CN109571487A - A kind of robotic presentation learning method of view-based access control model - Google Patents
A kind of robotic presentation learning method of view-based access control model Download PDFInfo
- Publication number
- CN109571487A CN109571487A CN201811064626.3A CN201811064626A CN109571487A CN 109571487 A CN109571487 A CN 109571487A CN 201811064626 A CN201811064626 A CN 201811064626A CN 109571487 A CN109571487 A CN 109571487A
- Authority
- CN
- China
- Prior art keywords
- robot
- tool
- coordinate system
- teaching display
- sticking plastic
- 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
Landscapes
- Engineering & Computer Science (AREA)
- Robotics (AREA)
- Mechanical Engineering (AREA)
- Manipulator (AREA)
- Numerical Control (AREA)
Abstract
The invention discloses a kind of robotic presentation learning method of view-based access control model, this method realizes the study of demonstration task using teaching display-tool with sticking plastic and visual sensor.Demonstrator holds teaching display-tool with sticking plastic demonstration operation task first, then visual sensor obtains the characteristics of image of teaching display-tool with sticking plastic, according to the intrinsic parameter of visual sensor, obtain teaching track of the teaching display-tool with sticking plastic in presentation process, robot motion is controlled, end movement track of the robot in presentation process is obtained, Kalman filtering finally is carried out to robot end's motion profile, the study track of robot is obtained, realizes the study of demonstration task；The present invention is by simple vision aid, it is easy to which the six-dimensional pose information for extracting teaching display-tool with sticking plastic, the real-time for demonstrating study are good.Present invention reduces the teaching difficulty of operator, the inexperienced operator of milli can also carry out the demonstration teaching of robot.
Description
Technical field
The invention belongs to robot control fields, relate more specifically to a kind of robotic presentation study side of view-based access control model
Method.
Background technique
Robot is a kind of mechanical device that repetitive operation is completed under programming and control, and main task is to replace people
Class carries out some repeatability, the manual operations that environment is poor, risk is high.With the continuous development of technology, robot is in many necks
Domain has been able to that the mankind is replaced to be engaged in some heavy, complicated, dangerous activities.It can be improved operating efficiency, reduce operation wind
Danger is used widely in the industrial productions such as welding, assembly.
However, current most of robots all work in the space separated with the mankind, the mankind can only pass through demonstrator
Or programming is to achieve specific track.This mode needs operator to be familiar with robot operating system in advance, has one
Fixed program capability.Also, operator is in different spaces from robot, and operating difficulties, accuracy rate is low, and time-consuming efficiency is not
It is high.
In order to improve the independence of robot behavior, the difficulty that layman participates in robot control is reduced, demonstration is learned
Habit is come into being.Demonstration study is also known as learning from instruction, it makes movement of the robot by " observation " demonstrator (people or robot)
Behavior learns Motion Control Strategies, and then obtains motor skill, generates the independent behaviour as people.Rozo etc. uses demonstration
Learning method, which realizes a Six-DOF industrial robot, can independently complete the power control operation task of Ball-In-Box
(Rozo L.,Jiménez P.,Torras C..A robot learning from demonstration framework
to perform force-based manipulation tasks[J].Intelligent service robotics,
2013,6(1):33-51.).The corresponding each joint velocity of phase acquisition mechanical arm tail end six-dimensional force vector sum is demonstrated in demonstrator,
Each joint angular speed is exported when inputting current end moment information by gained model in the movement reproduction stage, to drive machinery
Arm controls the movement of bead in box and falls it at hole.This method needs to build action sequence with Hidden Markov Model
Mould, computationally intensive, real-time is not high.Liu Kun et al. is passed using Universal Robot as research object by a power/torque
Sensor perceives the teaching power of operator, is collected into power/torque voltage analog signal using data collecting card, carries out in host computer
Power/torque is converted to, the conversion of power and position is then carried out, realizes study (Liu Kun, Lee's generation that robot acts operator
In, direct teaching system research science and technology and engineering of the Wang Baoxiang based on UR robot, 2015,15 (28): 22-26).It should
Method is not filtered to the signal of force snesor acquisition and temperature-compensating, and the teaching fluctuation of people is larger, so showing
Teach precision not high, the study precision of robot is difficult to ensure.Wang Zhaoyang proposes a kind of class people's mechanical arm demonstration based on Kinect
Learning method obtains body motion information by Kinect camera, establishes the mapping relations mould between human arm and robot
Type realizes study (class people's mechanical arm demonstration Learning Studies [D] the master of Wang Zhaoyang based on Kinect to human arm motion
Academic dissertation, Heilungkiang: Harbin Institute of Technology, 2017.).This method is carried out by the human body motion capture function of Kinect
Robot arm motion tracking, but there are larger noises for the data of somatosensory device acquisition, are easy to cause the motion profile of study unstable.
