CN114310917A - Joint track error compensation method for oil pipe transfer robot - Google Patents
Joint track error compensation method for oil pipe transfer robot Download PDFInfo
- Publication number
- CN114310917A CN114310917A CN202210238378.XA CN202210238378A CN114310917A CN 114310917 A CN114310917 A CN 114310917A CN 202210238378 A CN202210238378 A CN 202210238378A CN 114310917 A CN114310917 A CN 114310917A
- Authority
- CN
- China
- Prior art keywords
- joint
- oil pipe
- transfer robot
- dynamic
- pipe transfer
- 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
Images
Landscapes
- Manipulator (AREA)
Abstract
The invention belongs to the technical field of joint track error compensation of an oil pipe transfer robot, and particularly relates to a joint track error compensation method of an oil pipe transfer robot. The joint track error compensation method for the oil pipe carrying robot improves the motion precision of the oil pipe carried by the tail end executing mechanism of the carrying robot, improves the processing and repairing efficiency of manufacturing equipment, and reduces the running cost of the manufacturing equipment. The joint track error compensation method of the oil pipe transfer robot comprises the steps of constructing a joint topological graph and a joint angle characteristic matrix of the oil pipe transfer robot, constructing a joint dynamic graph network model of the oil pipe transfer robot, constructing a joint dynamic graph circulating network model of the oil pipe transfer robot, predicting joint angles of future sequences based on joint angles of historical sequences and the like.
Description
Technical Field
The invention belongs to the technical field of joint track error compensation of an oil pipe transfer robot, and particularly relates to a joint track error compensation method of an oil pipe transfer robot.
Background
The petroleum oil pipe processing or repairing equipment is widely applied to the procedures of oil field oil pipe processing, repairing, paint spraying, detecting and the like. It is worth noting that because oil pipe length is long, weight is big, and artifical material loading is wasted time and energy, consequently the link mechanism of robot is usually adopted at present stage to carry to fundamentally solve the transport difficult problem of oil pipe.
However, after further research, the inventor finds that, because the oil pipe is long and heavy, the robot has many joints, the joints of the robot have slight deviation, the positions of all the joints behind the joints are changed, and all the components are offset; and because no feedback exists, even if all joint variables are set, the tail end of the oil pipe transfer robot cannot be accurately positioned at a given position, so that the movement precision is easily reduced in the transfer process, and the equipment processing and repairing precision is directly influenced. With the development of modern science and technology, because the error compensation precision and the joint motion have a direct important relationship and are influenced by various geometric errors and non-geometric errors, the spatial correlation between joints is highly dynamic and is determined by a real-time joint motion state and a connecting rod topological structure, and meanwhile, because of nonlinear change and periodic cross mixing, the connecting rod motion has strong nonlinear dynamic correlation in space and time dimensions, so that a simple, reliable and accurate error compensation method must be established in order to improve the motion precision of a carrying robot in the carrying process and reduce the carrying, processing, testing and maintenance costs.
Disclosure of Invention
The invention provides a joint track error compensation method of an oil pipe transfer robot, which improves the motion precision of an oil pipe transferred by a tail end execution mechanism of the transfer robot, improves the processing and repairing efficiency of manufacturing equipment and reduces the running cost of the manufacturing equipment.
In order to solve the technical problems, the invention adopts the following technical scheme:
a joint track error compensation method for an oil pipe transfer robot comprises the following steps:
step 1: constructing a joint topological graph and a joint angle characteristic matrix of the oil pipe carrying robot;
step 2: constructing a joint dynamic graph network model of the oil pipe transfer robot to obtain dynamic node characteristics; inputting the dynamic node characteristics into graph convolution for dynamic filtering;
and step 3: constructing a joint dynamic graph circulation network model of the oil pipe transfer robot;
and 4, step 4: predicting joint angles of a future sequence based on joint angles of a historical sequence according to a joint dynamic graph circulation network model of the oil pipe carrying robot;
preferably, the method further comprises the following steps:
and 5: and performing deviation correction on a predicted value obtained by the joint dynamic graph circulation network model of the oil pipe transfer robot.
