in-hole assembly, and record cooperates in assembling process
The value of torque and angular speed, carries out repeating teaching M times according to the method described above, and acquires torque and angular speed formation data set;
Step 2: model learning, will be repeated several times the data set that teaching generates and encodes teaching number by using gauss hybrid models
According to, and the relational model of power and angular speed is obtained, gauss hybrid models training is carried out based on expectation-maximization algorithm, power can be obtained
The Function Mapping relationship of square and angular speed;
Step 3: peg-in-hole assembly, considers that there is deviation in the direction of axis hole and attitude angle, mechanical arm drives pin to shaft hole part table
Face position is moved, and is then carried out axis hole search on shaft hole part surface and is positioned, after positioning shaft hole, passes through six-dimensional force/torque
Sensor feedback control carries out pose adjustment, according to the Function Mapping relationship of torque in model learning and angular speed, according to real-time
The output position coordinate of the magnitude of angular velocity of the Calculating Torque during Rotary mechanical arm tail end of acquisition, binding site controller passes through inverse kinematics meter
Six shaft mechanical shoulder joint angles are calculated, to control mechanical arm tail end movement, complete peg-in-hole assembly operation.
4. Robot Peg-in-Hole assembly method as claimed in claim 3, it is characterised in that: in step 1, mechanical arm drives pin zero
For part during mobile search axis hole in the plane of hole, pin carries out axis hole position in hole plane with spiral of Archimedes motion profile
Search show that axis hole position has been found when six-dimensional force/torque sensor detects the jump of the power in the direction z, stop searching
Hole.
5. Robot Peg-in-Hole assembly method as claimed in claim 3, it is characterised in that: the mechanical arm is estimated with joint moment
Function is counted, may be implemented to be driven in the reverse direction, external environment power can be estimated according to joint moment, and be mapped in robot motion,
People's driving machinery arm can be achieved and complete free movement.
6. Robot Peg-in-Hole assembly method as claimed in claim 3, it is characterised in that: for data set ξ=[F, the X] of teachingT,
Wherein F={ Mx,My, X={ ωx,ωy, training data is encoded using gauss hybrid models, and obtain the relationship of power and position
Model, wherein some data point ξ ∈ RD×NProbability are as follows:
Wherein πk∈ [0,1] is prior probability, andK is the number of Gaussian Profile, and R is real number field, and N is data point
Sum, D be data dimension, μk∈RD,∑k∈RD×DThe mean value and covariance matrix for respectively indicating k-th of Gaussian Profile, give
Determine input variable ξX, export ξFConditional probability distribution are as follows:
Wherein Indicate the mean value and variance of k-th of Gaussian Profile in posterior probability:
AndK-th of Gaussian Profile mean value of respectively X and F,Covariance matrix between X and F:
ξFProbability in k-th of Gaussian Profile is
Gaussian Mixture returns as given ξFLower ξXConditional probability distribution expectation
Therefore, gauss hybrid models/Gaussian Mixture is returned by parameterIt is determined, the determination of parameter value is generally adopted
With expectation-maximization algorithm, hyper parameter K is the number of Gaussian Profile, is determined by bayesian information criterion, obtains optimal model
Parameter, final torque and angular speedBetween mapping relations, adjustment when fittage can be completed
Strategy.
7. Robot Peg-in-Hole assembly method as claimed in claim 3, it is characterised in that: in the step 1 and step 3, into
When row axis hole is searched for, power control tracking is realized as positioner using admittance controller, is being kept and hole plane vertical direction
While contact force is constant, reduce the contact force in the direction x and y, realizes flexible contact.
8. Robot Peg-in-Hole assembly method as claimed in claim 3, it is characterised in that: the mechanical arm passes through torque controller
Carry out the fine tuning of posture, torque controller is using the mapping relations between the above-mentioned torque learnt and angular speed:According to real-time collected torque, target end magnitude of angular velocity, end spin matrix R (t) are calculated
By R (t+ Δ t)=(Δ tS (ωa)+I3×3) R (t) calculating, ωx,ωy,ωzIt is mechanical arm tail end axis in x, y, the angle speed in the direction z
Degree, t indicate the time, and Δ t indicates the control period of control system, I3×3Indicate three rank unit matrixs, wherein backslash symmetrical matrix:
After calculating spin matrix according to the control period, the position output coordinate x of binding site controller passes through inverse kinematics meter
Six shaft mechanical shoulder joint angles are calculated, to control robot motion, realize the flexible axis hole operation that personalizes.