Summary of the invention
Based on the above background, the present invention provides a kind of robotic presentation learning method of view-based access control model.This method includes step
It is rapid as follows:
Step S0: demonstrator holds the teaching display-tool with sticking plastic demonstration robot operation task to be learnt；
Step S1: using the teaching display-tool with sticking plastic image in a visual sensor acquisition presentation process, from the vision figure of acquisition
The characteristic information of teaching display-tool with sticking plastic is extracted as in；
Step S2: it according to the characteristics of image of S1 and the intrinsic parameter of visual sensor, obtains teaching display-tool with sticking plastic and is sat in video camera
Posture information under mark system；
Step S3: according between camera coordinate system and robot coordinate system relationship and S2 in teaching display-tool with sticking plastic taking the photograph
Posture information under camera coordinate system obtains posture information of the teaching display-tool with sticking plastic under robot coordinate system；
Step S4: according to pose of the teaching display-tool with sticking plastic of S3 under robot coordinate system, the movement of robot next step is obtained
Adjustment amount controls robot motion, the end pose of recorder people；
Step S5: repeating step S0 to S4, and until operation task demonstration terminates, the robot for obtaining entire presentation process is transported
Dynamic rail mark；
Step S6: Kalman filtering is carried out to the robot end track of S5, the study track of robot is obtained, will learn
Track is sent to robot, realizes the reproduction to demo content.
It is further described that wherein the visual sensor is a RGB-D camera, the teaching display-tool with sticking plastic is a cross
Frame, the upper end of cross, left end, right end and center is each fixes a bead, and four beads vary in color.
It is further described that wherein the characteristics of image of teaching display-tool with sticking plastic described in step S1 is as follows:
Visual pattern based on acquisition, using color segmentation, then image-region where obtaining four beads respectively exists
The pixel of bead is extracted in each region respectively, and then obtains the characteristic information of teaching display-tool with sticking plastic, the centre of sphere including four beads
Image coordinate (ui,vi) (i=1,2,3,4) and four beads centre of sphere depth zi(i=1,2,3,4).
It is further described that wherein posture information of the teaching display-tool with sticking plastic described in step S2 under camera coordinate system calculates such as
Under:
Using the bead centre of sphere in teaching display-tool with sticking plastic center as coordinate origin, using the right end of cross as X-axis positive direction, with ten
The upper end of cabinet frame is Y-axis positive direction, establishes teaching display-tool with sticking plastic coordinate system.According to the characteristic information of S1, teaching display-tool with sticking plastic coordinate system is obtained
In the position [p of camera coordinate systemx,py,pz]TIt is as follows:
Wherein, TinIt is the intrinsic parameter of visual sensor, (u0,v0) be astrosphere image coordinate, z0It is the depth of astrosphere
Degree.
Three top, left end and right end beads are obtained in camera coordinates using formula (1) according to the characteristic information of S1
The coordinate of system.According to the definition of teaching display-tool with sticking plastic coordinate system, and then teaching display-tool with sticking plastic coordinate system X-axis, Y-axis and Z axis can be obtained in video camera
Normalization direction vector n, o, a of coordinate system, in conjunction with the position vector [p of teaching display-tool with sticking plastic coordinate systemx,py,pz]T, obtain teaching work
Has the position auto―control T in camera coordinate systemcIt is as follows:
It is further described that wherein posture information of the teaching display-tool with sticking plastic described in step S3 under robot coordinate system is as follows:
According to position auto―control T of the teaching display-tool with sticking plastic of S2 under camera coordinate systemcAnd visual sensor and robot are sat
Mark the relational matrix T of systemm, it is as follows to obtain position auto―control T of the teaching display-tool with sticking plastic under robot coordinate system:
T=TcTm (3)
According to general rotation transformation, can by the position auto―control T equivalence transformation of formula (3) at six-dimensional pose vector [dx, dy,
dz,rx,ry,rz]T。
It is further described that wherein the movement adjustment amount of robot next step described in step S4 is as follows:
Using formula (3), teaching display-tool with sticking plastic is obtained in current pose [dx, dy, dz, the r of robot coordinate systemx,ry,rz]T。
Using the feature for demonstrating start time as initial characteristics, teaching display-tool with sticking plastic can be obtained in the initial pose [dx of robot coordinate system0,dy0,
dz0,rx0,ry0,rz0]T.According to current pose and initial pose, teaching display-tool with sticking plastic is obtained in the pose variable quantity of robot coordinate system
Are as follows:
Therefore, movement adjustment amount [x, y, z, the θ of robot next step are obtainedx,θy,θz]TIt is as follows:
Wherein, λpIt is regulation coefficient.