Preferably, the step 5 can be specifically described as:
preferably, the step of constructing the joint topology map of the pipe handling robot in step 1 may be specifically described as follows:
defining a joint topological graph of the oil pipe transfer robot as follows:(ii) a Wherein V represents the set of joint nodes of the oil pipe transfer robot,(ii) a n is the number of nodes of the joint node; e represents the set of edges where the joints are located; a represents the static adjacency matrix between joint nodes.
Preferably, the step of constructing the joint angle feature matrix of the pipe handling robot in step 1 may be specifically described as follows:
defining a joint angle characteristic matrix of the oil pipe transfer robot as follows:(ii) a Wherein R is the diameter of the joint node; and p is the attribute characteristic number of the joint node.
Preferably, the step of the network model of the joint dynamic diagram of the oil pipe transfer robot in step 2 may be specifically described as:
at each time step, the joint angle is connected in seriesTime t and hidden stateObtaining dynamic node characteristics I as input of joint dynamic graph network modelt(ii) a Wherein the content of the first and second substances,;b is the block size, n is the number of nodes, d is the characteristic dimension, TtIs the articulation cycle time.
Preferably, the step of inputting the dynamic node characteristics into the graph convolution in the step 2 to perform dynamic filtering may be specifically described as:
dynamic node characteristics ItThe input graph is convolved and the output is:(ii) a Wherein the dynamic filtering tensor,Representing graph convolution operationsIn order to do so,are the learned filter parameters.
Preferably, the step 3 can be specifically described as:
the joint dynamic graph circulation network model of the oil pipe transfer robot constructed by adopting the k-hop algorithm is expressed as follows:
wherein HinInputs for each node in the oil pipe transfer robot,in order to control the input signal(s),in order to be the parameters of the dynamic map of the joint,for the pre-defined static map parameters of the joint,a degree matrix after node self-loop is introduced;
preferably, the step 4 may be specifically described as:
at time t, the angle value (theta) of each joint is obtained1,θ2,…, θi) Giving a training step length M; will be provided withToPredicting output sequences as training inputsTo。
The invention provides an oil pipe transfer robot joint track error compensation method which comprises the steps of constructing a joint topological graph and a joint angle characteristic matrix of an oil pipe transfer robot, constructing a joint dynamic graph network model of the oil pipe transfer robot, constructing a joint dynamic graph circulating network model of the oil pipe transfer robot, predicting joint angles of future sequences based on joint angles of historical sequences and the like. The joint trajectory error compensation method of the oil pipe transfer robot with the steps and characteristics carries out real-time compensation on joint trajectory errors of the oil pipe transfer robot by establishing a nonlinear dynamic space-time correlation model, reduces repeated motion of repair equipment, improves motion precision of a tail end execution mechanism, increases machining and repair efficiency of the equipment, and reduces operation cost of the equipment.
Drawings
The accompanying drawings, which are included to provide a further understanding of the invention and are incorporated in and constitute a part of this specification, illustrate embodiments of the invention and together with the description serve to explain the principles of the invention and not to limit the invention.
FIG. 1 is a schematic flow chart of a joint trajectory error compensation method for an oil pipe transfer robot according to the present invention;
FIG. 2 is a flow chart schematic of constructing a joint dynamic graph network model of the oil pipe transfer robot;
FIG. 3 is a diagram convolution flow schematic of a joint dynamics diagram cycle network model;
FIG. 4 is a schematic representation of joint prediction for a joint dynamics graph cycle network model.
Detailed Description
The invention provides a joint track error compensation method of an oil pipe transfer robot, which improves the motion precision of an oil pipe transferred by a tail end execution mechanism of the transfer robot, improves the processing and repairing efficiency of manufacturing equipment and reduces the running cost of the manufacturing equipment.
In order to make the objects, technical solutions and advantages of the present invention more apparent, the present invention is described in further detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative of the invention and are not intended to limit the invention.
The invention provides a joint track error compensation method of an oil pipe transfer robot, which specifically comprises the following steps:
step 1: and constructing a joint topological graph and a joint angle characteristic matrix of the oil pipe transfer robot.
