Method and apparatus for geometric error modeling and identification of hybrid robots characterized by axial displacement screw force

By constructing a realistic kinematic model of the hybrid robot using axial displacement screws and eliminating redundant parameters, and combining it with the nonlinear least squares method, the problems of completeness and minimization of geometric error modeling of the hybrid robot were solved, and high-precision error compensation was achieved.

CN123103655APending Publication Date: 2026-11-17TSINGHUA UNIVERSITY
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202611387129.1
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2026-09-08
Publication Date
2026-11-17

AI Technical Summary

Technical Problem

Existing geometric error modeling methods for hybrid robots cannot simultaneously satisfy completeness and minimization, leading to unstable error parameter identification and affecting the accuracy of the end effector.

Method used

A realistic kinematic model is constructed using the screw of axial displacement. Through error embedding and redundant parameter reduction, combined with the identification equation in the form of pose residual screw and nonlinear least squares method, the stable identification and compensation of geometric error parameters are achieved.

Benefits of technology

A geometric error model satisfying completeness and minimization was established, which improved the accuracy compensation effect of the end-effector pose of the hybrid robot and enhanced the stability and anti-interference ability of error parameter identification.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN123103655A_ABST
    Figure CN123103655A_ABST
Patent Text Reader

Abstract

This invention relates to the field of robot calibration technology, and particularly to a method and apparatus for modeling and identifying geometric errors in hybrid robots using axial displacement screw representation. The method includes: constructing a complete axial displacement screw representing the relative pose relationship between joint axes, and establishing a true kinematic model of the hybrid robot; embedding error parameters into the true kinematic model of the hybrid robot to construct a geometric error model that can represent any non-ideal geometric relationship without redundancy; measuring several poses of the hybrid robot in the workspace and constructing identification equations in the form of pose residual screws, reducing low-sensitivity error parameters in the parallel section and differential screws of passive joint motions in each branch, establishing the Jacobian matrix of pose residuals with respect to geometric errors, and identifying geometric error parameters using a nonlinear least squares algorithm; embedding the identified error parameters into the geometric error model to achieve error compensation, and ensuring that the constructed geometric error model meets the requirements of minimization and completeness.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] This invention relates to the field of robot calibration technology, and in particular to a method and apparatus for modeling and identifying geometric errors of hybrid robots by characterizing axial displacement screws. Background Technology

[0002] Hybrid robots combine the advantages of serial and parallel robots, offering a large workspace, high flexibility, and good rigidity and dynamic response capabilities. In recent years, they have been widely used in high-end manufacturing fields such as aerospace, automotive manufacturing, and energy equipment, and have been successfully deployed in various industrial tasks including machining and precision assembly. In precision machining scenarios, the absolute pose accuracy of the end effector is the core factor restricting the actual machining accuracy of hybrid robots. However, geometric errors are inevitably introduced during the manufacturing and assembly of mechanical parts. These errors accumulate and propagate step-by-step in the complex kinematic chain of the hybrid mechanism, ultimately causing a systematic deviation between the actual pose of the end effector and the commanded pose. This deviation often exceeds the tolerance range allowed by precision machining. Therefore, research on error compensation for hybrid robots is of great significance for effectively improving their precision machining performance.

[0003] Geometric error models are a crucial factor affecting the accuracy of error compensation. A geometric error model describes the mapping relationship between geometric errors and end effector pose deviations. Ideally, the error model should possess both completeness and minimization. Completeness requires that the parameters in the model can fully characterize any geometric deviation, reflecting the model's ability to cover geometric errors. A model containing a complete set of geometric error parameters can more adequately compensate for end effector pose deviations, thus achieving a higher accuracy ceiling. Minimization requires that the error parameters are free of redundancy, and its local criterion is that the column vectors of the corresponding Jacobian matrix are linearly independent, i.e., full column rank, which is a necessary condition for achieving stable identification. Eliminating redundant parameters helps obtain locally unique identification results and improves identification stability and anti-interference capabilities. Constructing an error model that simultaneously satisfies these two properties is not only an important goal in the field of robot calibration but also a key approach to achieving high-quality error compensation from the modeling level.

[0004] However, existing modeling methods still have shortcomings in simultaneously satisfying completeness and minimality. Methods based on DH parameters impose strict constraints on the definition of the local coordinate system, inherently lacking completeness, and can introduce parameter singularities when specific geometric relationships exist between adjacent joint axes. While CPC models and traditional POE models can meet completeness requirements, their error parameters exhibit linear correlation, weakening the stability of the identification process. Closed-loop vector methods cannot extract angular error parameters at closed loops, and some angular error components are easily missed during perturbation derivation, leading to insufficient model completeness. While screw theory methods can comprehensively describe various errors between adjacent joint axes, they still contain redundant parameters and are difficult to naturally embed into the kinematic model, thus limiting the feasibility of achieving high-precision compensation through iterative error parameter updates. Furthermore, the multi-closed-loop structure in hybrid robots introduces low-sensitivity errors, causing the identified Jacobian matrix to approach singularity, thereby affecting the stability of the identification algorithm.

[0005] In summary, how to accurately model the geometric errors of hybrid robots that meet the requirements of minimization and completeness, and stably identify the error parameters, remains a bottleneck problem that urgently needs to be solved in the field. Summary of the Invention

[0006] This invention provides a method and apparatus for modeling and identifying geometric errors in hybrid robots using axial displacement screws. The method utilizes axial displacement screws to establish a generalized and standardized real kinematic model. Through error embedding and redundant parameter reduction, a geometric error model that simultaneously satisfies completeness and minimization is obtained. Combining the identification equation in the form of pose residual screws with the nonlinear least squares method, the geometric error parameters are stably identified. Finally, the identification results are embedded into the inverse kinematics solution to achieve high-precision and stable error compensation.

[0007] A first aspect of the present invention provides a method for modeling and identifying geometric errors of a hybrid robot by characterizing the spinor of axial displacement, comprising the following steps: Based on the Lie algebra subspace, an axial displacement screw that can completely characterize the joint relationships of the target hybrid robot is constructed to establish a real kinematic model, and the parameterization form of the actual geometric parameters is determined according to the real kinematic model. The actual geometric parameters are represented as a superposition of nominal geometric parameters and error parameters to construct a geometric error model that can represent any non-ideal geometric relationship without redundancy. The geometric error model is then explicitly decomposed into an error-pose mapping model and an error-drive mapping model. To reduce the low-sensitivity error parameters of the parallel branch of the target hybrid robot, the measured pose data of the robot end effector in the workspace is obtained, and the measured pose data of the robot end effector is combined with the error-pose mapping model to construct a geometric error parameter identification equation based on the pose residual spinor. Eliminate the ineffective differential spins of passive joints in the parallel chain of the target hybrid robot, and extract the identification Jacobian matrix of the hybrid robot pose residual relative to the geometric error parameters from the geometric error parameter identification equation; Based on the nonlinear least squares algorithm, the geometric error parameter estimates are iteratively solved using the identified Jacobian matrix, and then input into the error-drive mapping model to compensate for the end-effector pose error of the hybrid robot.

[0008] Optionally, the target hybrid robot includes a series section and a parallel section connected in sequence, wherein, The series connection is in the form of an open chain mechanism. Each joint in the open chain mechanism is classified into a rotary joint or a traversing joint according to the type of motion. Each joint in the open chain mechanism is an active joint. The parallel part is in the form of a closed-loop mechanism. Each joint in the closed-loop mechanism is classified into a rotary joint or a traverse joint according to the type of motion. Each joint in the open-loop mechanism is classified into an active joint or a passive joint according to the driving method.

[0009] Optionally, the construction of axial displacement screws based on Lie algebra subspaces to fully characterize the joint relationships of the target hybrid robot, in order to establish a true kinematic model, and the determination of the parameterized form of the actual geometric parameters based on the true kinematic model, includes: Extracting Lie Algebra In a four-dimensional subspace, given spinor coordinates and a selected range of rotation angles, the elements within the four-dimensional subspace can locally and uniquely represent the relationship between the helical axis and the coordinate system. z Arbitrary helical motion with axes perpendicularly intersecting; All elements within the four-dimensional subspace are represented using six-dimensional coordinates and defined as axial displacement spinors; The same screw exponential product modeling method is used for the serial and parallel parts of the target hybrid robot to establish the kinematic equations of the serial part and the kinematic equations of each branch of the parallel part. The real kinematic model is constructed based on the kinematic equations of the series part and the kinematic equations of each branch of the parallel part, and the parameterized form of the actual geometric parameters is determined based on the real kinematic model.