Movement adjustment amount shown in formula (5) is sent to robot, controls robot motion, recorder people's post exercise
End pose J.
It is further described that wherein the robot motion track of entire presentation process described in step S5 is as follows:
Each control period repeats step S0 to S4, the end pose of recorder people.After operation task is demonstrated,
Obtaining robot motion track is are as follows:
W=(J0,J1,…,Jm) (6)
Wherein, m is the control periodicity of presentation process.
It is further described that wherein robot learning track described in step S6 is as follows:
Establish the prediction model of Kalman filtering:
Wherein,It is the robot pose estimated value of i+1 time, Ki+1It is the kalman gain coefficient of i+1 time, Ji+1
It is the robot pose true value of i+1 time.
Kalman gain coefficient update is as follows:
Ki+1=(Pi+Q)/(Pi+Q+R) (8)
Wherein, PiIt is the variance of last estimated value, Q is the variance of Gaussian noise, and R is the variance of true value.
The variance of estimated value calculates as follows:
Pi+1=(1-Ki+1)Pi (9)
According to the robot motion track W of S5, using formula (7)~(9), karr is carried out to the robot motion track of S5
Graceful filtering obtains the study track L of robot are as follows:
Study track L is sent to robot, the reproduction to demonstration task can be realized.
Based on the above-mentioned technical proposal it is found that the invention has the following advantages: the teachings skill such as traditional demonstrator, programming
Art is more demanding to operator, and teaching process is cumbersome, time-consuming inefficient.Current demonstration learning method mostly uses power/torque
Sensor, it is at high cost, and also collection process is complicated, and needs to carry out temperature-compensating to the data of acquisition.Based on body-sensing camera into
The method of row demonstration study obtains body motion information and is easier to, but learning effect is limited to the motion-captured essence of body-sensing camera
Degree.
In order to improve the independence of robot behavior, the difficulty that layman participates in robot control is reduced, the present invention
For the demonstration study of robot, demonstrator is held the teaching display-tool with sticking plastic demonstration robot operation task to be learnt, is regarded using one
Feel the teaching display-tool with sticking plastic image in sensor acquisition presentation process, extract the motion information of teaching display-tool with sticking plastic, realizes robot to demonstration
The study of task.
The visual sensor and teaching display-tool with sticking plastic that the present invention uses are cheap, at low cost.The present invention is to demonstrate start time
State be original state, demonstration study can start under any position and posture, significantly very high demonstration learning efficiency.This hair
It is bright by simple vision aid, it is easy to the six-dimensional pose information for extracting teaching display-tool with sticking plastic, the real-time for demonstrating study are good.This hair
The bright teaching difficulty for reducing operator, the inexperienced operator of milli can also carry out the demonstration teaching of robot.
Detailed description of the invention
Fig. 1 is the robotic presentation learning method flow chart of view-based access control model of the invention.
Specific embodiment
Be described in detail with reference to the accompanying drawing to the embodiment of the present invention: the present embodiment is being with technical solution of the present invention
Under the premise of implemented, in conjunction with detailed embodiment and specific operating process, but protection scope of the present invention is not limited to down
State embodiment.
The invention discloses a kind of robotic presentation learning method of view-based access control model, the present invention holds teaching by demonstrator
The tool demonstration robot operation task to be learnt utilizes the teaching display-tool with sticking plastic figure in a visual sensor acquisition presentation process
Picture extracts the motion information of teaching display-tool with sticking plastic, realizes study of the robot to demonstration task.