As a preferred embodiment of the present invention, the step of constructing the joint topology map of the pipe-handling robot in step 1 may be specifically described as follows: defining a joint topological graph of the oil pipe transfer robot as follows: g = (V, E, a). V represents a set of joint nodes of the oil pipe transfer robot, and V = { V = { (V)1, V2,…, Vn}; n is the number of nodes of the joint node; e represents the set of edges where the joints are located; a represents the static adjacency matrix between joint nodes.
It should be noted that, by constructing the joint topological diagram of the oil pipe transfer robot, the definition of the characteristic information of each joint of the oil pipe transfer robot can be realized; and the related characteristic information can be defined as any motion information of the joint, such as joint angle, velocity, acceleration and the like according to requirements, and the aim of the related characteristic information is to finally realize error compensation by predicting the joint angle.
As another preferred embodiment of the present invention, the step of constructing the joint angle feature matrix of the pipe handling robot in step 1 may be specifically described as follows: defining a joint angle characteristic matrix of the oil pipe transfer robot as follows:(ii) a Wherein R is the diameter of the joint node; and p is the attribute feature number (namely the historical time sequence length of the training learning data set) of the joint node.
After the constructed joint topological graph G and the joint angle characteristic matrix theta are completed, a mapping function f is learned and generated by utilizing a graph convolution network, so that joint angle motion information in a certain period of time in the future can be predicted, and the specific relevant formula can refer to the following formula:
step 2: constructing a joint dynamic graph network model of the oil pipe transfer robot to obtain dynamic node characteristics; and inputting the dynamic node characteristics into graph convolution for dynamic filtering.
It should be noted that, since the influence of non-geometric errors such as the spatial clearance and friction of each joint of the pipe-handling robot is complicated, in order to ensure that the described result is clear and natural, a dynamic node feature is adopted here, so as to perform description in a dynamic angle (see fig. 2).
Specifically, as a preferred embodiment of the present invention, the step of the network model of the joint dynamic diagram of the oil pipe handling robot in step 2 can be specifically described as follows:
at each time step, the joint angle is connected in seriesTime t and hidden state Ht-1Obtaining dynamic node characteristics I as input of joint dynamic graph network modelt(ii) a Wherein the content of the first and second substances,;b is the block size, n is the number of nodes, d is the characteristic dimension, TtIs the articulation cycle time.
The step of inputting the dynamic node characteristics into the graph convolution in the step 2 for dynamic filtering can be specifically described as follows:
dynamic node characteristics ItThe input image is convolved, anThe output is:(ii) a Wherein the dynamic filtering tensor,A graph convolution operation is shown in a graph convolution operation,are the learned filter parameters.
It should be noted that, by performing a graph convolution operation on a predefined adjacency matrix a (including a static distance relationship between nodes), message passing on a dynamic node state can be realized, that is, the adjacency matrix is updated.
And step 3: and constructing a joint dynamic graph circulation network model of the oil pipe transfer robot.
In order to improve the accuracy of joint prediction, a joint dynamic graph circulation network model of the oil pipe transfer robot needs to be constructed; in the joint dynamic graph cyclic network model, as shown in fig. 3, when graph convolution is performed, a predefined distance-based static graph and a node attribute-based dynamic graph are combined, and weighted summation of graph convolution results of an input graph signal, the dynamic graph and the predefined static graph is performed on each graph convolution layer.
Specifically, as a preferred embodiment of the present invention, step 3 can be specifically described as:
the joint dynamic graph circulation network model of the oil pipe transfer robot constructed by adopting the k-hop algorithm is expressed as follows:
wherein HinInputs for each node in the oil pipe transfer robot,in order to control the input signal(s),in order to be the parameters of the dynamic map of the joint,for the pre-defined static map parameters of the joint,a degree matrix after node self-loop is introduced;
it is noted that, considering the bi-directional propagation of the gap error, one skilled in the art may also consider using a bipartite graph convolution, i.e. the output representation may be chosen as:。
and 4, step 4: and predicting the joint angle of a future sequence based on the joint angle of the historical sequence according to the joint dynamic graph circulation network model of the oil pipe transfer robot.