[0010] Optionally, the step of performing the spinor exponential product method on the serial portion of the target hybrid robot to establish the kinematic equations of the serial portion includes: Assume the serial part of the target hybrid robot is composed of Composed of several components, the initial configuration of the series section is known, and a base coordinate system 0 is defined, wherein, It is a positive integer; Let coordinate system of z axis and corresponding joint axis The coordinate system is determined by the exponential product of the rotational displacement of the corresponding joint axis and the rotational displacement of the joint motion. With coordinate system The homogeneous transformation matrix between them, wherein the joint motion spinor is defined in six-dimensional coordinate form, representing the rotation around the joint. z Rotation of the shaft or along z Axial translation element, ; Let coordinate system The pose of the end of the series section is represented by the coordinate system determined by the exponential mapping of the six-dimensional displacement spinor. To coordinate system The homogeneous transformation matrix, wherein the six-dimensional displacement spinor is defined as any... in six-dimensional coordinate form. element; Multiply the homogeneous transformation matrices between all adjacent coordinate systems in sequence to obtain the kinematic equations of the cascaded part.

[0011] Optionally, the step of characterizing the actual geometric parameters as a superposition of nominal geometric parameters and error parameters to construct a geometric error model that can characterize any non-ideal geometric relationship without redundancy, and explicitly decomposing the geometric error model into an error-pose mapping model and an error-drive mapping model, includes: The actual geometric parameters are represented as a superposition of nominal geometric parameters and error parameters; Based on the superposition of the nominal geometric parameters and error parameters, the axial displacement screw of each moving joint is synchronously expressed as the sum of the nominal axial displacement screw and the axial displacement screw error; The movement component in the axial displacement screw error of each moving joint is removed to obtain a geometric error model that can characterize any non-ideal geometric relationship without redundancy. The geometric error model is explicitly decomposed into the error-pose mapping model and the error-driven mapping model.

[0012] Optionally, the process of reducing the low-sensitivity error parameters of the parallel branches of the target hybrid robot, obtaining the measured pose data of the robot's end effector within the workspace, and combining the measured pose data of the robot's end effector with the error-pose mapping model to construct a geometric error parameter identification equation based on the pose residual spinor includes: Under the conditions of the nominal geometric parameters, the power screw, constraint screw, and differential screw corresponding to each error parameter of the parallel chain of the target hybrid robot are solved. The nominal geometric parameter condition means that the screw error of all axis displacements is zero. The differential screw corresponding to each error parameter is used to characterize the helical motion caused by small changes in the error parameters. Perform a reciprocal product operation between the differential spinor corresponding to each error parameter and the spinor of the transmitted force and the spinor of the constraint force of the branch. If the reciprocal product is always zero, then the current error parameter is determined to be the low-sensitivity error parameter and removed from the set of parameters to be identified. The robot end effector measured pose data in the workspace is obtained using a preset external measuring device, and the measured pose data is represented in the form of a homogeneous transformation matrix to obtain the measured pose homogeneous transformation matrix. Based on the current configuration and geometric error parameter estimates, the nominal pose homogeneous transformation matrix of the robot end effector is obtained by solving the real kinematic model. The inverse matrix of the measured pose homogeneous transformation matrix is ​​multiplied by the measured pose homogeneous transformation matrix to obtain the pose residual homogeneous transformation matrix. Based on the pose information obtained from the error-pose mapping model, a Lie group logarithmic mapping operation is performed on the pose residual homogeneous transformation matrix to construct the pose residual spinor under a single configuration. Within the workspace, the pose of the robot end effector under multiple configurations is measured according to the principle of traversing working conditions, so as to calculate the pose residual screw of each measurement configuration, and the pose residual screw of each measurement configuration is set into the geometric error parameter identification equation.

[0013] Optionally, the step of reducing the ineffective differential spins of passive joints in the parallel branches of the target hybrid robot, and extracting the identification Jacobian matrix of the hybrid robot pose residuals relative to the geometric error parameters from the geometric error parameter identification equation, includes: Differentiate the pose residual screw of the hybrid robot and decompose it into the pose residual differential screw of the serial part and the pose residual differential screw of the parallel part. The transmission force screw and constraint force screw of each branch in the parallel part are applied to the corresponding pose residual differential screw. The invalid differential screw corresponding to the passive joint motion of each branch is determined and eliminated by the reciprocity product operation in order to extract the Jacobian matrix of the parallel part. Solve for the Jacobian matrix of the series connection, and combine the Jacobian matrix of the series connection with the Jacobian matrix of the parallel connection to form the Jacobian matrix of the whole machine; The identification Jacobian matrix is ​​formed by grouping the whole machine Jacobian matrices corresponding to different measurement poses.

[0014] A second aspect of the present invention provides a device for modeling and identifying geometric errors of a hybrid robot by characterizing the spinor of axial displacement, comprising: The module for constructing and acquiring is used to construct axial displacement screws that can completely characterize the inter-joint relationships of the target hybrid robot based on the Lie algebra subspace, so as to establish a real kinematic model and determine the parameterization form of the actual geometric parameters according to the real kinematic model. The decomposition module is used to characterize the actual geometric parameters as a superposition of nominal geometric parameters and error parameters, so as to construct a geometric error model that can characterize any non-ideal geometric relationship without redundancy, and to explicitly decompose the geometric error model into an error-pose mapping model and an error-drive mapping model. The first reduction and construction module is used to reduce the low-sensitivity error parameters of the parallel branch of the target hybrid robot, obtain the actual pose data of the robot end in the workspace, and combine the actual pose data of the robot end with the error-pose mapping model to construct the geometric error parameter identification equation based on the pose residual spinor. The second reduction and construction module is used to reduce the ineffective differential spins of the passive joints in the parallel chain of the target hybrid robot, and to extract the identification Jacobian matrix of the hybrid robot pose residual relative to the geometric error parameters from the geometric error parameter identification equation. The solution and compensation module is used to iteratively solve the geometric error parameter estimates based on the nonlinear least squares algorithm and the identified Jacobian matrix, and input the geometric error parameter estimates into the error-drive mapping model to compensate for the end-effector pose error of the hybrid robot.

[0015] A third aspect of the present invention provides an electronic device, including: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the program to implement the method for geometric error modeling and identification of hybrid robots with axial displacement spinor characterization as described in the above embodiments.

[0016] A fourth aspect of the present invention provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the above-described method for geometric error modeling and identification of a hybrid robot with axial displacement spinor characterization.

[0017] The present invention proposes a method and apparatus for modeling and identifying geometric errors in hybrid robots using axial displacement screw representation. This method constructs axial displacement screws using Lie algebra subspaces, establishes a generalized and standardized real kinematic model of the hybrid robot, embeds error parameters into the kinematic model, and obtains a geometric error model that simultaneously satisfies completeness and minimization by reducing redundant parameters. Several poses are measured in the workspace, and identification equations in the form of pose residual screws are constructed. Low-sensitivity error parameters and passive joint motion differential screws in the parallel part are reduced, and a Jacobian matrix of pose residuals with respect to geometric errors is established. A nonlinear least squares algorithm is used to stably identify geometric error parameters. Finally, the identification results are embedded into the inverse kinematics solution to complete error compensation. The established geometric error model can completely represent any geometric deviation while ensuring the non-redundancy of error parameters. The reduction of low-sensitivity error parameters and passive joint motion differential screws helps improve the stability of error parameter identification, ultimately achieving high-precision compensation for geometric errors in hybrid robots.