More specifically, as a preferred embodiment of the present invention, the robot of view-based access control model of the invention as shown in figure 1
Demonstrate learning method flow chart.It demonstrates in learning process, demonstrator holds teaching display-tool with sticking plastic demonstration operation task first, then vision
The characteristics of image that sensor obtains teaching display-tool with sticking plastic in presentation process obtains teaching display-tool with sticking plastic and exists according to the intrinsic parameter of visual sensor
Teaching track is finally converted into robot end track by the teaching track under camera coordinate system, and to robot end's rail
Mark carries out Kalman filtering, obtains the study track of robot, to realize study of the robot to demonstration task.This method packet
Include following steps:
Step 1: demonstrator holds the teaching display-tool with sticking plastic demonstration robot operation task to be learnt, a visual sensing is utilized
Device acquires the teaching display-tool with sticking plastic image in presentation process, and the characteristic information of teaching display-tool with sticking plastic is extracted from the visual pattern of acquisition；
Step 2: obtaining teaching display-tool with sticking plastic according to the characteristics of image of the first step and the intrinsic parameter of visual sensor and imaging
Posture information under machine coordinate system；
Step 3: according between camera coordinate system and robot coordinate system relationship and second step in teaching display-tool with sticking plastic
Posture information under camera coordinate system obtains posture information of the teaching display-tool with sticking plastic under robot coordinate system；
Step 4: the pose according to the teaching display-tool with sticking plastic of third step under robot coordinate system, obtains robot next step
Adjustment amount is moved, robot motion, the end pose of recorder people are controlled；
Step 5: repeating the first step to the 4th step, demonstrating until operation task terminates, and obtains the machine of entire presentation process
People's motion profile；
Step 6: the robot end track to the 5th step carries out Kalman filtering, the study track of robot is obtained, it will
Study track is sent to robot, realizes the reproduction to demo content.
The first step, specific as follows:
The teaching display-tool with sticking plastic image of view-based access control model sensor acquisition, using color segmentation, where obtaining four beads respectively
Image-region, then extracts the pixel of bead respectively in each region, and then obtains the characteristic information of teaching display-tool with sticking plastic, including
Centre of sphere image coordinate (the u of four beadsi,vi) (i=1,2,3,4) and four beads centre of sphere depth zi(i=1,2,3,
4)。
The second step, specific as follows:
Teaching display-tool with sticking plastic coordinate system is obtained in the position of camera coordinate system using formula (1) according to the characteristic information of the first step
Set [px,py,pz]T.According to the definition of teaching display-tool with sticking plastic coordinate system, teaching display-tool with sticking plastic coordinate system is obtained in the posture of camera coordinate system,
And then teaching display-tool with sticking plastic coordinate system shown in formula (2) is obtained in the position auto―control of camera coordinate system.
Wherein formula (1) and formula (2) are obtained by step in detail below:
Using the bead centre of sphere in teaching display-tool with sticking plastic center as coordinate origin, using the right end of cross as X-axis positive direction, with ten
The upper end of cabinet frame is Y-axis positive direction, establishes teaching display-tool with sticking plastic coordinate system.According to the characteristic information of S1, teaching display-tool with sticking plastic coordinate system is obtained
In the position [p of camera coordinate systemx,py,pz]TIt is as follows:
Wherein, TinIt is the intrinsic parameter of visual sensor, (u0,v0) be astrosphere image coordinate, z0It is the depth of astrosphere
Degree.
Three top, left end and right end beads are obtained in video camera using formula (1) according to the characteristic information of the first step
The coordinate of coordinate system.According to the definition of teaching display-tool with sticking plastic coordinate system, and then teaching display-tool with sticking plastic coordinate system X-axis, Y-axis and Z axis can be obtained and taken the photograph
Normalization direction vector n, o, a of camera coordinate system, in conjunction with the position vector [p of teaching display-tool with sticking plastic coordinate systemx,py,pz]T, shown
Teaching and administrative staff has the position auto―control T in camera coordinate systemcIt is as follows:
The third step, specific as follows:
According to position auto―control of the teaching display-tool with sticking plastic of second step under camera coordinate system and visual sensor and robot
The relational matrix of coordinate system obtains pose of the teaching display-tool with sticking plastic under robot coordinate system using formula (3).
Wherein formula (3) is obtained by step in detail below:
According to position auto―control T of the teaching display-tool with sticking plastic of second step under camera coordinate systemcAnd visual sensor and machine
The relational matrix T of people's coordinate systemm, it is as follows to obtain position auto―control T of the teaching display-tool with sticking plastic under robot coordinate system:
T=TcTm (3)
According to general rotation transformation, can by the position auto―control T equivalence transformation of formula (3) at six-dimensional pose vector [dx, dy,
dz,rx,ry,rz]T。
4th step, specific as follows:
It obtains teaching display-tool with sticking plastic using formula (4) according to pose of the teaching display-tool with sticking plastic of third step under robot coordinate system and exists
The pose variable quantity of robot coordinate system.Using formula (5), the movement adjustment amount of robot next step is obtained, controls robot
Movement, recorder people's post exercise end pose.