Specifically, as a preferred embodiment of the present invention, the step 4 can be specifically described as:
as shown in fig. 4, the dynamic adjacency matrix of each time step is synchronously generated according to the recursive operation of the DGCRN; and at the time t, the angle value of each joint is obtained according to the obtained angle valueGiving a training step length M; will be provided withToPredicting output sequences as training inputsTo。
It is noted that, since the joint dynamic graph loop network model described above mainly makes the predicted angle infinitely close to the actual angle of the joint, in order to prevent overfitting, and at the same time add the L2 regularization term, the specific definition loss function may be as follows:
wherein λ in the above formula is a regularization hyper-parameter. And (3) updating and training the weight threshold values of various learning parameters of the model by calculating a loss function between the predicted value and the actual value and adopting an adaptive moment estimation optimization algorithm Adam until the model meets the design requirements.
In addition, as a preferred embodiment, the method for compensating joint trajectory error of a pipe handling robot according to the present invention further includes step 5: and performing deviation correction on a predicted value obtained by the joint dynamic graph circulation network model of the oil pipe transfer robot.
It should be noted that, in consideration of the prediction deviation of the predicted value itself and the influence caused by time lag during the joint compensation process, in order to improve the accuracy of error compensation, after the training of the joint dynamic graph loop network model in step 4 is completed, during actual compensation, the predicted value of the joint dynamic graph loop network model may be further subjected to deviation correction, specifically, the correction formula may refer to:。
therefore, according to the joint trajectory error compensation method for the oil pipe transfer robot, provided by the invention, the joint angle is subjected to fine adjustment compensation by calculating and finally predicting the difference value between the joint angle and the target value, so that the tail end error is reduced.
The invention provides an oil pipe transfer robot joint track error compensation method which comprises the steps of constructing a joint topological graph and a joint angle characteristic matrix of an oil pipe transfer robot, constructing a joint dynamic graph network model of the oil pipe transfer robot, constructing a joint dynamic graph circulating network model of the oil pipe transfer robot, predicting joint angles of future sequences based on joint angles of historical sequences and the like. The joint trajectory error compensation method of the oil pipe transfer robot with the steps and characteristics carries out real-time compensation on joint trajectory errors of the oil pipe transfer robot by establishing a nonlinear dynamic space-time correlation model, reduces repeated motion of repair equipment, improves motion precision of a tail end execution mechanism, increases machining and repair efficiency of the equipment, and reduces operation cost of the equipment.
The above description is only for the specific embodiments of the present invention, but the scope of the present invention is not limited thereto, and any person skilled in the art can easily conceive of the changes or substitutions within the technical scope of the present invention, and all the changes or substitutions should be covered within the scope of the present invention. Therefore, the protection scope of the present invention shall be subject to the protection scope of the appended claims.
Claims (9)
1. A joint track error compensation method of an oil pipe transfer robot is characterized by comprising the following steps:
step 1: constructing a joint topological graph and a joint angle characteristic matrix of the oil pipe carrying robot;
step 2: constructing a joint dynamic graph network model of the oil pipe transfer robot to obtain dynamic node characteristics; inputting the dynamic node characteristics into graph convolution for dynamic filtering;
and step 3: constructing a joint dynamic graph circulation network model of the oil pipe transfer robot;
and 4, step 4: and predicting the joint angle of a future sequence based on the joint angle of the historical sequence according to the joint dynamic graph circulation network model of the oil pipe transfer robot.
2. The method for compensating for joint trajectory error of a pipe handling robot of claim 1, further comprising the steps of:
and 5: and performing deviation correction on a predicted value obtained by the joint dynamic graph circulation network model of the oil pipe transfer robot.
4. the method for compensating for joint trajectory error of a tubing handling robot according to claim 1, wherein the step of constructing the joint topology map of the tubing handling robot in the step 1 can be specifically described as:
defining a joint topological graph of the oil pipe transfer robot as follows: g = (V, E, a); wherein V represents a set of joint nodes of the oil pipe transfer robot, and V = { V = { (V)1, V2,…, Vn}; n is the number of nodes of the joint node; e represents the set of edges where the joints are located; a represents the static adjacency matrix between joint nodes.