9. Robot Peg-in-Hole assembly method as claimed in claim 3, it is characterised in that: the range of the teaching number M is 3-9
It is secondary.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201811275792.8A CN109382828B (en) | 2018-10-30 | 2018-10-30 | Robot shaft hole assembling system and method based on teaching learning |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201811275792.8A CN109382828B (en) | 2018-10-30 | 2018-10-30 | Robot shaft hole assembling system and method based on teaching learning |
Publications (2)
Publication Number | Publication Date |
---|---|
CN109382828A true CN109382828A (en) | 2019-02-26 |
CN109382828B CN109382828B (en) | 2021-04-16 |
Family
ID=65427237
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201811275792.8A Active CN109382828B (en) | 2018-10-30 | 2018-10-30 | Robot shaft hole assembling system and method based on teaching learning |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN109382828B (en) |
Cited By (32)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN109732610A (en) * | 2019-03-01 | 2019-05-10 | 北京航空航天大学 | Man-machine collaboration robot grasping system and its working method |
CN110355557A (en) * | 2019-07-05 | 2019-10-22 | 清华大学 | A kind of spiral insertion method of large-size axis parts hole workpiece assembly |
CN110449882A (en) * | 2019-08-02 | 2019-11-15 | 珞石(北京)科技有限公司 | The search assembly method of binding force control |
CN110450143A (en) * | 2019-08-02 | 2019-11-15 | 珞石(北京)科技有限公司 | workpiece fatigue testing method based on cooperative robot |
CN110497423A (en) * | 2019-08-22 | 2019-11-26 | 泉州装备制造研究所 | A kind of manipulator adaptive machining method |
CN110501421A (en) * | 2019-07-24 | 2019-11-26 | 武汉大学 | A kind of track profiling method of detection based on mechanical arm |
CN110561421A (en) * | 2019-08-09 | 2019-12-13 | 哈尔滨工业大学(深圳) | Mechanical arm indirect dragging demonstration method and device |
CN110561430A (en) * | 2019-08-30 | 2019-12-13 | 哈尔滨工业大学(深圳) | robot assembly track optimization method and device for offline example learning |
CN110625611A (en) * | 2019-08-27 | 2019-12-31 | 上海卫星装备研究所 | Mechanical arm auxiliary component assembling method and system based on laser tracking measurement and force sensing combined control |
CN111230873A (en) * | 2020-01-31 | 2020-06-05 | 武汉大学 | Teaching learning-based collaborative handling control system and method |
CN111251277A (en) * | 2020-01-31 | 2020-06-09 | 武汉大学 | Human-computer collaboration tool submission system and method based on teaching learning |
CN111424380A (en) * | 2020-03-31 | 2020-07-17 | 山东大学 | Robot sewing system and method based on skill learning and generalization |
CN111452039A (en) * | 2020-03-16 | 2020-07-28 | 华中科技大学 | Robot posture adjusting method and device under dynamic system, electronic equipment and medium |
CN111515928A (en) * | 2020-04-15 | 2020-08-11 | 上海工程技术大学 | Mechanical arm motion control system |
CN111546035A (en) * | 2020-04-07 | 2020-08-18 | 大连理工大学 | Online rapid gear assembly method based on learning and prediction |
CN111993422A (en) * | 2020-08-11 | 2020-11-27 | 上海交通大学 | Robot axis and hole alignment control method based on uncalibrated vision |
CN112060627A (en) * | 2020-09-08 | 2020-12-11 | 武汉大学 | Digital intelligent laying method and system for composite material |
CN112223303A (en) * | 2020-08-27 | 2021-01-15 | 西安交通大学 | Robot automatic shaft hole axis alignment method and system based on plane surface friction |
CN112264998A (en) * | 2020-10-28 | 2021-01-26 | 上海非夕机器人科技有限公司 | Method for assembling operation member and adapting member by robot, robot and controller |
CN112428263A (en) * | 2020-10-16 | 2021-03-02 | 北京理工大学 | Mechanical arm control method and device and cluster model training method |
CN112605973A (en) * | 2020-11-19 | 2021-04-06 | 广东省科学院智能制造研究所 | Robot motor skill learning method and system |
CN112631201A (en) * | 2020-12-28 | 2021-04-09 | 佛山科学技术学院 | Hole searching control method and system for shaft hole assembly |
CN112665476A (en) * | 2020-12-29 | 2021-04-16 | 东风模具冲压技术有限公司 | Precision detection device for gripper of welding robot |
