CN115648228B - Industrial robot multi-source error compensation method, device, equipment and storage medium - Google Patents
Industrial robot multi-source error compensation method, device, equipment and storage medium Download PDFInfo
- Publication number
- CN115648228B CN115648228B CN202211689043.6A CN202211689043A CN115648228B CN 115648228 B CN115648228 B CN 115648228B CN 202211689043 A CN202211689043 A CN 202211689043A CN 115648228 B CN115648228 B CN 115648228B
- Authority
- CN
- China
- Prior art keywords
- error
- industrial robot
- model
- source
- positioning error
- 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.)
- Active
Links
Images
Classifications
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02P—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN THE PRODUCTION OR PROCESSING OF GOODS
- Y02P90/00—Enabling technologies with a potential contribution to greenhouse gas [GHG] emissions mitigation
- Y02P90/02—Total factory control, e.g. smart factories, flexible manufacturing systems [FMS] or integrated manufacturing systems [IMS]
Landscapes
- Numerical Control (AREA)
- Manipulator (AREA)
Abstract
The invention discloses a multi-source error compensation method, a device, equipment and a storage medium for an industrial robot, and relates to the technical field of industrial robots, wherein the method comprises the following steps: acquiring an initial working instruction, and correcting the error of the motion position of the initial working instruction according to a preset kinematic error model to acquire a target working instruction; acquiring pose information according to a target working instruction, inputting the pose information into a preset positioning error model to obtain positioning error prediction information related to the pose information, and establishing the positioning error model based on a migration learning process; and generating an error correction instruction according to the positioning error prediction information, and performing positioning error compensation correction on an end effector of the industrial robot based on the error correction instruction. The invention compensates the track error in the motion process of the industrial robot through the kinematic model, thereby improving the accuracy of error compensation.
Description
Technical Field
The present invention relates to the field of industrial robots, and in particular, to a method, an apparatus, a device, and a storage medium for multi-source error compensation of an industrial robot.
Background
Along with the intelligent development trend of manufacturing industry, industrial robots are widely applied in intelligent manufacturing, and the statues and shadows of the industrial robots are common in welding, polishing, drilling, carrying, spraying, polishing and other works. The application range of the industrial robot is expanded to the fields of aerospace and the like with extremely high requirements on working complexity and precision.
The positioning accuracy of the industrial robot is an important index for measuring the performance of the robot, however, factors influencing the accuracy of the industrial robot in the manufacturing, assembling and using processes are numerous, and it is particularly important to correspondingly compensate the motion error of the industrial robot. The existing compensation method often has the problem of low compensation precision.
Disclosure of Invention
The invention mainly aims to provide an industrial robot multi-source error compensation method, an industrial robot multi-source error compensation device, industrial robot multi-source error compensation equipment and a storage medium, and aims to solve the problem that an industrial robot error compensation method is low in compensation precision.
To achieve the above object, the present invention provides a multi-source error compensation method for an industrial robot, the method comprising:
acquiring an initial working instruction, and correcting the error of a motion position of the initial working instruction according to a preset kinematic error model to obtain a target working instruction;
acquiring pose information according to the target working instruction, inputting the pose information into a preset positioning error model to obtain positioning error prediction information associated with the pose information, wherein the positioning error model is established based on a transfer learning process;
and generating an error correction instruction according to the positioning error prediction information, and performing positioning error compensation correction on an end effector of the industrial robot based on the error correction instruction.
Optionally, the step of correcting the error of the motion position of the initial working instruction according to a preset kinematic error model includes:
extracting position information in the initial working instruction, and performing kinematic inversion processing on the position information to obtain nominal initial joint rotation information;
performing kinematic parameter error identification based on the kinematic error model to obtain a kinematic parameter error;
and fitting a position error according to the kinematic parameter error and the nominal initial joint rotation information in a Cartesian space, and correcting the position information according to the position error.
Optionally, the step of identifying the kinematic parameter error based on the kinematic error model to obtain the kinematic parameter error includes:
constructing a jacobian matrix according to the kinematic parameters of each connecting rod of the industrial robot;
determining a kinematic parameter error matrix of each connecting rod according to the jacobian matrix and the assumed pose errors;
and carrying out loop iteration update on the kinematic parameter errors based on the kinematic parameter error matrix, and obtaining the kinematic parameter errors after the loop iteration is ended.
Optionally, the step of acquiring pose information according to the target working instruction, inputting the pose information into a preset positioning error model, and obtaining positioning error prediction information associated with the pose information includes:
driving the industrial robot to reach a target pose state according to the target working instruction, and collecting pose information under the target pose state;
and carrying out Gaussian regression analysis on the pose information in a working space through the positioning error model to obtain positioning error information of a region corresponding to the pose information.
Optionally, the step of performing positioning error compensation correction on the end effector of the industrial robot based on the error correction instruction includes:
identifying the type of the current motion trail operation task;
if the current motion trail operation task is a discrete point position operation task, inserting the error correction instruction into the target operation instruction to compensate the positioning error of the end effector.