Wherein formula (4) and formula (5) are obtained by step in detail below:
Using the feature for demonstrating start time as initial characteristics, teaching display-tool with sticking plastic can be obtained in the initial pose of robot coordinate system
[dx0,dy0,dz0,rx0,ry0,rz0]T.According to current pose and initial pose, teaching display-tool with sticking plastic is obtained in the position of robot coordinate system
Appearance variable quantity are as follows:
Therefore, movement adjustment amount [x, y, z, the θ of robot next step are obtainedx,θy,θz]TIt is as follows:
Wherein, λpIt is regulation coefficient.
5th step, specific as follows:
Each control period repeats the step first step to the 4th step, the end pose of recorder people.Operation task demonstration
After, obtain robot motion track shown in formula (6).
W=(J0,J1,…,Jm) (6)
Wherein, m is the control periodicity of presentation process.
6th step, specific as follows:
Based on the robot motion track that the 5th step obtains, Kalman prediction model is established according to formula (7), according to
Formula (8) and (9) update kalman gain coefficient, carry out Kalman filtering to robot motion track, obtain shown in formula (10)
Robot learning track, study track is sent to robot, realizes the reproduction of demonstration task.
Establish the prediction model of Kalman filtering:
Wherein,It is the robot pose estimated value of i+1 time, Ki+1It is the kalman gain coefficient of i+1 time, Ji+1
It is the robot pose true value of i+1 time.
Kalman gain coefficient update is as follows:
Ki+1=(Pi+Q)/(Pi+Q+R) (8)
Wherein, PiIt is the variance of last estimated value, Q is the variance of Gaussian noise, and R is the variance of true value.
The variance of estimated value calculates as follows:
Pi+1=(1-Ki+1)Pi (9)
The robot motion track of S5 is carried out using formula (7)~(9) according to the robot motion track W of step 5
Kalman filtering obtains the study track L of robot are as follows:
Particular embodiments described above has carried out further in detail the purpose of the present invention, technical scheme and beneficial effects
Describe in detail bright, it should be understood that the above is only a specific embodiment of the present invention, is not intended to restrict the invention, it is all
Within the spirit and principles in the present invention, any modification, equivalent substitution, improvement and etc. done should be included in protection of the invention
Within the scope of.
Claims (8)
1. a kind of robotic presentation learning method of view-based access control model, comprising the following steps:
Step S0: demonstrator holds the teaching display-tool with sticking plastic demonstration robot operation task to be learnt；
Step S1: using the teaching display-tool with sticking plastic image in a visual sensor acquisition presentation process, from the visual pattern of acquisition
Extract the characteristic information of teaching display-tool with sticking plastic；
Step S2: according to the characteristics of image of S1 and the intrinsic parameter of visual sensor, teaching display-tool with sticking plastic is obtained in camera coordinate system
Under posture information；
Step S3: according between camera coordinate system and robot coordinate system relationship and S2 in teaching display-tool with sticking plastic in video camera
Posture information under coordinate system obtains posture information of the teaching display-tool with sticking plastic under robot coordinate system；
Step S4: according to pose of the teaching display-tool with sticking plastic of S3 under robot coordinate system, the movement adjustment of robot next step is obtained
Amount controls robot motion, the end pose of recorder people；
Step S5: repeating step S0 to S4, and demonstrating until operation task terminates, and obtains robot motion's rail of entire presentation process
Mark；
Step S6: Kalman filtering is carried out to the robot end track of S5, the study track of robot is obtained, track will be learnt
It is sent to robot, realizes the reproduction to demo content.
2. the robotic presentation learning method of view-based access control model according to claim 1, wherein the visual sensor is one
A RGB-D camera, the teaching display-tool with sticking plastic are a crosses, the upper end of cross, left end, right end and center it is each fix it is one small
Ball, and four beads vary in color.
3. the robotic presentation learning method of view-based access control model according to claim 1, wherein teaching described in step S1
The characteristics of image of tool is as follows:
Visual pattern based on acquisition, using color segmentation, image-region where obtaining four beads respectively, then each
The pixel of bead is extracted in region respectively, and then obtains the characteristic information of teaching display-tool with sticking plastic.Centre of sphere image including four beads
Coordinate (ui,vi) (i=1,2,3,4) and four beads centre of sphere depth zi(i=1,2,3,4).