CN112706161A (en) * | 2020-11-17 | 2021-04-27 | 中国航空工业集团公司北京长城航空测控技术研究所 | Gluing control system with intelligent sensing capability |
CN112720478A (en) * | 2020-12-22 | 2021-04-30 | 深圳市优必选科技股份有限公司 | Robot torque control method and device, readable storage medium and robot |
CN112917474A (en) * | 2021-01-19 | 2021-06-08 | 哈尔滨工业大学 | Skill extraction platform and method for arm-hand operation hexagonal wrench tool |
CN114161479A (en) * | 2021-12-24 | 2022-03-11 | 上海机器人产业技术研究院有限公司 | Robot dragging demonstration performance test system and test method |
CN114786888A (en) * | 2020-01-16 | 2022-07-22 | 欧姆龙株式会社 | Control device, control method, and control program |
CN115338610A (en) * | 2022-07-04 | 2022-11-15 | 中国科学院自动化研究所 | Biaxial hole assembling method and device, electronic device and storage medium |
CN115685872A (en) * | 2022-09-05 | 2023-02-03 | 大连交通大学 | Robot assembly algorithm based on compliance control |
US11833666B2 (en) | 2020-10-28 | 2023-12-05 | Shanghai Flexiv Robotics Technology Co., Ltd. | Method for assembling an operating member and an adapting member by a robot, robot, and controller |
WO2024098787A1 (en) * | 2022-11-07 | 2024-05-16 | 节卡机器人股份有限公司 | Shaft hole assembly method and system, and storage medium |
Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN104057290A (en) * | 2014-06-24 | 2014-09-24 | 中国科学院自动化研究所 | Method and system for assembling robot based on visual sense and force feedback control |
CN104625676A (en) * | 2013-11-14 | 2015-05-20 | 沈阳新松机器人自动化股份有限公司 | Shaft hole assembly industrial robot system and working method thereof |
WO2016066616A1 (en) * | 2014-10-27 | 2016-05-06 | Kuka Systems Gmbh | Method and robot system for using an industrial robot for test jobs |
CN106799728A (en) * | 2017-03-16 | 2017-06-06 | 天津工业大学 | A kind of passive compliance device |
CN106826822A (en) * | 2017-01-25 | 2017-06-13 | 南京阿凡达机器人科技有限公司 | A kind of vision positioning and mechanical arm crawl implementation method based on ROS systems |
-
2018
- 2018-10-30 CN CN201811275792.8A patent/CN109382828B/en active Active
Patent Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN104625676A (en) * | 2013-11-14 | 2015-05-20 | 沈阳新松机器人自动化股份有限公司 | Shaft hole assembly industrial robot system and working method thereof |
CN104057290A (en) * | 2014-06-24 | 2014-09-24 | 中国科学院自动化研究所 | Method and system for assembling robot based on visual sense and force feedback control |
WO2016066616A1 (en) * | 2014-10-27 | 2016-05-06 | Kuka Systems Gmbh | Method and robot system for using an industrial robot for test jobs |
CN106826822A (en) * | 2017-01-25 | 2017-06-13 | 南京阿凡达机器人科技有限公司 | A kind of vision positioning and mechanical arm crawl implementation method based on ROS systems |
CN106799728A (en) * | 2017-03-16 | 2017-06-06 | 天津工业大学 | A kind of passive compliance device |
Cited By (47)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN109732610A (en) * | 2019-03-01 | 2019-05-10 | 北京航空航天大学 | Man-machine collaboration robot grasping system and its working method |
CN110355557A (en) * | 2019-07-05 | 2019-10-22 | 清华大学 | A kind of spiral insertion method of large-size axis parts hole workpiece assembly |
WO2021003869A1 (en) * | 2019-07-05 | 2021-01-14 | 清华大学 | Screw insertion method for large-size peg-in-hole workpiece assembly |
CN110501421A (en) * | 2019-07-24 | 2019-11-26 | 武汉大学 | A kind of track profiling method of detection based on mechanical arm |
CN110449882A (en) * | 2019-08-02 | 2019-11-15 | 珞石(北京)科技有限公司 | The search assembly method of binding force control |
CN110450143A (en) * | 2019-08-02 | 2019-11-15 | 珞石(北京)科技有限公司 | workpiece fatigue testing method based on cooperative robot |
CN110561421A (en) * | 2019-08-09 | 2019-12-13 | 哈尔滨工业大学(深圳) | Mechanical arm indirect dragging demonstration method and device |
CN110497423B (en) * | 2019-08-22 | 2022-08-16 | 泉州装备制造研究所 | Self-adaptive machining method for manipulator |
CN110497423A (en) * | 2019-08-22 | 2019-11-26 | 泉州装备制造研究所 | A kind of manipulator adaptive machining method |
CN110625611A (en) * | 2019-08-27 | 2019-12-31 | 上海卫星装备研究所 | Mechanical arm auxiliary component assembling method and system based on laser tracking measurement and force sensing combined control |