Optionally, the industrial robot multi-source error compensation method further includes:
constructing a training working space of the industrial robot, and dividing the training working space into a source domain and a target domain;
sampling in the source domain and the target domain respectively to construct a data set;
and migrating the tag data in the source domain to the target domain, carrying out data fusion on the data set, and establishing a source domain positioning error prediction model.
Optionally, after the establishing the source domain prediction model, the method further includes:
establishing a model error of the source domain positioning error prediction model based on Gaussian process regression;
calculating a data weighted fitting error of the target domain according to the model error;
and when the data weighted fitting error exceeds a preset error threshold, a target domain positioning error prediction model is obtained, and the source domain positioning error prediction model and the target domain positioning error prediction model are integrated into the positioning error model.
In addition, in order to achieve the above object, the present invention also provides an industrial robot multi-source error compensation apparatus comprising:
the acquisition module is used for acquiring an initial working instruction, and correcting the error of the motion position of the initial working instruction according to a preset kinematic error model to obtain a target working instruction;
the acquisition module is used for acquiring pose information according to the target working instruction, inputting the pose information into a preset positioning error model to obtain positioning error prediction information related to the pose information, and the positioning error model is built based on a transfer learning process;
and the correction module is used for generating an error correction instruction according to the positioning error prediction information and carrying out positioning error compensation correction on the end effector of the industrial robot based on the error correction instruction.
In addition, to achieve the above object, the present invention also provides an electronic device including: a memory, a processor, and an industrial robot multi-source error compensation program stored on the memory and executable on the processor, the industrial robot multi-source error compensation program configured to implement the steps of the industrial robot multi-source error compensation method as described above.
In addition, in order to achieve the above object, the present invention also provides a computer-readable storage medium having stored thereon an industrial robot multi-source error compensation program which, when executed by a processor, implements the steps of the industrial robot multi-source error compensation method as described above.
According to the industrial robot multi-source error compensation method, device, equipment and storage medium, an initial working instruction is obtained, error correction of a motion position is carried out on the initial working instruction according to the preset kinematic error model, a target working instruction is obtained, pose information is collected according to the target working instruction, the pose information is input into the preset positioning error model, positioning error prediction information related to the pose information is obtained, the positioning error model is built based on a migration learning process, an error correction instruction is generated according to the positioning error prediction information, positioning error compensation correction is carried out on an end effector of the industrial robot based on the error correction instruction, track errors in the motion process of the industrial robot are compensated through the kinematic model, terminal positioning accuracy is further compensated through the positioning error model, errors from different sources are respectively compensated in a hierarchical compensation mode, and error compensation accuracy is improved.
Drawings
FIG. 1 is a schematic diagram of an electronic device in a hardware operating environment according to an embodiment of the present invention;
FIG. 2 is a flow chart of a first embodiment of the multi-source error compensation method of the industrial robot according to the present invention;
FIG. 3 is a flow chart of a multi-source error compensation method for an industrial robot according to a second embodiment of the present invention;
FIG. 4a is a schematic diagram of a column working space involved in the multi-source error compensation method of the industrial robot of the present invention;
FIG. 4b is a schematic diagram of a cubic workspace involved in the industrial robot multi-source error compensation method of the present invention;
FIG. 5 is a schematic diagram of an industrial robot multi-source error compensation device according to the present invention.
The achievement of the objects, functional features and advantages of the present invention will be further described with reference to the accompanying drawings, in conjunction with the embodiments.
Detailed Description
It should be understood that the specific embodiments described herein are for purposes of illustration only and are not intended to limit the scope of the invention.
The positioning accuracy of the industrial robot can be divided into repeated positioning accuracy and absolute positioning accuracy, the repeated positioning accuracy of the conventional industrial robot can reach below 0.1mm, and the absolute positioning accuracy is usually 1-3mm, so that the requirement of precision machining cannot be met. The absolute positioning accuracy of the industrial robot can be improved through error compensation.
The main technical scheme of the invention is as follows: acquiring an initial working instruction, and correcting the error of a motion position of the initial working instruction according to a preset kinematic error model to obtain a target working instruction; acquiring pose information according to the target working instruction, inputting the pose information into a preset positioning error model to obtain positioning error prediction information associated with the pose information, wherein the positioning error model is established based on a transfer learning process; and generating an error correction instruction according to the positioning error prediction information, and performing positioning error compensation correction on an end effector of the industrial robot based on the error correction instruction.
Referring to fig. 1, fig. 1 is a schematic structural diagram of an electronic device in a hardware running environment according to an embodiment of the present invention.