4. the robotic presentation learning method of view-based access control model according to claim 1, wherein teaching described in step S2
Posture information of the tool under camera coordinate system calculates as follows:
Using the bead centre of sphere in teaching display-tool with sticking plastic center as coordinate origin, using the right end of cross as X-axis positive direction, with cross
Upper end be Y-axis positive direction, establish teaching display-tool with sticking plastic coordinate system.According to the characteristic information of S1, obtains teaching display-tool with sticking plastic coordinate system and taking the photograph
Position [the p of camera coordinate systemx,py,pz]TIt is as follows:
Wherein, TinIt is the intrinsic parameter of visual sensor, (u0,v0) be astrosphere image coordinate, z0It is the depth of astrosphere.
Three top, left end and right end beads are obtained in camera coordinate system using formula (1) according to the characteristic information of S1
Coordinate.According to the definition of teaching display-tool with sticking plastic coordinate system, and then teaching display-tool with sticking plastic coordinate system X-axis, Y-axis and Z axis can be obtained in camera coordinates
Normalization direction vector n, o, a of system, in conjunction with the position vector [p of teaching display-tool with sticking plastic coordinate systemx,py,pz]T, obtain teaching display-tool with sticking plastic and exist
The position auto―control T of camera coordinate systemcIt is as follows:
5. the robotic presentation learning method of view-based access control model according to claim 1, wherein teaching described in step S3
Posture information of the tool under robot coordinate system is as follows:
According to position auto―control T of the teaching display-tool with sticking plastic of S2 under camera coordinate systemcAnd visual sensor and robot coordinate system
Relational matrix Tm, it is as follows to obtain position auto―control T of the teaching display-tool with sticking plastic under robot coordinate system:
T=TcTm (3)
It, can be by the position auto―control T equivalence transformation of formula (3) at six-dimensional pose vector [dx, dy, dz, r according to general rotation transformationx,
ry,rz]T。
6. the robotic presentation learning method of view-based access control model according to claim 1, wherein machine described in step S4
The movement adjustment amount of people's next step is as follows:
Using formula (3), teaching display-tool with sticking plastic is obtained in current pose [dx, dy, dz, the r of robot coordinate systemx,ry,rz]T.With demonstration
The feature of start time is initial characteristics, can obtain teaching display-tool with sticking plastic in the initial pose [dx of robot coordinate system0,dy0,dz0,rx0,
ry0,rz0]T.According to current pose and initial pose, teaching display-tool with sticking plastic is obtained in the pose variable quantity of robot coordinate system are as follows:
Therefore, movement adjustment amount [x, y, z, the θ of robot next step are obtainedx,θy,θz]TIt is as follows:
Wherein, λpIt is regulation coefficient.
Movement adjustment amount shown in formula (5) is sent to robot, controls robot motion, recorder people's post exercise end
Pose J.
7. the robotic presentation learning method of view-based access control model according to claim 1, wherein entire described in step S5
The robot motion track of presentation process is as follows:
Each control period repeats step S0 to S4, the end pose of recorder people.After operation task is demonstrated, obtain
Robot motion track is are as follows:
W=(J0,J1,…,Jm) (6)
Wherein, m is the control periodicity of presentation process.
8. the robotic presentation learning method of view-based access control model according to claim 1, wherein machine described in step S6
It is as follows that people learns track:
Establish the prediction model of Kalman filtering:
Wherein,It is the robot pose estimated value of i+1 time, Ki+1It is the kalman gain coefficient of i+1 time, Ji+1Be i-th+
1 robot pose true value.
Kalman gain coefficient update is as follows:
Ki+1=(Pi+Q)/(Pi+Q+R) (8)
Wherein, PiIt is the variance of last estimated value, Q is the variance of Gaussian noise, and R is the variance of true value.