[0018] Additional aspects and advantages of the invention will be set forth in part in the description which follows, and in part will be obvious from the description, or may be learned by practice of the invention. Attached Figure Description

[0019] The above and / or additional aspects and advantages of the present invention will become apparent and readily understood from the following description of the embodiments taken in conjunction with the accompanying drawings, wherein: Figure 1 A flowchart of a method for geometric error modeling and identification of a hybrid robot based on axial displacement screw quantity characterization according to an embodiment of the present invention; Figure 2 This is a schematic diagram of a seven-degree-of-freedom hybrid robot structure used in a method for geometric error modeling and identification of a hybrid robot based on an axial displacement screw quantity characterization according to an embodiment of the present invention. Figure 3 A simplified kinematic diagram of the serial and parallel parts of a seven-degree-of-freedom hybrid robot according to an embodiment of the present invention; Figure 4 This is a schematic diagram illustrating a process for establishing a realistic kinematic model according to an embodiment of the present invention. Figure 5 This is a schematic diagram illustrating the principle of an axial displacement spinor construction according to an embodiment of the present invention; Figure 6 This is a schematic diagram illustrating the establishment of a joint coordinate system for a hybrid robot according to an embodiment of the present invention; Figure 7 This is a schematic diagram of a process for establishing a geometric error model of a hybrid robot that can characterize any non-ideal geometric relationship without redundancy, according to an embodiment of the present invention. Figure 8This is a schematic diagram of a process for reducing low-sensitivity error parameters in the parallel section of a hybrid robot and constructing identification equations according to an embodiment of the present invention. Figure 9 A flowchart illustrating a process for reducing the differential spinor of the passive joint motion of each branch in a parallel section and constructing the Jacobian matrix required for identification, according to an embodiment of the present invention. Figure 10 This is a schematic diagram of an iterative process for identifying and compensating for geometric error parameters according to an embodiment of the present invention; Figure 11 This is a block diagram of a hybrid robot geometric error modeling and identification device for representing axial displacement screw quantity according to an embodiment of the present invention; Figure 12 This is a schematic diagram of the structure of an electronic device according to an embodiment of the present invention.

[0020] Explanation of reference numerals in the attached figures: 1-Moving platform, 2-RPS branch, 3-Fixed platform, 4-Horizontal guide rail, 5-Angled guide rail, 6-Horizontal guide rail, 7-Turntable, 1201-Memory, 1202-Processor and 1203-Communication interface. Detailed Implementation

[0021] Embodiments of the present invention are described in detail below, examples of which are illustrated in the accompanying drawings, wherein the same or similar reference numerals denote the same or similar elements or elements having the same or similar functions throughout. The embodiments described below with reference to the accompanying drawings are exemplary and intended to explain the present invention, and should not be construed as limiting the present invention.

[0022] The following description, with reference to the accompanying drawings, describes a method and apparatus for modeling and identifying geometric errors in a hybrid robot by characterizing the spinor displacement of the axis of rotation, according to an embodiment of the present invention.

[0023] Figure 1 This is a flowchart illustrating a method for geometric error modeling and identification of a hybrid robot based on axial displacement screw representation according to an embodiment of the present invention.

[0024] like Figure 1 As shown, the method for modeling and identifying geometric errors of hybrid robots characterized by the spinor displacement of the axis includes the following steps: In step S101, an axial displacement screw that can completely characterize the joint relationships of the target hybrid robot is constructed based on the Lie algebra subspace to establish a real kinematic model, and the parameterization form of the actual geometric parameters is determined based on the real kinematic model.

[0025] In some embodiments, an axial displacement screw that can completely characterize the inter-joint relationships of the target hybrid robot is constructed based on the Lie algebra subspace to establish a true kinematic model, and the parameterized form of the actual geometric parameters is determined based on the true kinematic model, including: Extracting Lie Algebra Given spinor coordinates and a range of rotation angles, the elements within the four-dimensional subspace can locally and uniquely represent the relationship between the helical axis and the coordinate system. z Arbitrary helical motion with axes perpendicularly intersecting; All elements in the four-dimensional subspace are represented using six-dimensional coordinates and defined as axis displacement screws; The same screw exponential product modeling method is used for both the serial and parallel parts of the target hybrid robot to establish the kinematic equations of the serial part and the kinematic equations of each branch of the parallel part. A true kinematic model is constructed based on the kinematic equations of the series section and the kinematic equations of each branch of the parallel section, and the parameterization form of the actual geometric parameters is determined based on the true kinematic model.

[0026] In some embodiments, the screw exponential product method is performed on the serial portion of the target hybrid robot to establish the kinematic equations of the serial portion, including: Assume the serial part of the target hybrid robot consists of It consists of several components. The initial configuration of the series section is known, and a base coordinate system 0 is defined, where, It is a positive integer; Let coordinate system of z axis and corresponding joint axis The coordinate system is determined by the exponential product of the rotational displacement of the corresponding joint axis and the rotational displacement of the joint motion. With coordinate system The homogeneous transformation matrix between them, where the joint rotational spinor is defined in six-dimensional coordinate form, representing the rotation around the joint. z Rotation of the shaft or along z Axial translation element, ; Let coordinate system The pose of the end of the series section is represented by the coordinate system determined by the exponential mapping of the six-dimensional displacement spinor. To coordinate system The homogeneous transformation matrix, where the six-dimensional displacement spinor is defined as any coordinate in six-dimensional form. element; Multiply the homogeneous transformation matrices between all adjacent coordinate systems in sequence to obtain the kinematic equations of the cascaded part.

[0027] It should be noted that the target hybrid robot includes sequentially connected series and parallel parts. The order in which the series and parallel parts are combined is arbitrary. The series part is an open-chain mechanism, and the joints in the open-chain mechanism are classified as either rotary joints or translating joints according to their motion type. All joints in the open-chain mechanism are active joints. The parallel part is a closed-loop mechanism, and the joints in the closed-loop mechanism are classified as either rotary joints or translating joints according to their motion type. The joints in the open-chain mechanism are classified as either active joints or passive joints according to their driving method.

[0028] In actual implementation, Lie algebra is extracted. The four-dimensional subspace, given the screw coordinates and the selected rotation angle range, can locally and uniquely represent any helical motion in which the helical axis intersects the z-axis of the coordinate system perpendicularly; the elements of the subspace are represented by six-dimensional coordinates and defined as the axial displacement screw.

[0029] Furthermore, let the series section be composed of The system consists of several components, with a known initial configuration and a defined base coordinate system 0; let the coordinate system... of Axis and axis coincide( ), coordinate system With coordinate system The homogeneous transformation matrix between them is given by the corresponding The twist of the axis's axial displacement and the twist of the joint motion are determined by the exponential product. The twist of the joint motion is defined as a six-dimensional coordinate representation around the axis. Rotation of the shaft or along Translation of axis Elements; coordinate system Indicates the pose of the end of the cascaded component, coordinate system To coordinate system The homogeneous transformation matrix is ​​determined by the exponential mapping of the six-dimensional displacement spinor, where the six-dimensional displacement spinor is defined as an arbitrary coordinate form. Elements; multiply the homogeneous transformation matrices between all adjacent coordinate systems in sequence to obtain the kinematic equations of the cascaded part.

[0030] Furthermore, the kinematic equations of each branch in the parallel section are established using the same method, and the kinematic equations of each branch in the series and parallel sections are combined to obtain the real kinematic model of the hybrid robot.

[0031] In step S102, the actual geometric parameters are represented as a superposition of nominal geometric parameters and error parameters to construct a geometric error model that can represent any non-ideal geometric relationship without redundancy. The geometric error model is then explicitly decomposed into an error-pose mapping model and an error-drive mapping model.

[0032] In some embodiments, the actual geometric parameters are represented as a superposition of nominal geometric parameters and error parameters to construct a geometric error model that can represent any non-ideal geometric relationship without redundancy. The geometric error model is then explicitly decomposed into an error-pose mapping model and an error-drive mapping model, including: The actual geometric parameters are represented as a superposition of the nominal geometric parameters and the error parameters; Based on the superposition of nominal geometric parameters and error parameters, the axial displacement screw of each moving joint is synchronously expressed as the sum of the nominal axial displacement screw and the axial displacement screw error. The movement component in the axial displacement screw error of each moving joint is removed to obtain a geometric error model that can characterize any non-ideal geometric relationship without redundancy. The geometric error model is explicitly decomposed into an error-pose mapping model and an error-drive mapping model.

[0033] In actual execution, the actual geometric parameters are represented as a superposition of nominal geometric parameters and error parameters. Based on the superposition of nominal geometric parameters and error parameters, the axial displacement screw of each joint is synchronously represented as the sum of the nominal axial displacement screw and the axial displacement screw error, realizing the explicit embedding of all error parameters in the exponential product form of the real kinematic model. For each moving joint, the movement component in its axial displacement screw error is removed to eliminate the corresponding redundant error parameters, obtaining a geometric error model that can characterize any non-ideal geometric relationship of the hybrid robot without redundancy. Finally, the geometric error model is explicitly transformed into an error-pose mapping model and an error-drive mapping model, where the error-pose mapping model is equivalent to the forward kinematics model and the error-drive mapping model is equivalent to the inverse kinematics model.