5. The method for compensating joint trajectory error of a tubing handling robot according to claim 1, wherein the step of constructing the joint angle feature matrix of the tubing handling robot in the step 1 can be specifically described as:
6. The method for compensating joint trajectory error of a tubing handling robot according to claim 1, wherein the step of the joint dynamic graph network model of the tubing handling robot in the step 2 is specifically described as follows:
at each time step, the tandem joint angle θtTime t and hidden state Ht-1Obtaining dynamic node characteristics I as input of joint dynamic graph network modelt(ii) a Wherein the content of the first and second substances,;b is the block size, n is the number of nodes, d is the characteristic dimension, TtIs the articulation cycle time.
7. The method for compensating for joint trajectory error of a pipe handling robot according to claim 1, wherein the step of inputting the dynamic node features into the graph convolution in the step 2 for dynamic filtering can be specifically described as:
8. The method for compensating for joint trajectory error of a pipe handling robot according to claim 1, wherein the step 3 can be specifically described as:
the joint dynamic graph circulation network model of the oil pipe transfer robot constructed by adopting the k-hop algorithm is expressed as follows:
wherein HinInputs for each node in the oil pipe transfer robot,in order to control the input signal(s),in order to be the parameters of the dynamic map of the joint,d ̃ is a degree matrix after the self-loop of the introduced nodes, which is a predefined static map parameter of the joint;
9. the method for compensating for joint trajectory error of a pipe handling robot according to claim 1, wherein the step 4 can be specifically described as:
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202210238378.XA CN114310917B (en) | 2022-03-11 | 2022-03-11 | Oil pipe transfer robot joint track error compensation method |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202210238378.XA CN114310917B (en) | 2022-03-11 | 2022-03-11 | Oil pipe transfer robot joint track error compensation method |
Publications (2)
Publication Number | Publication Date |
---|---|
CN114310917A true CN114310917A (en) | 2022-04-12 |
CN114310917B CN114310917B (en) | 2022-06-14 |
Family
ID=81033949
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN202210238378.XA Active CN114310917B (en) | 2022-03-11 | 2022-03-11 | Oil pipe transfer robot joint track error compensation method |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN114310917B (en) |
Citations (9)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JP2013248682A (en) * | 2012-05-30 | 2013-12-12 | Kobe Steel Ltd | Device and method for trajectory control of articulated robot |
US20150306761A1 (en) * | 2014-04-29 | 2015-10-29 | Brain Corporation | Trainable convolutional network apparatus and methods for operating a robotic vehicle |
CN108393892A (en) * | 2018-03-05 | 2018-08-14 | 厦门大学 | A kind of robot feedforward torque compensation method |
CN110202581A (en) * | 2019-06-28 | 2019-09-06 | 南京博蓝奇智能科技有限公司 | Compensation method, device and the electronic equipment of end effector of robot operating error |
CN110712201A (en) * | 2019-09-20 | 2020-01-21 | 同济大学 | Robot multi-joint self-adaptive compensation method based on perceptron model and stabilizer |
CN111300406A (en) * | 2020-01-17 | 2020-06-19 | 浙江理工大学 | Industrial robot track precision compensation system and method based on kinematic analysis |
CN112699771A (en) * | 2020-12-26 | 2021-04-23 | 南京理工大学 | Abnormal behavior detection algorithm based on human body posture prediction |
CN113361334A (en) * | 2021-05-18 | 2021-09-07 | 山东师范大学 | Convolutional pedestrian re-identification method and system based on key point optimization and multi-hop attention intention |
CN113705402A (en) * | 2021-08-18 | 2021-11-26 | 中国科学院自动化研究所 | Video behavior prediction method, system, electronic device and storage medium |
-
2022
- 2022-03-11 CN CN202210238378.XA patent/CN114310917B/en active Active
Patent Citations (9)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JP2013248682A (en) * | 2012-05-30 | 2013-12-12 | Kobe Steel Ltd | Device and method for trajectory control of articulated robot |