The variance of estimated value calculates as follows:
Pi+1=(1-Ki+1)Pi (9)
According to the robot motion track W of S5, using formula (7)~(9), Kalman's filter is carried out to the robot motion track of S5
Wave obtains the study track L of robot are as follows:
Study track L is sent to robot, the reproduction to demonstration task can be realized.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811064626.3A CN109571487B (en) | 2018-09-12 | 2018-09-12 | Robot demonstration learning method based on vision |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811064626.3A CN109571487B (en) | 2018-09-12 | 2018-09-12 | Robot demonstration learning method based on vision |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN109571487A true CN109571487A (en) | 2019-04-05 |
| CN109571487B CN109571487B (en) | 2020-08-28 |
Family
ID=65919729
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201811064626.3A Active CN109571487B (en) | 2018-09-12 | 2018-09-12 | Robot demonstration learning method based on vision |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109571487B (en) |
Cited By (12)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110065068A (en) * | 2019-04-08 | 2019-07-30 | 浙江大学 | A kind of robotic asssembly operation programming by demonstration method and device based on reverse-engineering |
| CN110170995A (en) * | 2019-05-09 | 2019-08-27 | 广西安博特智能科技有限公司 | A kind of quick teaching method of robot based on stereoscopic vision |
| CN110315544A (en) * | 2019-06-24 | 2019-10-11 | 南京邮电大学 | A kind of robot manipulation's learning method based on video image demonstration |
| CN110480642A (en) * | 2019-10-16 | 2019-11-22 | 遨博（江苏）机器人有限公司 | Industrial robot and its method for utilizing vision calibration user coordinate system |
| 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 |
| CN110587579A (en) * | 2019-09-30 | 2019-12-20 | 厦门大学嘉庚学院 | Kinect-based robot teaching programming guiding method |
| CN110900609A (en) * | 2019-12-11 | 2020-03-24 | 浙江钱江机器人有限公司 | Robot teaching device and method thereof |
| CN110919626A (en) * | 2019-05-16 | 2020-03-27 | 广西大学 | Robot handheld teaching device and method based on stereoscopic vision |
| CN111002289A (en) * | 2019-11-25 | 2020-04-14 | 华中科技大学 | Robot online teaching method and device, terminal device and storage medium |
| CN111152230A (en) * | 2020-04-08 | 2020-05-15 | 季华实验室 | Robot teaching method, system, teaching robot and storage medium |
| CN112008692A (en) * | 2019-05-31 | 2020-12-01 | 精工爱普生株式会社 | Teaching method |
| CN112509392A (en) * | 2020-12-16 | 2021-03-16 | 复旦大学 | Robot behavior teaching method based on meta-learning |
Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102135776A (en) * | 2011-01-25 | 2011-07-27 | 解则晓 | Industrial robot control system based on visual positioning and control method thereof |
| CN102581445A (en) * | 2012-02-08 | 2012-07-18 | 中国科学院自动化研究所 | Visual real-time deviation rectifying system and visual real-time deviation rectifying method for robot |
| CN105196292A (en) * | 2015-10-09 | 2015-12-30 | 浙江大学 | Visual servo control method based on iterative duration variation |
| CN106142092A (en) * | 2016-07-26 | 2016-11-23 | 张扬 | A kind of method robot being carried out teaching based on stereovision technique |
| CN106553195A (en) * | 2016-11-25 | 2017-04-05 | 中国科学技术大学 | Object 6DOF localization method and system during industrial robot crawl |
| CN107160364A (en) * | 2017-06-07 | 2017-09-15 | 华南理工大学 | A kind of industrial robot teaching system and method based on machine vision |
| CN108161882A (en) * | 2017-12-08 | 2018-06-15 | 华南理工大学 | A kind of robot teaching reproducting method and device based on augmented reality |
| EP3366433A1 (en) * | 2017-02-09 | 2018-08-29 | Canon Kabushiki Kaisha | Method of controlling robot, method of teaching robot, and robot system |
-
2018
- 2018-09-12 CN CN201811064626.3A patent/CN109571487B/en active Active
Patent Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102135776A (en) * | 2011-01-25 | 2011-07-27 | 解则晓 | Industrial robot control system based on visual positioning and control method thereof |
| CN102581445A (en) * | 2012-02-08 | 2012-07-18 | 中国科学院自动化研究所 | Visual real-time deviation rectifying system and visual real-time deviation rectifying method for robot |
| CN105196292A (en) * | 2015-10-09 | 2015-12-30 | 浙江大学 | Visual servo control method based on iterative duration variation |
| CN106142092A (en) * | 2016-07-26 | 2016-11-23 | 张扬 | A kind of method robot being carried out teaching based on stereovision technique |
| CN106553195A (en) * | 2016-11-25 | 2017-04-05 | 中国科学技术大学 | Object 6DOF localization method and system during industrial robot crawl |
| EP3366433A1 (en) * | 2017-02-09 | 2018-08-29 | Canon Kabushiki Kaisha | Method of controlling robot, method of teaching robot, and robot system |
| CN107160364A (en) * | 2017-06-07 | 2017-09-15 | 华南理工大学 | A kind of industrial robot teaching system and method based on machine vision |