[0034] In step S103, the low-sensitivity error parameters of the parallel branches of the target hybrid robot are reduced, the actual pose data of the robot end effector in the workspace is obtained, and the actual pose data of the robot end effector is combined with the error-pose mapping model to construct the geometric error parameter identification equation based on the pose residual spinor.

[0035] In some embodiments, the low-sensitivity error parameters of the parallel branches of the target hybrid robot are reduced, the measured pose data of the robot end effector in the workspace is obtained, and the measured pose data of the robot end effector is combined with the error-pose mapping model to construct a geometric error parameter identification equation based on the pose residual spinor, including: Under the condition of nominal geometric parameters, solve for the transmitted force screw, constraint force screw, and differential screw corresponding to each error parameter of the parallel chain of the target hybrid robot. The condition of nominal geometric parameters means that the screw error of all axis displacements is zero. The differential screw corresponding to each error parameter is used to characterize the helical motion caused by small changes in the error parameters. Perform reciprocal product operations between the differential spinor corresponding to each error parameter and the spinor of the transmitted force and the spinor of the constraint force of the branch. If the reciprocal product is always zero, the current error parameter is determined to be a low-sensitivity error parameter and removed from the set of parameters to be identified. The robot end effector's measured pose data in the workspace is obtained using a preset external measuring device, and the measured pose data is represented in the form of a homogeneous transformation matrix to obtain the measured pose homogeneous transformation matrix. Based on the current configuration and geometric error parameter estimates, the nominal pose homogeneous transformation matrix of the robot end effector is obtained by solving the real kinematic model. The inverse matrix of the measured pose homogeneous transformation matrix is ​​multiplied by the measured pose homogeneous transformation matrix to obtain the pose residual homogeneous transformation matrix. Based on the pose information obtained from the error-pose mapping model, a Lie group logarithmic mapping operation is performed on the pose residual homogeneous transformation matrix to construct the pose residual spinor under a single configuration. Within the workspace, the pose of the robot end effector under multiple configurations is measured according to the principle of traversing working conditions, so as to calculate the pose residual screw of each measurement configuration and set the pose residual screw of each measurement configuration into a geometric error parameter identification equation.

[0036] In actual execution, under the nominal geometric parameter condition that all axis displacement screw errors are zero, for each independent parallel branch of the hybrid robot, based on the reciprocity relationship between the branch motion screw and the force screw, the transmitted force screw and constraint force screw of the branch are solved. At the same time, all geometric error parameters to be identified are traversed, and the helical motion form corresponding to each error parameter when a small disturbance occurs is solved one by one to obtain the differential screw corresponding to each error parameter.

[0037] Furthermore, the differential screw quantity corresponding to each error parameter is subjected to a six-dimensional screw reciprocity product operation with the screw quantities of the transmitted force and constraint force of its respective branch. If the result of the reciprocity product of the differential screw quantity with the two types of screw quantities is always zero in all configurations of the robot's entire workspace, then the corresponding geometric error parameter is determined to be a low-sensitivity error parameter and is removed from the set of parameters to be identified. The reduction of low-sensitivity error parameters is a process in the parameter identification stage and does not change the established geometric error model or its completeness.

[0038] Furthermore, a laser tracker is used as a pre-set external measurement device to collect measured pose data of the robot's end effector within the workspace under no-load conditions. The collected pose data is represented as a homogeneous transformation matrix, yielding the measured pose homogeneous transformation matrix. Based on the current configuration and estimated geometric error parameters, a non-redundant real kinematic model in exponential product form is used to forward solve for the calculated pose homogeneous transformation matrix of the robot's end effector. The inverse of the calculated pose homogeneous transformation matrix is ​​multiplied by the measured pose homogeneous transformation matrix to obtain the pose residual homogeneous transformation matrix, which characterizes the end effector pose deviation. The pose residual homogeneous transformation matrix is ​​then converted into a Lie algebra through a Lie group logarithmic mapping. The elements in the equation are then represented using six-dimensional spinor coordinates to construct the pose residual spinor for a single configuration.

[0039] Finally, within the robot's entire workspace, following the principle of traversing the joint space and covering extreme and intermediate poses, dozens of evenly distributed measured pose data of the robot's end effector were collected. The pose residual screw of each pose point was calculated for each group. The six-dimensional pose residual screws of all pose points were stacked and grouped sequentially according to the row dimension to finally obtain the geometric error parameter identification equation with dimension matching and no redundant parameters.

[0040] In step S104, the ineffective differential spins of the passive joints in the parallel branches of the target hybrid robot are reduced, and the identification Jacobian matrix of the hybrid robot pose residuals relative to the geometric error parameters is extracted from the geometric error parameter identification equation.

[0041] In some embodiments, reducing the ineffective differential spins of passive joints in the parallel branches of the target hybrid robot, and extracting the identification Jacobian matrix of the hybrid robot pose residuals relative to the geometric error parameters from the geometric error parameter identification equation, includes: Differentiate the pose residual screw of the hybrid robot and decompose it into the pose residual differential screw of the serial part and the pose residual differential screw of the parallel part. The transmission force screw and constraint force screw of each branch in the parallel part are applied to the corresponding pose residual differential screw. The invalid differential screw corresponding to the passive joint motion of each branch is determined and eliminated through reciprocal product operation in order to extract the Jacobian matrix of the parallel part. Solve for the Jacobian matrix of the series connection and combine the Jacobian matrix of the series connection with the Jacobian matrix of the parallel connection to obtain the Jacobian matrix of the whole machine; The whole-machine Jacobian matrix corresponding to different measurement poses is grouped together to form the identification Jacobian matrix.

[0042] In actual execution, total differential operations are performed on the six-dimensional pose residual screws under a single configuration. Based on the structure of the serial and parallel parts of the hybrid robot, the pose residual differential screws are decomposed into contributions from the serial and parallel parts. The transmitted force screws and constraint force screws corresponding to each branch of the parallel part are reciprocally multiplied with the pose residual differential screws of that branch. Utilizing the property that both types of force screws are reciprocal with the passive joint motion screws of the corresponding branches, the differential screws introduced by the passive joint motion are eliminated, thereby obtaining the Jacobian matrix of the parallel part.

[0043] Furthermore, based on the axial displacement screws of each joint in the series section, the Jacobian matrix of the series section is solved segment by segment. After unifying the Jacobian matrix of the series section and the Jacobian matrix of the parallel section to the same coordinate system by using the adjoint transformation, the columns are spliced ​​according to the arrangement order of the geometric error parameters to obtain the Jacobian matrix of the entire hybrid robot under this configuration. Finally, all the pose points that have been measured in the workspace are traversed, and the whole-machine Jacobian matrix corresponding to each pose point is stacked and grouped in order according to the row dimension. Finally, the identification Jacobian matrix of the hybrid robot pose residual relative to the geometric error parameter covering all measured pose points is obtained.

[0044] In step S105, based on the nonlinear least squares algorithm, the estimated values ​​of geometric error parameters are iteratively solved using the identified Jacobian matrix, and the estimated values ​​of geometric error parameters are input into the error-drive mapping model to compensate for the end-effector pose error of the hybrid robot.

[0045] In actual execution, the six components of the pose residual spinor are weighted and normalized according to the standard deviation of the pose error in each direction to eliminate the dimensional differences between different pose residual components, and the differences in measurement accuracy are also taken into account. Based on the objective function of the weighted and normalized residual, combined with the identified Jacobian matrix completed in the previous steps, a nonlinear least squares method (such as the Levenberg-Marquardt algorithm) is adopted and appropriate regularization parameters are determined to improve the numerical stability of the iterative solution, promote the convergence of the algorithm, and iteratively identify the actual geometric error parameters. The identified geometric error parameters are directly input into the error-drive mapping model to compensate for the pose error of the hybrid robot end effector.

[0046] The following detailed description of the method for geometric error modeling and identification of hybrid robots based on the spinor characterization of axial displacement proposed in this invention will be provided through a specific embodiment.