As shown in fig. 1, the electronic device may include: a processor 1001, such as a central processing unit (Central Processing Unit, CPU), a communication bus 1002, a user interface 1003, a network interface 1004, a memory 1005. Wherein the communication bus 1002 is used to enable connected communication between these components. The user interface 1003 may include a Display, an input unit such as a Keyboard (Keyboard), and the optional user interface 1003 may further include a standard wired interface, a wireless interface. The network interface 1004 may optionally include a standard wired interface, a WIreless interface (e.g., a WIreless-FIdelity (WI-FI) interface). The Memory 1005 may be a high-speed random access Memory (Random Access Memory, RAM) Memory or a stable nonvolatile Memory (NVM), such as a disk Memory. The memory 1005 may also optionally be a storage device separate from the processor 1001 described above.
Those skilled in the art will appreciate that the structure shown in fig. 1 is not limiting of the electronic device and may include more or fewer components than shown, or may combine certain components, or may be arranged in different components.
As shown in fig. 1, an operating system, a network communication module, a user interface module, and an industrial robot multi-source error compensation program may be included in the memory 1005 as one type of storage medium.
In the electronic device shown in fig. 1, the network interface 1004 is mainly used for data communication with other devices; the user interface 1003 is mainly used for data interaction with a user; the processor 1001 and the memory 1005 in the electronic device of the present invention may be provided in the electronic device, where the electronic device invokes the industrial robot multi-source error compensation program stored in the memory 1005 through the processor 1001, and executes the industrial robot multi-source error compensation method provided by the embodiment of the present invention.
The embodiment of the invention provides an industrial robot multi-source error compensation method, and referring to fig. 2, fig. 2 is a flow chart of a first embodiment of the industrial robot multi-source error compensation method.
In this embodiment, the industrial robot multi-source error compensation method includes:
step S10, an initial working instruction is obtained, and error correction of a motion position is carried out on the initial working instruction according to a preset kinematic error model, so that a target working instruction is obtained;
the body structure of the industrial robot has certain errors in the manufacturing and assembling processes, so that the actual motion parameters and the nominal parameters in the controller have deviation, and the actual pose and the theoretical pose of the industrial robot are error due to the transmission and amplification effects of the connecting rod. The error compensation mode of the industrial robot can be divided into online real-time compensation and offline calibration compensation. The online compensation mode depends on an external measurement system, and can directly measure an error source to correct the pose of the industrial robot, but has extremely high development cost and is difficult to apply to field processing scenes. Off-line calibration compensation is often more common.
The industrial robot may move in multiple degrees of freedom directions. The initial work order may be regarded as an order for controlling the industrial robot to perform a corresponding operation. And the established kinematic error model is used for identifying kinematic parameters, the track error corresponding to the initial working instruction is compensated, and the working accuracy of working according to the obtained target working instruction is higher.
In some possible embodiments, the step of correcting the error of the motion position of the initial working instruction according to a preset kinematic error model may include:
step a, extracting position information in the initial working instruction, and performing kinematic inversion and solution processing on the position information to obtain nominal initial joint rotation information;
step b, identifying the kinematic parameter errors based on the kinematic error model to obtain the kinematic parameter errors;
and c, fitting a position error according to the kinematic parameter error and the nominal initial joint rotation information in a Cartesian space, and correcting the position information according to the position error.
The kinematic parameter may be at least one of a link length, a link offset, a joint torsion angle, and a joint rotation angle. The kinematic parameters are key to accurately controlling the no-load motion of the robot. The industrial robot can move to corresponding space coordinates and poses according to the position information in the initial working instruction. In the non-geometric error source of the industrial robot, the positioning error caused by the dead weight of the robot structure is relatively large, and the positioning error is mainly represented by elastic deformation of a connecting rod and torsional deformation of a joint. In the kinematic error model, it is assumed that the links of the industrial robot are rigid bodies, and the joint rotation angle offset due to the dead weight is a main factor.
The kinematic error compensation in the joint space often needs to derive an inverse kinematic solving process, the calculating process is complex, and the kinematic error model can be used for carrying out the error identification of the kinematic parameters in the Cartesian space to obtain the kinematic parameter errors. The kinematic parameter error can be further used for correcting the kinematic parameter to obtain the corrected kinematic parameter. The industrial robot can reach a designated position or a designated pose by controlling the torsion of the joints and the movement of the connecting rod, and the position information can comprise the movement information of the joints and the movement information of the connecting rod. For the position information in the initial working instruction, the initial joint rotation angle of the name of the industrial robot can be solved through kinematic inverse solution, a jacobian matrix is built by using the nominal rotation angle, the position error of the industrial robot in Cartesian space is obtained by using the corrected kinematic parameters, and the position information is corrected by using the position error, so that the offline correction of the initial working instruction is realized.
In some possible embodiments, the step of identifying the kinematic parameter error based on the kinematic error model may include:
step b1, constructing a jacobian matrix according to the kinematic parameters of each connecting rod of the industrial robot;
step b2, determining a kinematic parameter error matrix of each connecting rod according to the jacobian matrix and the assumed pose errors;
and b3, carrying out loop iteration update on the kinematic parameter errors based on the kinematic parameter error matrix, and ending the loop iteration to obtain the kinematic parameter errors.