| CN108161882A (en) * | 2017-12-08 | 2018-06-15 | 华南理工大学 | A kind of robot teaching reproducting method and device based on augmented reality |
Non-Patent Citations (1)
| Title |
|---|
| 倪自强: "基于视觉引导的工业机器人示教编程系统", 《北京航空航天大学学报》 * |
Cited By (16)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110065068A (en) * | 2019-04-08 | 2019-07-30 | 浙江大学 | A kind of robotic asssembly operation programming by demonstration method and device based on reverse-engineering |
| CN110170995A (en) * | 2019-05-09 | 2019-08-27 | 广西安博特智能科技有限公司 | A kind of quick teaching method of robot based on stereoscopic vision |
| CN110919626B (en) * | 2019-05-16 | 2023-03-14 | 广西大学 | Robot handheld teaching device and method based on stereoscopic vision |
| CN110919626A (en) * | 2019-05-16 | 2020-03-27 | 广西大学 | Robot handheld teaching device and method based on stereoscopic vision |
| CN112008692A (en) * | 2019-05-31 | 2020-12-01 | 精工爱普生株式会社 | Teaching method |
| CN110315544A (en) * | 2019-06-24 | 2019-10-11 | 南京邮电大学 | A kind of robot manipulation's learning method based on video image demonstration |
| 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 |
| CN110587579A (en) * | 2019-09-30 | 2019-12-20 | 厦门大学嘉庚学院 | Kinect-based robot teaching programming guiding method |
| CN110480642A (en) * | 2019-10-16 | 2019-11-22 | 遨博（江苏）机器人有限公司 | Industrial robot and its method for utilizing vision calibration user coordinate system |
| CN111002289A (en) * | 2019-11-25 | 2020-04-14 | 华中科技大学 | Robot online teaching method and device, terminal device and storage medium |
| CN111002289B (en) * | 2019-11-25 | 2021-08-17 | 华中科技大学 | Robot online teaching method and device, terminal device and storage medium |
| CN110900609A (en) * | 2019-12-11 | 2020-03-24 | 浙江钱江机器人有限公司 | Robot teaching device and method thereof |
| CN111152230B (en) * | 2020-04-08 | 2020-09-04 | 季华实验室 | Robot teaching method, system, teaching robot and storage medium |
| CN111152230A (en) * | 2020-04-08 | 2020-05-15 | 季华实验室 | Robot teaching method, system, teaching robot and storage medium |
| CN112509392A (en) * | 2020-12-16 | 2021-03-16 | 复旦大学 | Robot behavior teaching method based on meta-learning |
Also Published As
| Publication number | Publication date |
|---|---|
| CN109571487B (en) | 2020-08-28 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109571487A (en) | A kind of robotic presentation learning method of view-based access control model | |
| Qi et al. | Contour moments based manipulation of composite rigid-deformable objects with finite time model estimation and shape/position control | |
| CN108284436B (en) | Remote mechanical double-arm system with simulation learning mechanism and method | |
| CN109102525A (en) | A kind of mobile robot follow-up control method based on the estimation of adaptive pose | |
| CN105512621A (en) | Kinect-based badminton motion guidance system | |
| CN106371442B (en) | A kind of mobile robot control method based on the transformation of tensor product model | |
| CN109108942A (en) | The mechanical arm motion control method and system of the real-time teaching of view-based access control model and adaptive DMPS | |
| CN105760894A (en) | Robot navigation method based on machine vision and machine learning | |
| Li et al. | Visual servoing of wheeled mobile robots without desired images | |
| Li et al. | Development of kinect based teleoperation of nao robot | |
| Chen et al. | Transferable active grasping and real embodied dataset | |
| CN109636856B (en) | Object six-dimensional pose information joint measurement method based on HOG feature fusion operator | |
| Ikeuchi et al. | Applying learning-from-observation to household service robots: three common-sense formulation | |
| Gao et al. | Kinect-based motion recognition tracking robotic arm platform | |
| CN108621164A (en) | Taiji push hands machine people based on depth camera | |
| Wang et al. | Design and implementation of humanoid robot behavior imitation system based on skeleton tracking | |
| Yin et al. | Monitoring-based visual servoing of wheeled mobile robots | |
| CN113134839A (en) | Robot precision flexible assembly method based on vision and force position image learning | |
| Lang et al. | Visual servoing with LQR control for mobile robots | |
| CN114434441A (en) | Mobile robot visual servo tracking control method based on self-adaptive dynamic programming | |
| CN113492404B (en) | Humanoid robot action mapping control method based on machine vision | |
| CN112257655B (en) | Method for robot to recognize human body sewing action | |
| Lepora et al. | Pose-based servo control with soft tactile sensing | |
| Jayasurya et al. | Gesture controlled AI-robot using Kinect | |
| CN205870565U (en) | Robot control system is felt to long -range body based on kinect |
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 |