[0047] like Figure 2As shown, taking a seven-axis hybrid robot as an example, the parallel section consists of a moving platform 1, three RPS branches 2, and a fixed platform 3, while the serial section consists of a transverse guide rail 4, an oblique guide rail 5, a horizontal guide rail 6, and a turntable 7. The parallel section has three degrees of freedom (1T2R), enabling omnidirectional 45° swing angle and high dynamic attitude adjustment. The bottom turntable provides the robot's rotational motion, while the horizontal and oblique guide rails enable the robot to perform a wide range of positioning in the vertical plane. The transverse guide rail can generate translation perpendicular to the vertical plane and is typically used as a redundant axis.

[0048] like Figure 3 As shown, the configuration of the series section is RPPP, and the configuration of the parallel section is 3-SPR. The axes of the first revolute joint and the first prismatic joint in the series section intersect perpendicularly, the axes of the first prismatic joint and the second prismatic joint intersect at an angle of 45°, and the axes of the second prismatic joint and the third prismatic joint are perpendicular to each other in opposite planes. ( ) represents the three RPS branches in the parallel section, where the ball joint is equivalent to three revolute joints with their axes perpendicular to each other and intersecting at a point. ( Evenly distributed around the center of the circle. , radius is On the circle, there is the center of the rotating joint where the three RPS branches are connected to the fixed platform. ( Evenly distributed around the center of the circle. , radius is On the circle, there is the center of the ball joint that represents the connection between the three RPS branches and the moving platform.

[0049] like Figure 4 As shown, extracting Lie algebras se (3) In the four-dimensional subspace, given the spinor coordinates and the selected rotation angle range, its elements can locally and uniquely represent the relationship between the helical axis and the coordinate system. Arbitrary helical motion with perpendicularly intersecting axes; the elements of the subspace are represented using six-dimensional coordinates and defined as the rotational displacement of the axes; assuming the series part consists of... The system consists of several components, with a known initial configuration and a defined base coordinate system 0; let the coordinate system... of Axis and axis coincide( The homogeneous transformation matrix of adjacent coordinate systems is determined by the exponential product of the rotations of axial displacement and joint motion; coordinate system Represents the pose of the end effector, coordinate system To coordinate system The homogeneous transformation matrix is ​​determined by the exponential mapping of the six-dimensional displacement spinor; the kinematic equations of the series part are obtained by multiplying the homogeneous transformation matrix in sequence, and the kinematic equations of each branch of the parallel part are established by the same method; the kinematic equations of each branch of the parallel part and the series part are combined to obtain the real kinematic model of the hybrid robot.

[0050] Representable with coordinate system Lie algebra for arbitrary helical motions with perpendicularly intersecting axes se (3) The four-dimensional subspace element is defined as the axis displacement spinor, and its six-dimensional coordinates are... The format is:

[0051] in, For rotational components, For the shift component, The translation distance of the helical axis. For rotational components coordinate, For rotational components coordinate, For the shift component coordinate, For the shift component coordinate.

[0052] Joint rotation The expression is:

[0053] in, For the driving variable of the rotating pair, This is the moving pair driving variable.

[0054] The tandem section has four joint axes, with corresponding axial displacement rotation and joint motion rotation as follows: ,

[0055] ,

[0056] ,

[0057] ,

[0058] in, This refers to the height of the slide on the horizontal guide rail.

[0059] The parallel section has 3 branches, each branch has 5 joint axes, and the corresponding axis displacement rotation and joint motion rotation are as follows: ,

[0060] ,

[0061] ,

[0062] ,

[0063] ,

[0064] The coordinate systems of each joint, determined by the aforementioned axial displacement screw and joint motion screw, have been marked as follows: Figure 3 In the context of homogeneous transformation matrices between adjacent coordinate systems Obtained through the exponential product:

[0065] in, .

[0066] The homogeneous transformation matrix between the parallel branch coordinate system 45 and the moving platform coordinate system can be expressed as:

[0067]

[0068] in, This is the coordinate system transformation matrix at the connection point between the series and parallel parts.

[0069] The kinematic equations of each branch in the parallel section can be expressed as:

[0070] like Figure 5 As shown, the rotational displacement of the axis is applied to the coordinate system. If the initial axes coincide, then the initial axes are about a distance perpendicular to the coordinate system. Axis and distance flat Rotation angle of the helical shaft Then translate the distance along the spiral axis. This allows the generation of axes in any orientation.

[0071] like Figure 6 As shown, the axial displacement screw and joint motion screw are applied sequentially to the first... Joint coordinate system, can generate the first Joint coordinate system.

[0072] like Figure 7 As shown, the axial displacement screw is expressed as the sum of the nominal axial displacement screw and the axial displacement screw error, thus embedding the error parameter into the real kinematic model. The movement component in the axial displacement screw error of the moving joint is removed to obtain a hybrid robot geometric error model that can characterize any non-ideal geometric relationship without redundancy. The geometric error model is explicitly expressed as an error-pose mapping model and an error-drive mapping model.

[0073] Rotor error of axial displacement after eliminating redundant error parameters Represented as:

[0074] in, For the rotational component of the spinor error of the axial displacement coordinate, For the rotational component of the spinor error of the axial displacement coordinate, The shift component of the rotational error of the axial displacement coordinate, The shift component of the rotational error of the axial displacement The components of the coordinate and axis displacement screw error are geometric error parameters.

[0075] The explicit error-pose mapping model and error-drive mapping model are respectively represented as follows:

[0076] in, For the end-effector pose vector of the hybrid robot, This is a vector composed of the variables of each drive joint of the hybrid robot. It is a vector consisting of all geometric error parameters. For positive kinematic functions, It is the inverse kinematic function.

[0077] like Figure 8As shown, under nominal geometric parameters, the power spinor and constraint spinor of each branch in the parallel part are solved; under nominal geometric parameters, the differential spinor corresponding to each error parameter in the parallel part is solved; the differential spinor corresponding to the error parameter is reciprocally multiplied with the power spinor and constraint spinor of the branch respectively. If the reciprocal product is always zero, the error parameter is removed from the set of parameters to be identified; the position and attitude of the hybrid robot are obtained through an external measuring device and expressed in the form of a homogeneous transformation matrix; the pose residual spinor under a single configuration is constructed through an error-pose mapping model and combined with logarithmic mapping; several pose points are measured in the workspace according to the principle of traversing the working conditions, the pose residual spinor at all pose points is calculated, and the equations are set together to form the identification equation.

[0078] The spinors of the transmitted forces and the spinors of the constraint forces in each branch can be solved by the following formula:

[0079] in, To transmit force spinor, For the first A permissible rotational quantity of motion For the first A constraint force spinor, The linear combination coefficients of the constraint force spinors, This refers to the serial number of the drive joint.

[0080] Pose residual spinor in a single configuration Represented as:

[0081] in, The homogeneous transformation matrix corresponding to the end-effector pose of the hybrid robot acquired by the external measurement device.

[0082] like Figure 9 As shown, the pose residual screw of the hybrid robot is differentiated and decomposed into the pose residual differential screw of the serial part and the pose residual differential screw of the parallel part. The transmission force screw and constraint force screw of each branch in the parallel part are applied to the corresponding pose residual differential screw to reduce the differential screw of the passive joint motion of each branch, and the Jacobian matrix of the parallel part is obtained. The Jacobian matrix of the serial part of the hybrid robot is solved and combined with the Jacobian matrix of the parallel part to form the whole Jacobian matrix of the hybrid robot. The whole Jacobian matrices corresponding to different measurement poses are grouped to form the Jacobian matrix required for error parameter identification.

[0083] Jacobian matrix of the concatenated part Represented as:

[0084] in, For the adjoint transformation operator, For the first part of the series Rotation of the axial displacement of each joint. For Li Qun se (3) is the right Jacobian matrix. This represents the number of joints in the series connection.

[0085] The Jacobian matrix for the parallel portion of the differential spins of the passive joint motions of each branch is expressed as:

[0086] in, ( ) indicates the parallel part. The Jacobian matrix of a branch is solved in the same way as that of the Jacobian matrix of the chained part.

[0087] The overall Jacobian matrix obtained by combining the Jacobian matrices of the serial and parallel components is expressed as follows:

[0088] like Figure 10 As shown, the residual components are weighted according to the standard deviation of the pose error in each direction and a suitable regularization parameter is determined; the actual geometric error parameters are identified iteratively using the nonlinear least squares method; and the identified geometric error parameters are directly input into the error-drive mapping model of the hybrid robot.