Using q k Representing the kinematic parameters of each link of the industrial robot, J (q k ) Representing jacobian matrices, the kinematic parameter error matrices for each link can be represented as formula 1 below:
equation 1:
。
wherein, the liquid crystal display device comprises a liquid crystal display device,
representing the kinematic parameter error matrix of each link,
representing the assumed pose error of the person,
represents the damping coefficient and I represents the connecting rod parameter.
Using
Updating kinematic parameter errors, repeating the above-mentioned overlappingAnd (3) in the process of substitution, when the difference value of the assumed pose errors of two adjacent iterations approaches zero, the convergence can be considered, or the cyclic iteration process is ended when the iteration times reach a preset iteration times threshold value. The process of identifying the kinematic parameter errors is similar to the process of calibrating the kinematic parameters, and can be converted into a process of solving the linear relation between the assumed pose errors and the kinematic error matrix. The assumed pose error can be corresponding to the kinematic parameters of the connecting rod, and the kinematic parameters can be input into a kinematic error model for different connecting rods of the industrial robot to obtain the assumed pose error, wherein the assumed pose error belongs to the kinematic parameter error.
Step S20, acquiring pose information according to the target working instruction, inputting the pose information into a preset positioning error model to obtain positioning error prediction information associated with the pose information, wherein the positioning error model is established based on a transfer learning process;
the pose information can comprise a space coordinate value and a pose angle, and can be acquired under different pose conditions in the working process of the industrial robot. The industrial robot has error similarity in the working space, and the positioning error model can be constructed based on error distribution conditions of the error similarity. The positioning error model can predict the positioning error of the corresponding pose sampling point.
In some possible embodiments, the step of acquiring pose information according to the target working instruction, inputting the pose information into a preset positioning error model, and obtaining positioning error prediction information associated with the pose information may include:
step d, driving the industrial robot to reach a target pose state according to the target working instruction, and collecting pose information under the target pose state;
and e, carrying out Gaussian regression analysis on the pose information in a working space through the positioning error model to obtain positioning error prediction information of a region corresponding to the pose information.
The industrial robot can reach the corresponding position according to the target working instruction, and then the laser tracker is used for acquiring the pose information of the industrial robot at the corresponding position under the target pose state. In the case where the number of samples obtained by sampling is small, gaussian regression analysis can be used.
And step S30, generating an error correction instruction according to the positioning error prediction information, and carrying out positioning error compensation correction on an end effector of the industrial robot based on the error correction instruction.
The positioning error value in the positioning error prediction information can be used as a correction value to compensate the position information in the target working instruction, the spatial coordinate value and the attitude parameter of the industrial robot are adjusted, the compensation of the positioning error is completed, and the positioning precision of the industrial robot is improved.
In some possible embodiments, the step of performing positioning error compensation correction on the end effector of the industrial robot based on the error correction instructions may include:
f, identifying the type of the current motion trail operation task;
and g, if the current motion trail operation task is a discrete point position operation task, inserting the error correction instruction into the target operation instruction to compensate the positioning error of the end effector.
The positioning error prediction information can comprise position coordinate error prediction information and attitude angle error prediction information, the positioning error prediction information can be used for correcting a corresponding position information item in the target working instruction, and the corrected value is used for replacing an original value to generate an error correction instruction.
The operation task refers to a task which needs to be completed in the working process of the industrial robot, and the types of motion tracks contained in different operation tasks can be different, such as a discrete point position operation task and a continuous motion operation task. For job tasks of different track types, different compensation methods may be used for compensation. For a discrete point position operation task, the industrial robot is in a non-driving state after reaching any discrete target position, and an error correction instruction can be inserted after the robot is driven according to the instruction to compensate positioning errors at the current position.
For continuous motion work tasks, the motion consistency of the industrial robot is affected due to the compensation hysteresis problem of the discrete point compensation mode. The motion trail can be approximately regarded as a set of linear motions between interpolation point positions, and is decomposed into linear motion trail by adopting a mode of segment compensation and dislocation correction. For example, the positioning error prediction information for the i-th target pose is applied in the i+2-th target pose, the positioning error prediction information for the i+1-th target pose is applied in the i+3-th target pose, and so on.
In this embodiment, an initial working instruction is acquired, motion error correction is performed on the initial working instruction according to a preset kinematic error model, a target working instruction is obtained, pose information is acquired according to the target working instruction, the pose information is input into a preset positioning error model, positioning error prediction information is obtained, positioning error compensation correction is performed on an end effector of an industrial robot according to the positioning error prediction information, track errors in the motion process of the industrial robot are compensated through the kinematic model, terminal positioning accuracy is compensated by combining the positioning error model, and errors from different sources are respectively compensated in a hierarchical compensation mode, so that the error compensation accuracy is improved.