CN110561430A (en) * | 2019-08-30 | 2019-12-13 | 哈尔滨工业大学(深圳) | robot assembly track optimization method and device for offline example learning |
CN110561430B (en) * | 2019-08-30 | 2021-08-10 | 哈尔滨工业大学(深圳) | Robot assembly track optimization method and device for offline example learning |
CN114786888A (en) * | 2020-01-16 | 2022-07-22 | 欧姆龙株式会社 | Control device, control method, and control program |
CN111251277A (en) * | 2020-01-31 | 2020-06-09 | 武汉大学 | Human-computer collaboration tool submission system and method based on teaching learning |
CN111230873B (en) * | 2020-01-31 | 2022-02-01 | 武汉大学 | Teaching learning-based collaborative handling control system and method |
CN111251277B (en) * | 2020-01-31 | 2021-09-03 | 武汉大学 | Human-computer collaboration tool submission system and method based on teaching learning |
CN111230873A (en) * | 2020-01-31 | 2020-06-05 | 武汉大学 | Teaching learning-based collaborative handling control system and method |
CN111452039A (en) * | 2020-03-16 | 2020-07-28 | 华中科技大学 | Robot posture adjusting method and device under dynamic system, electronic equipment and medium |
CN111452039B (en) * | 2020-03-16 | 2022-05-17 | 华中科技大学 | Robot posture adjusting method and device under dynamic system, electronic equipment and medium |
CN111424380A (en) * | 2020-03-31 | 2020-07-17 | 山东大学 | Robot sewing system and method based on skill learning and generalization |
CN111546035A (en) * | 2020-04-07 | 2020-08-18 | 大连理工大学 | Online rapid gear assembly method based on learning and prediction |
CN111546035B (en) * | 2020-04-07 | 2021-07-02 | 大连理工大学 | Online rapid gear assembly method based on learning and prediction |
CN111515928A (en) * | 2020-04-15 | 2020-08-11 | 上海工程技术大学 | Mechanical arm motion control system |
CN111515928B (en) * | 2020-04-15 | 2023-03-31 | 上海工程技术大学 | Mechanical arm motion control system |
CN111993422B (en) * | 2020-08-11 | 2022-02-18 | 上海交通大学 | Robot axis and hole alignment control method based on uncalibrated vision |
CN111993422A (en) * | 2020-08-11 | 2020-11-27 | 上海交通大学 | Robot axis and hole alignment control method based on uncalibrated vision |
CN112223303B (en) * | 2020-08-27 | 2022-02-01 | 西安交通大学 | Robot automatic shaft hole axis alignment method and system based on plane surface friction |
CN112223303A (en) * | 2020-08-27 | 2021-01-15 | 西安交通大学 | Robot automatic shaft hole axis alignment method and system based on plane surface friction |
CN112060627A (en) * | 2020-09-08 | 2020-12-11 | 武汉大学 | Digital intelligent laying method and system for composite material |
CN112060627B (en) * | 2020-09-08 | 2022-01-14 | 武汉大学 | Digital intelligent laying method and system for composite material |
CN112428263A (en) * | 2020-10-16 | 2021-03-02 | 北京理工大学 | Mechanical arm control method and device and cluster model training method |
CN112264998A (en) * | 2020-10-28 | 2021-01-26 | 上海非夕机器人科技有限公司 | Method for assembling operation member and adapting member by robot, robot and controller |
US11833666B2 (en) | 2020-10-28 | 2023-12-05 | Shanghai Flexiv Robotics Technology Co., Ltd. | Method for assembling an operating member and an adapting member by a robot, robot, and controller |
CN112706161A (en) * | 2020-11-17 | 2021-04-27 | 中国航空工业集团公司北京长城航空测控技术研究所 | Gluing control system with intelligent sensing capability |
CN112605973A (en) * | 2020-11-19 | 2021-04-06 | 广东省科学院智能制造研究所 | Robot motor skill learning method and system |
CN112720478A (en) * | 2020-12-22 | 2021-04-30 | 深圳市优必选科技股份有限公司 | Robot torque control method and device, readable storage medium and robot |
CN112631201A (en) * | 2020-12-28 | 2021-04-09 | 佛山科学技术学院 | Hole searching control method and system for shaft hole assembly |
CN112665476A (en) * | 2020-12-29 | 2021-04-16 | 东风模具冲压技术有限公司 | Precision detection device for gripper of welding robot |
CN112917474B (en) * | 2021-01-19 | 2022-08-09 | 哈尔滨工业大学 | Skill extraction platform and method for arm-hand operation hexagonal wrench tool |
CN112917474A (en) * | 2021-01-19 | 2021-06-08 | 哈尔滨工业大学 | Skill extraction platform and method for arm-hand operation hexagonal wrench tool |
CN114161479A (en) * | 2021-12-24 | 2022-03-11 | 上海机器人产业技术研究院有限公司 | Robot dragging demonstration performance test system and test method |