[0089] Geometric error parameters The iterative formula for the nonlinear least squares method is:

[0090] in, , , It is the residual Jacobian matrix obtained by combining the whole-machine Jacobian matrices of different measurement poses. It is the total residual of the residual spinor of the set of different measurement poses. The weighting matrix is ​​constructed based on the standard deviation of the pose error in each direction. For regularization parameters, It is the identity matrix. For iteration rounds.

[0091] In summary, the method for modeling and identifying geometric errors of hybrid robots based on the spinor characterization of axial displacement proposed in this embodiment of the invention has the following beneficial effects: (1) The established geometric error model simultaneously satisfies the two core properties of completeness and minimization. Completeness ensures that the parameters in the error model can fully represent any geometric deviation, reflecting the model's ability to cover geometric errors. The reduction of low-sensitivity error parameters is only used in the identification process and does not change the established geometric error model or its completeness. Minimization requires that the error parameters have no redundancy and that the column vectors of the corresponding Jacobian matrix are linearly independent, i.e., full column rank, which is a necessary condition for stable identification. (2) Compared with existing modeling methods, it avoids the problem of parameter singularity and eliminates the linear correlation caused by redundant parameters, thus enhancing the anti-interference capability of the identification process; it naturally embeds complete geometric error parameters into the kinematic model framework, supports high-precision compensation through iterative parameter updates, thereby improving the end pose accuracy of the hybrid robot.

[0092] Next, referring to the accompanying drawings, a hybrid robot geometric error modeling and identification device for representing axial displacement screw quantity according to an embodiment of the present invention is described.

[0093] Figure 11 This is a block diagram of a hybrid robot geometric error modeling and identification device for representing axial displacement screw quantity according to an embodiment of the present invention.

[0094] like Figure 11 As shown, the hybrid robot geometric error modeling and identification device 110 characterized by the spinor of the axial displacement includes: a construction and acquisition module 1101, a decomposition module 1102, a first reduction and construction module 1103, a second reduction and construction module 1104, and a solution and compensation module 1105.

[0095] The module 1101 is used to construct axial displacement screws that can completely characterize the joint relationships of the target hybrid robot based on the Lie algebra subspace, in order to establish a real kinematic model and determine the parameterized form of the actual geometric parameters based on the real kinematic model. The decomposition module 1102 is used to characterize the actual geometric parameters as a superposition of nominal geometric parameters and error parameters, in order to construct a geometric error model that can characterize any non-ideal geometric relationship without redundancy, and to explicitly decompose the geometric error model into an error-pose mapping model and an error-drive mapping model. The first reduction and construction module 1103 is used to reduce the low-sensitivity error parameters of the parallel branches of the target hybrid robot, acquire the measured pose data of the robot end effector in the workspace, and combine the measured pose data of the robot end effector with the error-pose mapping model to construct a geometric error parameter identification equation based on the pose residual screw. The second reduction and construction module 1104 is used to reduce the ineffective differential spins of passive joints in the parallel branches of the target hybrid robot, and extracts the identification Jacobian matrix of the hybrid robot pose residuals relative to the geometric error parameters from the geometric error parameter identification equation. The solution and compensation module 1105 is used to iteratively solve the geometric error parameter estimates using the identification Jacobian matrix based on a nonlinear least squares algorithm, and inputs the geometric error parameter estimates into the error-drive mapping model to compensate for the end-effector pose error of the hybrid robot.

[0096] In some embodiments, the target hybrid robot includes sequentially connected series and parallel portions, wherein, The series part is in the form of an open chain mechanism. The joints in the open chain mechanism are classified into rotary joints or translating joints according to the type of motion. All joints in the open chain mechanism are active joints. The parallel part is a closed-loop mechanism. The joints in the closed-loop mechanism are classified into rotary joints or translating joints according to their motion type. The joints in the open-loop mechanism are classified into active joints or passive joints according to their driving method.

[0097] In some embodiments, the construction and acquisition module 1101 includes: Extraction unit, used to extract Lie algebras Given spinor coordinates and a range of rotation angles, the elements within the four-dimensional subspace can locally and uniquely represent the relationship between the helical axis and the coordinate system. z Arbitrary helical motion with axes perpendicularly intersecting; Define a unit for representing all elements in a four-dimensional subspace using six-dimensional coordinates, and define it as the axis displacement screw; The execution unit is used to apply the same screw exponential product modeling method to the serial and parallel parts of the target hybrid robot respectively, so as to establish the kinematic equations of the serial part and the kinematic equations of each branch of the parallel part. The building unit is used to construct a real kinematic model based on the kinematic equations of the series part and the kinematic equations of each branch of the parallel part, and to determine the parameterized form of the actual geometric parameters based on the real kinematic model.

[0098] In some embodiments, the execution unit includes: Define sub-units to define the serial part of the target hybrid robot as consisting of It consists of several components. The initial configuration of the series section is known, and a base coordinate system 0 is defined, where, It is a positive integer; The first defined sub-unit is used to define the coordinate system. of z axis and corresponding joint axis The coordinate system is determined by the exponential product of the rotational displacement of the corresponding joint axis and the rotational displacement of the joint motion. With coordinate system The homogeneous transformation matrix between them, where the joint rotational spinor is defined in six-dimensional coordinate form, representing the rotation around the joint. z Rotation of the shaft or along z Axial translation element, ; The second defined sub-unit is used to define the coordinate system. The pose of the end of the series section is represented by the coordinate system determined by the exponential mapping of the six-dimensional displacement spinor. To coordinate system The homogeneous transformation matrix, where the six-dimensional displacement spinor is defined as any coordinate in six-dimensional form. element; The multiplication sub-unit is used to multiply the homogeneous transformation matrices between all adjacent coordinate systems in sequence to obtain the kinematic equations of the cascaded part.

[0099] In some embodiments, the decomposition module 1102 includes: The characterization unit is used to characterize the actual geometric parameters as a superposition of the nominal geometric parameters and the error parameters; The synchronous representation unit is used to synchronously represent the axial displacement screw of each moving joint as the sum of the nominal axial displacement screw and the axial displacement screw error, based on the superposition of nominal geometric parameters and error parameters. The first elimination unit is used to eliminate the moving component in the axial displacement screw error of each moving joint in order to obtain a geometric error model that can characterize any non-ideal geometric relationship without redundancy. Decomposition units are used to explicitly decompose the geometric error model into an error-pose mapping model and an error-drive mapping model.

[0100] In some embodiments, the first reduction and construction module 1103 includes: The solver unit is used to solve the transmitted force screw, constraint force screw, and differential screw corresponding to each error parameter of the parallel chain of the target hybrid robot under the condition of nominal geometric parameters. The nominal geometric parameter condition means that the screw error of all axis displacements is zero, and the differential screw corresponding to each error parameter is used to characterize the helical motion caused by the small changes in the error parameters. The operation unit is used to perform reciprocal product operation between the differential spinor corresponding to each error parameter and the transmission force spinor and constraint force spinor of the branch respectively. If the reciprocal product is always zero, the current error parameter is determined to be a low-sensitivity error parameter and removed from the set of parameters to be identified. The acquisition unit is used to acquire the measured pose data of the robot end effector in the workspace using a preset external measuring device, and to represent the measured pose data of the robot end effector in the form of a homogeneous transformation matrix to obtain the measured pose homogeneous transformation matrix. The multiplication unit is used to obtain the nominal pose homogeneous transformation matrix of the robot end effector by solving the real kinematic model based on the current configuration and geometric error parameter estimates. The inverse matrix of the measured pose homogeneous transformation matrix is ​​multiplied with the measured pose homogeneous transformation matrix to obtain the pose residual homogeneous transformation matrix. The mapping unit is used to perform Lie group logarithmic mapping operation on the pose residual homogeneous transformation matrix based on the pose information obtained from the error-pose mapping model, so as to construct the pose residual spinor under a single configuration. The first set of units is used to measure the pose of the robot end effector in multiple configurations within the workspace according to the principle of traversing working conditions, to calculate the pose residual screw of each measurement configuration, and to set the pose residual screw of each measurement configuration into a geometric error parameter identification equation.

[0101] In some embodiments, the second reduction and construction module 1104 includes: Differential unit, used to differentiate the pose residual screw of the hybrid robot, decompose it into the pose residual differential screw of the serial part and the pose residual differential screw of the parallel part. The reduction unit is used to make the transmission force screw and constraint force screw of each branch in the parallel part act on the corresponding pose residual differential screw. The invalid differential screw corresponding to the passive joint motion of each branch is determined and reduced through reciprocal product operation in order to extract the Jacobian matrix of the parallel part. The combined unit is used to solve the Jacobian matrix of the series section and combine the Jacobian matrix of the series section with the Jacobian matrix of the parallel section to form the Jacobian matrix of the whole machine. The first and second set units are used to set up the whole machine Jacobian matrix corresponding to different measurement poses to form the identification Jacobian matrix.