Further, in a second embodiment of the industrial robot multi-source error compensation method of the present invention, as shown in fig. 3, the method includes:
step S40, constructing a training working space of the industrial robot, and dividing the training working space into a source domain and a target domain;
the positioning error model can be trained by using a migration learning mode. The transfer learning is a machine learning method applied to new problems according to the learned knowledge transfer, and can use the marked sample of the source domain to assist the target domain to construct a prediction model under the condition that the marked sample of the source domain is sufficient and the data quantity of the marked sample of the target domain is small. Industrial robots have error similarities in training workspaces, providing a basis for application of transfer learning.
The training workspace may be a cylinder workspace or a cube workspace. Fig. 4a is a schematic view of a column working space in which zones may be performed along joint axis corners, zone 1, zone 2 and zone 3 representing different zones respectively. Fig. 4b is a schematic view of a cubic working space in which partitions can be performed along the direction of the spatial coordinate axis, and regions 1, 2 and 3 represent different partitions, respectively. The partition intervals may be uniform intervals. The motion error distribution rule of the industrial robot can be analyzed in the cylinder working space or the cube working space, for example, one partition is selected as a reference area, the other partitions are selected as comparison areas, and the error distribution similarity of the comparison areas and the reference area is obtained through comparison.
Step S50, sampling is carried out in the source domain and the target domain respectively, and a data set is constructed;
in the embodiment, a positioning error model is constructed by taking a cylinder working space as an example, an axis corner of an end joint of an industrial robot is used as a discrete value for partitioning, the cylinder working space is divided into a source domain and a target domain space, a complete industrial robot pose and error data set is constructed in the source domain by adopting a hierarchical sampling method, and a small amount of industrial robot pose and error data set is constructed in the target domain based on migration analysis. The construction of the target domain dataset is related to the error distribution characteristics. On the basis of guaranteeing the pose and error sample diversity of the industrial robot, the measurement efficiency and the measurement precision are considered, and the similarity of the source domain and the target domain can be improved by designing proper source domain size and position and optimizing the sampling point distribution of the source domain and the working domain. The sampling interval is an important factor influencing the model prediction precision and the source domain space size, when the space point interval is larger, the error change of adjacent pose points is larger, and the error change of the industrial robot is difficult to describe accurately. When the space point interval is smaller, the measurement points are too many, the difference value of the positioning errors between the adjacent poses is smaller, and the model is easy to be interfered by the measurement errors of the measuring instrument during training.
And step S60, migrating the label data in the source domain to the target domain, carrying out data fusion on the data set, and establishing a source domain positioning error prediction model.
The industrial robot has similarity in error distribution in different areas, a target domain space can be divided into k subspaces according to the space size of a source domain, the characteristic spaces of the ith subspace of the source domain and the target domain are respectively aligned in a subspace alignment mode, a source domain label dataset is aligned with the label space of a designated subspace on the basis, a migration dataset of the ith subspace is formed, the k subspace migration dataset and the target domain data are fused to form a target domain training dataset, and a source domain positioning error prediction model is obtained through training. The tag data refers to data obtained by sampling the industrial robot in space, and may include spatial position data and pose data, that is, represent the position and pose of the industrial robot at the sampling point. In the process of dividing the subspaces, the overlapping rate between the adjacent subspaces can be determined according to actual requirements. The source domain feature space and the target domain feature space can be converted by a conversion matrix.
In some possible embodiments, after the step of establishing the source domain positioning error prediction model, the method may further include:
step i, establishing a model error of the source domain positioning error prediction model based on Gaussian process regression;
step j, calculating a data weighted fitting error of the target domain according to the model error;
and k, when the data weighted fitting error exceeds a preset error threshold, obtaining a target domain positioning error prediction model, and integrating the source domain positioning error prediction model and the target domain positioning error prediction model into the positioning error model.
For a source domain positioning error prediction model, an initial weight matrix and a weighted fitting error initial value of the source domain positioning error prediction model can be determined, a model error is established based on GPR (Gaussian Process Regression ), a target domain data weighted fitting error is calculated, iteration is stopped when the iterative model error exceeds an error threshold value, and the weight matrix is updated according to the weighted fitting error, so that the target domain positioning error prediction model is obtained. Gaussian Process regression is a non-parametric model that uses Gaussian Process (GP) priors to perform regression analysis on data, with versatility and resolvable. And combining the source domain positioning error prediction model and the target domain positioning error prediction model to obtain the positioning error model of the industrial robot in the whole working space. The combination of the source domain positioning error prediction model and the target domain positioning error prediction model refers to an integration process under the same feature space, and the source domain or the target domain is converted into the same feature space through a conversion matrix, so that corresponding training data can be integrated into the same representation mode in the same feature space. The weight matrix can be determined according to the measurement accuracy and measurement error of the laser tracker used for measurement.
In the embodiment, the positioning error model is trained by adopting a transfer learning mode, so that the number of poses required in the sampling process can be reduced, and compared with a space grid compensation method, the positioning accuracy is ensured, and meanwhile, the compensation efficiency of the positioning error of the industrial robot is improved due to the reduction of the number of the sampling poses.