CN114161479B (en) * | 2021-12-24 | 2023-10-20 | 上海机器人产业技术研究院有限公司 | Robot dragging teaching performance test system and test method |
CN115338610A (en) * | 2022-07-04 | 2022-11-15 | 中国科学院自动化研究所 | Biaxial hole assembling method and device, electronic device and storage medium |
CN115338610B (en) * | 2022-07-04 | 2024-02-13 | 中国科学院自动化研究所 | Double-shaft hole assembly method, device, electronic equipment and storage medium |
CN115685872A (en) * | 2022-09-05 | 2023-02-03 | 大连交通大学 | Robot assembly algorithm based on compliance control |
CN115685872B (en) * | 2022-09-05 | 2024-05-14 | 大连交通大学 | Robot assembly algorithm based on compliant control |
WO2024098787A1 (en) * | 2022-11-07 | 2024-05-16 | 节卡机器人股份有限公司 | Shaft hole assembly method and system, and storage medium |
Also Published As
Publication number | Publication date |
---|---|
CN109382828B (en) | 2021-04-16 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN109382828A (en) | A kind of Robot Peg-in-Hole assembly system and method based on learning from instruction | |
CN109483556B (en) | Robot polishing system and method based on teaching learning | |
CN106041926B (en) | A kind of industrial machinery arm strength/Position Hybrid Control method based on Kalman filter | |
Wang et al. | A hybrid visual servo controller for robust grasping by wheeled mobile robots | |
CN106625665B (en) | A kind of drilling milling machine device people's system of packaged type automatic addressing | |
CN109159151A (en) | A kind of mechanical arm space tracking tracking dynamic compensation method and system | |
Wen et al. | Elman fuzzy adaptive control for obstacle avoidance of mobile robots using hybrid force/position incorporation | |
CN106584093A (en) | Self-assembly system and method for industrial robots | |
CN106041927A (en) | Hybrid vision servo system and method combining eye-to-hand and eye-in-hand structures | |
Wang et al. | A framework of hybrid force/motion skills learning for robots | |
Cui et al. | A Darboux-frame-based formulation of spin-rolling motion of rigid objects with point contact | |
CN106383495B (en) | Curved surface profile constant force tracking method and application apparatus based on non-linear double-closed-loop control | |
Conticelli et al. | Nonlinear controllability and stability analysis of adaptive image-based systems | |
Giusti et al. | Flexible automation driven by demonstration: Leveraging strategies that simplify robotics | |
CN111515928B (en) | Mechanical arm motion control system | |
Roveda et al. | A control framework definition to overcome position/interaction dynamics uncertainties in force-controlled tasks | |
Liao et al. | Analysis of impact in robotic peg-in-hole assembly | |
CN114055467B (en) | Space pose online simulation system based on five-degree-of-freedom robot | |
Xu et al. | Nonlinear sliding mode control of manipulator based on iterative learning algorithm. | |
CN112045664A (en) | General mechanical arm controller based on ROS system | |
CN106092053A (en) | A kind of robot resetting system and localization method thereof | |
Campbell et al. | Superpositioning of behaviors learned through teleoperation | |
Lämmle et al. | Simulation-based learning of the peg-in-hole process using robot-skills | |
Merkt et al. | Towards shared autonomy applications using whole-body control formulations of locomanipulation | |
CN114800523B (en) | Mechanical arm track correction method, system, computer and readable storage medium |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
PB01 | Publication | ||
PB01 | Publication | ||
SE01 | Entry into force of request for substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
GR01 | Patent grant | ||
GR01 | Patent grant | ||
TR01 | Transfer of patent right |
Effective date of registration: 20220922 Address after: 3rd Floor, Building E2, Future Science and Technology City, No. 999 Gaoxin Avenue, Donghu New Technology Development Zone, Wuhan City, Hubei Province 430205 Patentee after: WUHAN COBOT TECHNOLOGY Co.,Ltd. Address before: 430072 Hubei Province, Wuhan city Wuchang District of Wuhan University Luojiashan Patentee before: WUHAN University |
|
TR01 | Transfer of patent right |