[0102] It should be noted that the explanation of the above-mentioned embodiment of the method for modeling and identifying geometric errors of hybrid robots by characterizing the spinor of axial displacement also applies to the device for modeling and identifying geometric errors of hybrid robots by characterizing the spinor of axial displacement in this embodiment, and will not be repeated here.

[0103] The hybrid robot geometric error modeling and identification device based on the axial displacement screw quantity characterization proposed in the embodiments of the present invention has the following beneficial effects: (1) The established geometric error model simultaneously satisfies the two core properties of completeness and minimization. Completeness ensures that the parameters in the error model can fully represent any geometric deviation, reflecting the model's ability to cover geometric errors. The reduction of low-sensitivity error parameters is only used in the identification process and does not change the established geometric error model or its completeness. Minimization requires that the error parameters have no redundancy and that the column vectors of the corresponding Jacobian matrix are linearly independent, i.e., full column rank, which is a necessary condition for stable identification. (2) Compared with existing modeling methods, it avoids the problem of parameter singularity and eliminates the linear correlation caused by redundant parameters, thus enhancing the anti-interference capability of the identification process; it naturally embeds complete geometric error parameters into the kinematic model framework, supports high-precision compensation through iterative parameter updates, thereby improving the end pose accuracy of the hybrid robot.

[0104] Figure 12 This is a schematic diagram of the structure of an electronic device according to an embodiment of the present invention.

[0105] The electronic device may include: a memory 1201, a processor 1202, and a computer program stored on the memory 1201 and executable on the processor 1202.

[0106] When the processor 1202 executes the program, it implements the method for geometric error modeling and identification of hybrid robots with axis displacement screw representation provided in the above embodiments.

[0107] Furthermore, electronic devices also include: Communication interface 1203 is used for communication between memory 1201 and processor 1202.

[0108] The memory 1201 is used to store computer programs that can run on the processor 1202.

[0109] The memory 1201 may include high-speed RAM memory, and may also include non-volatile memory, such as at least one disk storage.

[0110] If the memory 1201, processor 1202, and communication interface 1203 are implemented independently, then the communication interface 1203, memory 1201, and processor 1202 can be interconnected via a bus to complete communication between them. The bus can be an Industry Standard Architecture (ISA) bus, a Peripheral Component Interconnect (PCI) bus, or an Extended Industry Standard Architecture (EISA) bus, etc. Buses can be divided into address buses, data buses, control buses, etc. For ease of representation, Figure 12 The bus is represented by a single thick line, but this does not mean that there is only one bus or one type of bus.

[0111] Optionally, in a specific implementation, if the memory 1201, processor 1202, and communication interface 1203 are integrated on a single chip, then the memory 1201, processor 1202, and communication interface 1203 can communicate with each other through an internal interface.

[0112] The processor 1202 may be a central processing unit (CPU), an application specific integrated circuit (ASIC), or one or more integrated circuits configured to implement embodiments of the present invention.

[0113] This invention also provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the above-described method for geometric error modeling and identification of hybrid robots with axial displacement screw representation.

[0114] In the description of this specification, the references to terms such as "one embodiment," "some embodiments," "example," "specific example," or "some examples," etc., indicate that a specific feature, structure, material, or characteristic described in connection with that embodiment or example is included in at least one embodiment or example of the present invention. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples. Moreover, without contradiction, those skilled in the art can combine and integrate the different embodiments or examples described in this specification, as well as the features of different embodiments or examples.

[0115] Furthermore, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of indicated technical features. Thus, a feature defined as "first" or "second" may explicitly or implicitly include at least one of that feature. In the description of this invention, "N" means at least two, such as two, three, etc., unless otherwise explicitly specified.

[0116] Any process or method description in the flowchart or otherwise herein can be understood as representing a module, segment, or portion of code comprising one or N executable instructions for implementing custom logic functions or processes, and the scope of preferred embodiments of the invention includes additional implementations in which functions may be performed not in the order shown or discussed, including substantially simultaneously or in reverse order depending on the functions involved, as will be understood by those skilled in the art to which embodiments of the invention pertain.

[0117] The logic and / or steps represented in the flowchart or otherwise described herein, for example, can be considered as a sequenced list of executable instructions for implementing logical functions, and can be embodied in any computer-readable medium for use by, or in conjunction with, an instruction execution system, apparatus, or device (such as a computer-based system, a processor-included system, or other system that can fetch and execute instructions from, an instruction execution system, apparatus, or device). For the purposes of this specification, "computer-readable medium" can be any means that can contain, store, communicate, propagate, or transmit programs for use by, or in conjunction with, an instruction execution system, apparatus, or device. More specific examples (a non-exhaustive list) of computer-readable media include: an electrical connection having one or more wires (electronic device), a portable computer disk drive (magnetic device), random access memory (RAM), read-only memory (ROM), erasable and editable read-only memory (EPROM or flash memory), fiber optic devices, and portable optical disc read-only memory (CDROM). Alternatively, the computer-readable medium may be paper or other suitable media on which the program can be printed, since the program can be obtained electronically by optically scanning the paper or other medium, followed by editing, interpreting, or otherwise processing as necessary, and then stored in a computer memory.

[0118] It should be understood that various parts of the present invention can be implemented in hardware, software, firmware, or a combination thereof. In the above embodiments, the N steps or methods can be implemented in software or firmware stored in memory and executed by a suitable instruction execution system. If implemented in hardware, as in another embodiment, it can be implemented using any one or a combination of the following techniques known in the art: discrete logic circuits having logic gates for implementing logical functions on data signals, application-specific integrated circuits (ASICs) having suitable combinational logic gates, programmable gate arrays (PGAs), field-programmable gate arrays (FPGAs), etc.

[0119] Those skilled in the art will understand that all or part of the steps of the methods in the above embodiments can be implemented by a program instructing related hardware. The program can be stored in a computer-readable storage medium, and when executed, the program includes one or a combination of the steps of the method embodiments.

[0120] Furthermore, the functional units in the various embodiments of the present invention can be integrated into a processing module, or each unit can exist physically separately, or two or more units can be integrated into a module. The integrated module can be implemented in hardware or as a software functional module. If the integrated module is implemented as a software functional module and sold or used as an independent product, it can also be stored in a computer-readable storage medium.

[0121] The storage medium mentioned above can be a read-only memory, a disk, or an optical disk, etc. Although embodiments of the present invention have been shown and described above, it is to be understood that the above embodiments are exemplary and should not be construed as limiting the present invention. Those skilled in the art can make changes, modifications, substitutions, and variations to the above embodiments within the scope of the present invention.

Claims

1. A method for modeling and identifying geometric errors in hybrid robots by characterizing axial displacement screw force, characterized in that, Includes the following steps: Based on the Lie algebra subspace, an axial displacement screw that can completely characterize the joint relationships of the target hybrid robot is constructed to establish a real kinematic model, and the parameterized form that determines the actual geometric parameters is obtained based on the real kinematic model. The actual geometric parameters are represented as a superposition of nominal geometric parameters and error parameters to construct a geometric error model that can represent any non-ideal geometric relationship without redundancy. The geometric error model is then explicitly decomposed into an error-pose mapping model and an error-drive mapping model. To reduce the low-sensitivity error parameters of the parallel branch of the target hybrid robot, the measured pose data of the robot end effector in the workspace is obtained, and the measured pose data of the robot end effector is combined with the error-pose mapping model to construct a geometric error parameter identification equation based on the pose residual spinor. Eliminate the ineffective differential spins of passive joints in the parallel chain of the target hybrid robot, and extract the identification Jacobian matrix of the hybrid robot pose residual relative to the geometric error parameters from the geometric error parameter identification equation; Based on the nonlinear least squares algorithm, the geometric error parameter estimates are iteratively solved using the identified Jacobian matrix, and then input into the error-drive mapping model to compensate for the end-effector pose error of the hybrid robot.