The embodiment of the present invention further provides an industrial robot multi-source error compensation device, as shown in fig. 5, which may be used to implement the steps of the industrial robot multi-source error compensation method described above, where the industrial robot multi-source error compensation device includes:
the acquisition module 101 is configured to acquire an initial working instruction, correct an error of a motion position of the initial working instruction according to a preset kinematic error model, and obtain a target working instruction;
the acquisition module 102 is configured to acquire pose information according to the target working instruction, input the pose information into a preset positioning error model, and obtain positioning error prediction information associated with the pose information, where the positioning error model is established based on a transfer learning process;
and the correction module 103 is used for generating an error correction instruction according to the positioning error prediction information and carrying out positioning error compensation correction on the end effector of the industrial robot based on the error correction instruction.
Optionally, the obtaining module 101 is further configured to:
extracting position information in the initial working instruction, and performing kinematic inversion processing on the position information to obtain nominal initial joint rotation information;
performing kinematic parameter error identification based on the kinematic error model to obtain a kinematic parameter error;
and fitting a position error according to the kinematic parameter error and the nominal initial joint rotation information in a Cartesian space, and correcting the position information according to the position error.
Optionally, the obtaining module 101 is further configured to:
constructing a jacobian matrix according to the kinematic parameters of each connecting rod of the industrial robot;
determining a kinematic parameter error matrix of each connecting rod according to the jacobian matrix and the assumed pose errors;
and carrying out loop iteration update on the kinematic parameter errors based on the kinematic parameter error matrix, and obtaining the kinematic parameter errors after the loop iteration is ended.
Optionally, the acquisition module 102 is further configured to:
driving the industrial robot to reach a target pose state according to the target working instruction, and collecting pose information under the target pose state;
and carrying out Gaussian regression analysis on the pose information in a working space through the positioning error model to obtain positioning error prediction information of a region corresponding to the pose information.
Optionally, the correction module 103 is further configured to:
identifying the type of the current motion trail operation task;
if the current motion trail operation task is a discrete point position operation task, inserting the error correction instruction into the target operation instruction to compensate the positioning error of the end effector.
Optionally, the industrial robot multi-source error compensation device further comprises a construction module for:
constructing a training working space of the industrial robot, and dividing the training working space into a source domain and a target domain;
sampling in the source domain and the target domain respectively to construct a data set;
and migrating the tag data in the source domain to the target domain, carrying out data fusion on the data set, and establishing a source domain positioning error prediction model.
Optionally, the building module is further configured to:
establishing a model error of the source domain positioning error prediction model based on Gaussian process regression;
calculating a data weighted fitting error of the target domain according to the model error;
and when the data weighted fitting error exceeds a preset error threshold, a target domain positioning error prediction model is obtained, and the source domain positioning error prediction model and the target domain positioning error prediction model are integrated into the positioning error model.
The embodiment of the invention also provides electronic equipment, which comprises: a memory, a processor, and an industrial robot multi-source error compensation program stored on the memory and executable on the processor, the industrial robot multi-source error compensation program configured to implement the steps of the industrial robot multi-source error compensation method as described above. The specific implementation manner of the electronic device in the embodiment of the present invention refers to each embodiment of the method for compensating multi-source error of the industrial robot, and is not described herein again.
The embodiment of the invention also provides a computer readable storage medium, wherein the computer readable storage medium is stored with an industrial robot multi-source error compensation program, and the industrial robot multi-source error compensation program realizes the steps of the industrial robot multi-source error compensation method when being executed by a processor. The specific implementation manner of the computer readable storage medium in the embodiment of the present invention refers to each embodiment of the above-mentioned multi-source error compensation method for an industrial robot, and is not described herein again.
It should be noted that, in this document, the terms "comprises," "comprising," or any other variation thereof, are intended to cover a non-exclusive inclusion, such that a process, method, article, or system that comprises a list of elements does not include only those elements but may include other elements not expressly listed or inherent to such process, method, article, or system. Without further limitation, an element defined by the phrase "comprising one … …" does not exclude the presence of other like elements in a process, method, article, or system that comprises the element.
The foregoing embodiment numbers of the present invention are merely for the purpose of description, and do not represent the advantages or disadvantages of the embodiments.
From the above description of the embodiments, it will be clear to those skilled in the art that the above-described embodiment method may be implemented by means of software plus a necessary general hardware platform, but of course may also be implemented by means of hardware, but in many cases the former is a preferred embodiment. Based on such understanding, the technical solution of the present invention may be embodied essentially or in a part contributing to the prior art in the form of a software product stored in a storage medium (e.g. ROM/RAM, magnetic disk, optical disk) as described above, comprising instructions for causing a terminal device (which may be a mobile phone, a computer, a server, or a network device, etc.) to perform the method according to the embodiments of the present invention.
The foregoing description is only of the preferred embodiments of the present invention, and is not intended to limit the scope of the invention, but rather is intended to cover any equivalents of the structures or equivalent processes disclosed herein or in the alternative, which may be employed directly or indirectly in other related arts.