US20150306761A1 (en) * | 2014-04-29 | 2015-10-29 | Brain Corporation | Trainable convolutional network apparatus and methods for operating a robotic vehicle |
CN108393892A (en) * | 2018-03-05 | 2018-08-14 | 厦门大学 | A kind of robot feedforward torque compensation method |
CN110202581A (en) * | 2019-06-28 | 2019-09-06 | 南京博蓝奇智能科技有限公司 | Compensation method, device and the electronic equipment of end effector of robot operating error |
CN110712201A (en) * | 2019-09-20 | 2020-01-21 | 同济大学 | Robot multi-joint self-adaptive compensation method based on perceptron model and stabilizer |
CN111300406A (en) * | 2020-01-17 | 2020-06-19 | 浙江理工大学 | Industrial robot track precision compensation system and method based on kinematic analysis |
CN112699771A (en) * | 2020-12-26 | 2021-04-23 | 南京理工大学 | Abnormal behavior detection algorithm based on human body posture prediction |
CN113361334A (en) * | 2021-05-18 | 2021-09-07 | 山东师范大学 | Convolutional pedestrian re-identification method and system based on key point optimization and multi-hop attention intention |
CN113705402A (en) * | 2021-08-18 | 2021-11-26 | 中国科学院自动化研究所 | Video behavior prediction method, system, electronic device and storage medium |
Non-Patent Citations (2)
Title |
---|
XINZE ZHANG: "《Error-feedback stochastic modeling strategy for time series forecasting with convolutional neural networks 》", 《NEUROCOMPUTING》 * |
刘蓉: "《基于图卷积的动作识别方法研究》", 《中国优秀硕士学位论文全文数据库 信息科技辑》 * |
Also Published As
Publication number | Publication date |
---|---|
CN114310917B (en) | 2022-06-14 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
Pottmann et al. | A nonlinear predictive control strategy based on radial basis function models | |
Lu et al. | Generalized radial basis function neural network based on an improved dynamic particle swarm optimization and AdaBoost algorithm | |
Zio et al. | Failure and reliability predictions by infinite impulse response locally recurrent neural networks | |
Zaki et al. | Deep learning controller for nonlinear system based on Lyapunov stability criterion | |
Wang | Robot algorithm based on neural network and intelligent predictive control | |
CN110039537B (en) | Online self-learning multi-joint motion planning method based on neural network | |
CN111459031A (en) | Learning-oriented disturbance observer design method | |
CN117576379B (en) | Target detection method based on meta-learning combined attention mechanism network model | |
Hwang et al. | A fuzzy CMAC learning approach to image based visual servoing system | |
Liu et al. | Neural network control system of cooperative robot based on genetic algorithms | |
Chen et al. | Novel adaptive neural networks control with event-triggered for uncertain nonlinear system | |
Shi et al. | Neural network-based iterative learning control for trajectory tracking of unknown SISO nonlinear systems | |
CN114310917B (en) | Oil pipe transfer robot joint track error compensation method | |
Hosen et al. | NN-based prediction interval for nonlinear processes controller | |
Chen et al. | Echo state network with probabilistic regularization for time series prediction | |
CN115700414A (en) | Robot motion error compensation method | |
Ying et al. | Neural network nonlinear predictive control based on tent-map chaos optimization | |
CN114012733B (en) | Mechanical arm control method for scribing of PC component die | |
US20220414283A1 (en) | Predictive Modeling of Aircraft Dynamics | |
CN115762182A (en) | Vehicle intelligent track prediction method based on kinematic constraint | |
CN114943182A (en) | Robot cable shape control method and device based on graph neural network | |
CN110766144B (en) | Scalar coupling constant prediction system between atoms based on multi-layer decomposition fuzzy neural network | |
CN114779792A (en) | Medical robot autonomous obstacle avoidance method and system based on simulation and reinforcement learning | |
Liu et al. | Data learning‐based model‐free adaptive control and application to an NAO robot | |
CN113052297B (en) | Towing cable attitude calculation method and system based on convolution neural network fusion EKF |
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 |