2. The method for geometric error modeling and identification of hybrid robots based on the rotational characteristic of axial displacement according to claim 1, characterized in that, The target hybrid robot comprises a series connection and a parallel connection, wherein, The series connection is in the form of an open chain mechanism. Each joint in the open chain mechanism is classified into a rotary joint or a traversing joint according to the type of motion. Each joint in the open chain mechanism is an active joint. The parallel part is in the form of a closed-loop mechanism. Each joint in the closed-loop mechanism is classified into a rotary joint or a traverse joint according to the type of motion. Each joint in the open-loop mechanism is classified into an active joint or a passive joint according to the driving method.

3. The method for geometric error modeling and identification of hybrid robots based on the rotational characteristic of axial displacement according to claim 1, characterized in that, The method involves constructing axial displacement screws based on Lie algebra subspaces to fully characterize the joint relationships of the target hybrid robot, thereby establishing a realistic kinematic model. The method then determines the parameterized form of the actual geometric parameters based on this realistic kinematic model, including: Extracting Lie Algebra In a four-dimensional subspace, given spinor coordinates and a selected range of rotation angles, the elements within the four-dimensional subspace can locally and uniquely represent the relationship between the helical axis and the coordinate system. z Arbitrary helical motion with axes perpendicularly intersecting; All elements within the four-dimensional subspace are represented using six-dimensional coordinates and defined as axial displacement screws; The same screw exponential product modeling method is used for the serial and parallel parts of the target hybrid robot to establish the kinematic equations of the serial part and the kinematic equations of each branch of the parallel part. The real kinematic model is constructed based on the kinematic equations of the series part and the kinematic equations of each branch of the parallel part, and the parameterized form of the actual geometric parameters is determined based on the real kinematic model.

4. The method for geometric error modeling and identification of hybrid robots based on the rotational characteristic of axial displacement according to claim 3, characterized in that, The step of performing the spinor exponential product method on the serial portion of the target hybrid robot to establish the kinematic equations of the serial portion includes: Assume the serial part of the target hybrid robot is composed of Composed of several components, the initial configuration of the series section is known, and a base coordinate system 0 is defined, wherein, It is a positive integer; Let coordinate system of z axis and corresponding joint axis The coordinate system is determined by the exponential product of the rotational displacement of the corresponding joint axis and the rotational displacement of the joint motion. With coordinate system The homogeneous transformation matrix between them, wherein the joint motion spinor is defined in six-dimensional coordinate form, representing the rotation around the joint. z Rotation of the shaft or along z Axial translation element, ; Let coordinate system The pose of the end of the series section is represented by the coordinate system determined by the exponential mapping of the six-dimensional displacement spinor. To coordinate system The homogeneous transformation matrix, wherein the six-dimensional displacement spinor is defined as any... in six-dimensional coordinate form. element; Multiply the homogeneous transformation matrices between all adjacent coordinate systems in sequence to obtain the kinematic equations of the cascaded part.

5. The method for geometric error modeling and identification of hybrid robots based on the rotational characteristic of axial displacement according to claim 1, characterized in that, The process of characterizing the actual geometric parameters as a superposition of nominal geometric parameters and error parameters to construct a geometric error model that can characterize any non-ideal geometric relationship without redundancy, and explicitly decomposing the geometric error model into an error-pose mapping model and an error-drive mapping model, includes: The actual geometric parameters are represented as a superposition of nominal geometric parameters and error parameters; Based on the superposition of the nominal geometric parameters and error parameters, the axial displacement screw of each moving joint is synchronously expressed as the sum of the nominal axial displacement screw and the axial displacement screw error; The movement component in the axial displacement screw error of each moving joint is removed to obtain a geometric error model that can characterize any non-ideal geometric relationship without redundancy. The geometric error model is explicitly decomposed into the error-pose mapping model and the error-driven mapping model.

6. The method for geometric error modeling and identification of hybrid robots based on the rotational characterization of axial displacement according to claim 1, characterized in that, The process of reducing the low-sensitivity error parameters of the parallel branches of the target hybrid robot involves acquiring the measured pose data of the robot's end effector within the workspace, and combining the measured pose data with the error-pose mapping model to construct a geometric error parameter identification equation based on the pose residual spinor. This includes: Under the conditions of the nominal geometric parameters, the power screw, constraint screw, and differential screw corresponding to each error parameter of the parallel chain of the target hybrid robot are solved. The nominal geometric parameter condition means that the screw error of all axis displacements is zero. The differential screw corresponding to each error parameter is used to characterize the helical motion caused by small changes in the error parameters. Perform a reciprocal product operation between the differential spinor corresponding to each error parameter and the spinor of the transmitted force and the spinor of the constraint force of the branch. If the reciprocal product is always zero, then the current error parameter is determined to be the low-sensitivity error parameter and removed from the set of parameters to be identified. The robot end effector measured pose data in the workspace is obtained using a preset external measuring device, and the measured pose data is represented in the form of a homogeneous transformation matrix to obtain the measured pose homogeneous transformation matrix. Based on the current configuration and geometric error parameter estimates, the nominal pose homogeneous transformation matrix of the robot end effector is obtained by solving the real kinematic model. The inverse matrix of the measured pose homogeneous transformation matrix is ​​multiplied by the measured pose homogeneous transformation matrix to obtain the pose residual homogeneous transformation matrix. Based on the pose information obtained from the error-pose mapping model, a Lie group logarithmic mapping operation is performed on the pose residual homogeneous transformation matrix to construct the pose residual spinor under a single configuration. Within the workspace, the pose of the robot end effector under multiple configurations is measured according to the principle of traversing working conditions, so as to calculate the pose residual screw of each measurement configuration, and the pose residual screw of each measurement configuration is set into the geometric error parameter identification equation.

7. The method for geometric error modeling and identification of hybrid robots based on the rotational characterization of axial displacement according to claim 1, characterized in that, The process of reducing the ineffective differential spins of passive joints in the parallel branches of the target hybrid robot, and extracting the identification Jacobian matrix of the hybrid robot pose residuals relative to the geometric error parameters from the geometric error parameter identification equation, includes: Differentiate the pose residual screw of the hybrid robot and decompose it into the pose residual differential screw of the serial part and the pose residual differential screw of the parallel part. The transmission force screw and constraint force screw of each branch in the parallel part are applied to the corresponding pose residual differential screw. The invalid differential screw corresponding to the passive joint motion of each branch is determined and eliminated by the reciprocity product operation in order to extract the Jacobian matrix of the parallel part. Solve for the Jacobian matrix of the series connection, and combine the Jacobian matrix of the series connection with the Jacobian matrix of the parallel connection to form the Jacobian matrix of the whole machine; The identification Jacobian matrix is ​​formed by grouping the whole machine Jacobian matrices corresponding to different measurement poses.

8. A device for modeling and identifying geometric errors of a hybrid robot by characterizing the spinor of axial displacement, characterized in that, include: The module for constructing and acquiring is used to construct axial displacement screws that can completely characterize the inter-joint relationships of the target hybrid robot based on the Lie algebra subspace, so as to establish a real kinematic model and determine the parameterization form of the actual geometric parameters according to the real kinematic model. The decomposition module is used to characterize the actual geometric parameters as a superposition of nominal geometric parameters and error parameters, so as to construct a geometric error model that can characterize any non-ideal geometric relationship without redundancy, and to explicitly decompose the geometric error model into an error-pose mapping model and an error-drive mapping model. The first reduction and construction module is used to reduce the low-sensitivity error parameters of the parallel branch of the target hybrid robot, obtain the actual pose data of the robot end in the workspace, and combine the actual pose data of the robot end with the error-pose mapping model to construct the geometric error parameter identification equation based on the pose residual spinor. The second reduction and construction module is used to reduce the ineffective differential spins of the passive joints in the parallel chain of the target hybrid robot, and to extract the identification Jacobian matrix of the hybrid robot pose residual relative to the geometric error parameters from the geometric error parameter identification equation. The solution and compensation module is used to iteratively solve the geometric error parameter estimates based on the nonlinear least squares algorithm and the identified Jacobian matrix, and input the geometric error parameter estimates into the error-drive mapping model to compensate for the end-effector pose error of the hybrid robot.

9. An electronic device, characterized in that, include: The method includes a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the program to implement the method for geometric error modeling and identification of hybrid robots with axial displacement spinor characterization as described in any one of claims 1-6.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that, The program is executed by the processor to implement the method for geometric error modeling and identification of hybrid robots with axial displacement spinor characterization as described in any one of claims 1-6.