Claims (9)
1. An industrial robot multi-source error compensation method is characterized by comprising the following steps:
acquiring an initial working instruction, and correcting the error of a motion position of the initial working instruction according to a preset kinematic error model to obtain a target working instruction;
acquiring pose information according to the target working instruction, inputting the pose information into a preset positioning error model to obtain positioning error prediction information associated with the pose information, wherein the positioning error model is established based on a transfer learning process;
generating an error correction instruction according to the positioning error prediction information, and performing positioning error compensation correction on an end effector of the industrial robot based on the error correction instruction;
the step of correcting the error of the motion position of the initial working instruction according to a preset kinematic error model comprises the following steps:
extracting position information in the initial working instruction, and performing kinematic inversion processing on the position information to obtain nominal initial joint rotation information;
performing kinematic parameter error identification based on the kinematic error model to obtain a kinematic parameter error;
and fitting a position error according to the kinematic parameter error and the nominal initial joint rotation information in a Cartesian space, and correcting the position information according to the position error.
2. The method for compensating for multi-source errors of an industrial robot according to claim 1, wherein the step of performing the identification of the kinematic parameter errors based on the kinematic error model, and obtaining the kinematic parameter errors comprises:
constructing a jacobian matrix according to the kinematic parameters of each connecting rod of the industrial robot;
determining a kinematic parameter error matrix of each connecting rod according to the jacobian matrix and the assumed pose errors;
and carrying out loop iteration update on the kinematic parameter errors based on the kinematic parameter error matrix, and obtaining the kinematic parameter errors after the loop iteration is ended.
3. The method for compensating for multi-source errors of an industrial robot according to claim 2, wherein the step of acquiring pose information according to the target work instruction, inputting the pose information into a preset positioning error model, and obtaining positioning error prediction information associated with the pose information comprises:
driving the industrial robot to reach a target pose state according to the target working instruction, and collecting pose information under the target pose state;
and carrying out Gaussian regression analysis on the pose information in a working space through the positioning error model to obtain positioning error prediction information of a region corresponding to the pose information.
4. The method of claim 3, wherein the step of performing positioning error compensation correction on the end effector of the industrial robot based on the error correction instruction comprises:
identifying the type of the current motion trail operation task;
if the current motion trail operation task is a discrete point position operation task, inserting the error correction instruction into the target operation instruction to compensate the positioning error of the end effector.
5. The industrial robot multi-source error compensation method of claim 4, further comprising:
constructing a training working space of the industrial robot, and dividing the training working space into a source domain and a target domain;
sampling in the source domain and the target domain respectively to construct a data set;
and migrating the tag data in the source domain to the target domain, carrying out data fusion on the data set, and establishing a source domain positioning error prediction model.
6. The method of claim 5, further comprising, after said establishing a source domain positioning error prediction model:
establishing a model error of the source domain positioning error prediction model based on Gaussian process regression;
calculating a data weighted fitting error of the target domain according to the model error;
and when the data weighted fitting error exceeds a preset error threshold, a target domain positioning error prediction model is obtained, and the source domain positioning error prediction model and the target domain positioning error prediction model are integrated into the positioning error model.
7. An industrial robot multi-source error compensation device, characterized in that the industrial robot multi-source error compensation device comprises:
the acquisition module is used for acquiring an initial working instruction, and correcting the error of the motion position of the initial working instruction according to a preset kinematic error model to obtain a target working instruction;
the acquisition module is used for acquiring pose information according to the target working instruction, inputting the pose information into a preset positioning error model to obtain positioning error prediction information related to the pose information, and the positioning error model is built based on a transfer learning process;
the correction module is used for generating an error correction instruction according to the positioning error prediction information and carrying out positioning error compensation correction on an end effector of the industrial robot based on the error correction instruction;
the acquisition module is also used for extracting the position information in the initial working instruction, and performing kinematic inversion and solution processing on the position information to obtain nominal initial joint rotation information;
performing kinematic parameter error identification based on the kinematic error model to obtain a kinematic parameter error;
and fitting a position error according to the kinematic parameter error and the nominal initial joint rotation information in a Cartesian space, and correcting the position information according to the position error.
8. An electronic device, the electronic device comprising: a memory, a processor and an industrial robot multi-source error compensation program stored on the memory and executable on the processor, the industrial robot multi-source error compensation program being configured to implement the steps of the industrial robot multi-source error compensation method of any one of claims 1 to 6.
9. A computer-readable storage medium, characterized in that the computer-readable storage medium has stored thereon an industrial robot multi-source error compensation program, which when executed by a processor, implements the steps of the industrial robot multi-source error compensation method according to any one of claims 1 to 6.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211689043.6A CN115648228B (en) | 2022-12-28 | 2022-12-28 | Industrial robot multi-source error compensation method, device, equipment and storage medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211689043.6A CN115648228B (en) | 2022-12-28 | 2022-12-28 | Industrial robot multi-source error compensation method, device, equipment and storage medium |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN115648228A CN115648228A (en) | 2023-01-31 |
| CN115648228B true CN115648228B (en) | 2023-05-26 |
Family
ID=85023227
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202211689043.6A Active CN115648228B (en) | 2022-12-28 | 2022-12-28 | Industrial robot multi-source error compensation method, device, equipment and storage medium |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN115648228B (en) |
Families Citing this family (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117340897B (en) * | 2023-12-05 | 2024-03-12 | 山东建筑大学 | Dynamic response prediction-oriented robot digital twin model modeling method and system |
| CN117688519A (en) * | 2024-02-04 | 2024-03-12 | 山东建筑大学 | Multi-degree-of-freedom series mechanism positioning error prediction method |
| CN117742239B (en) * | 2024-02-19 | 2024-05-14 | 南京超颖新能源科技有限公司 | Vertical correction system and correction method for machine tool |
Family Cites Families (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104535027B (en) * | 2014-12-18 | 2017-06-13 | 南京航空航天大学 | A kind of robot precision's compensation method of variable element error identification |
| CN108908327B (en) * | 2018-04-25 | 2021-09-07 | 成都飞机工业（集团）有限责任公司 | Robot positioning error grading compensation method |
| CN110385720B (en) * | 2019-07-26 | 2020-08-04 | 南京航空航天大学 | Robot positioning error compensation method based on deep neural network |
| CN110842927B (en) * | 2019-11-30 | 2022-07-08 | 天津大学 | Robot joint geometric error compensation method based on multiple regression |
| CN111590569A (en) * | 2020-05-14 | 2020-08-28 | 河北工业大学 | Industrial robot positioning error compensation method |
| CN111590581B (en) * | 2020-05-26 | 2021-10-22 | 珠海格力智能装备有限公司 | Positioning compensation method and device for robot |
| CN112536797A (en) * | 2020-12-01 | 2021-03-23 | 南京航空航天大学 | Comprehensive compensation method for position and attitude errors of industrial robot |
| CN114131611B (en) * | 2021-12-16 | 2023-10-24 | 华中科技大学 | Off-line compensation method, system and terminal for joint errors of robot gravity pose decomposition |
| CN115401699B (en) * | 2022-10-31 | 2023-03-24 | 广东隆崎机器人有限公司 | Industrial robot precision reliability analysis method, device, equipment and storage medium |
-
2022
- 2022-12-28 CN CN202211689043.6A patent/CN115648228B/en active Active
Also Published As
| Publication number | Publication date |
|---|---|
| CN115648228A (en) | 2023-01-31 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN115648228B (en) | Industrial robot multi-source error compensation method, device, equipment and storage medium | |
| CN108908327B (en) | Robot positioning error grading compensation method | |
| Marie et al. | Elasto-geometrical modeling and calibration of robot manipulators: Application to machining and forming applications | |
| Zhu et al. | Positioning error compensation on two-dimensional manifold for robotic machining | |
| CN115401699B (en) | Industrial robot precision reliability analysis method, device, equipment and storage medium | |
| Klimchik et al. | Geometric and elastostatic calibration of robotic manipulator using partial pose measurements | |
| JP3349652B2 (en) | Offline teaching method | |
| Hafezipour et al. | An uncertainty analysis method for error reduction in end-effector of spatial robots with joint clearances and link dimension deviations | |
| CN116038702B (en) | Seven-axis robot inverse solution method and seven-axis robot | |
| CN113910218A (en) | Robot calibration method and device based on kinematics and deep neural network fusion | |
| Hu et al. | Optimal synthesis of pose repeatability for collaborative robots based on the ISO 9283 standard | |
| Omodei et al. | Calibration of a measuring robot: Experimental results on a 5 DOF structure | |
| Wu et al. | Review of industrial robot stiffness identification and modelling | |
| Tian et al. | A general approach for robot pose error compensation based on an equivalent joint motion error model | |
| Li et al. | Model-based recurrent neural network for redundancy resolution of manipulator with remote centre of motion constraints | |
| Pott et al. | A simplified force-based method for the linearization and sensitivity analysis of complex manipulation systems | |
| US6785624B2 (en) | Method for calibrating of machine units located in machine tools or robotic devices | |
| CN115816458A (en) | Robot kinematics parameter and gravity integrated calibration and compensation method and device | |
| Wang et al. | An ANN‐Based Precision Compensation Method for Industrial Manipulators via Optimization of Point Selection | |
| EP2359989A1 (en) | Robot control with bootstrapping inverse kinematics | |
| Skopec et al. | Calibration using adaptive model complexity for parallel and fiber-driven mechanisms | |
| Dastangoo et al. | Non-linear Parameter Identification for Humanoid Robot Components | |
| Qi et al. | Trajectory evaluation for manipulators with motion and sensor uncertainties | |
| Tian et al. | A dual-six-dimensional force sensors calibration method for pHRI based on ridge regression | |
| Solanes et al. | Generalization of reference filtering control strategy for 2D/3D visual feedback control of industrial robot manipulators |
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 |