Performance-aware motion planning method and system for series-parallel redundant manipulator

Abstract

The application provides a performance-aware motion planning method and system for a series-parallel hybrid redundant manipulator, the method comprising the following steps: S1, establishing an ideal kinematic model of the series-parallel hybrid robot and defining a plurality of coordinate systems; S2, collecting a plurality of sets of measurement data under a preset excitation trajectory, introducing corresponding parallel platform poses as implicit local pose variables, and constructing parallel chain constraint residuals and end pose measurement residuals; S3, uniformly linearizing the residuals on SE(3) to obtain a joint identification Jacobian matrix of global error parameters and implicit local pose variables, and then obtaining a calibration result; and S4, updating the ideal kinematic model of the series-parallel hybrid robot using the calibration result and performing error compensation. The application can complete unified calibration of the parallel and series subsystems without explicitly solving the parallel platform forward kinematics, can significantly improve the parameter observability and calibration accuracy, and is suitable for high-precision assembly, machining and operation scenes.

Description

Technical Field

[0001] This invention relates to the field of robot kinematics calibration technology, specifically to a performance-aware motion planning method and system for a serial-parallel hybrid redundant robotic arm. Background Technology

[0002] Parallel-serial robots combine the advantages of high rigidity in parallel mechanisms and large workspace in serial mechanisms. However, their complex mechanisms and multiple sources of assembly and geometric errors can easily lead to decreased end-effector positioning accuracy. For parallel mechanisms such as the Stewart platform, the forward kinematics often lacks a closed-form solution, and the platform pose cannot be directly obtained analytically from the driving variables. Traditional calibration often requires numerically solving the parallel forward kinematics for each measurement configuration or introducing additional platform measurement devices, resulting in complex calibration modeling and high computational overhead. Inconsistent rotational linearization leads to non-geometrical Jacobian inconsistency, affecting convergence and accuracy. It is difficult to jointly identify installation and geometric errors in serial and parallel subsystems within a unified framework. Therefore, a hybrid parallel-serial robot calibration method is needed that does not require explicit solution of the forward kinematics of the parallel mechanism, can handle consistent linearization of multiple rotational variables within a unified framework, can jointly identify multi-source geometric / assembly errors, and possesses good numerical stability and identifiability. Summary of the Invention

[0003] To address the shortcomings of existing technologies, the purpose of this invention is to provide a performance-aware motion planning method and system for serial-parallel hybrid redundant robotic arms.

[0004] A performance-aware motion planning method for a serial-parallel hybrid redundant robotic arm, provided by the present invention, includes: Step S1: Establish an ideal kinematic model of the serial-parallel hybrid robot and define multiple coordinate systems; Step S2: Collect multiple sets of measurement data under the preset excitation trajectory, and based on the ideal kinematic model of the serial-parallel hybrid robot, introduce the parallel platform pose corresponding to each set of measurement data as an implicit local pose variable to construct the parallel branch constraint residual and the end pose measurement residual. Step S3: Perform uniform linearization on SE(3) to obtain the joint identification Jacobian matrix of global error parameters and implicit local pose variables, and then construct a weighted nonlinear least squares objective function to solve for global error parameters and implicit local pose variables of each measurement configuration to obtain calibration results. Step S4: Update the ideal kinematic model of the serial-parallel hybrid robot using the calibration results and perform error compensation.

[0005] Preferably, the plurality of coordinate systems include the parallel mechanism moving platform coordinate system {P}, the base coordinate system {G}, the serial manipulator base coordinate system {B}, and the end effector coordinate system {F}; the plurality of sets of measurement data include the measured actuator variables of each drive branch of the parallel mechanism, the joint variables of the serial manipulator, and the measured pose of the end effector in the base coordinate system {G}.

[0006] Preferably, step S2 includes constructing an error parameter set; the error parameter set includes the position deviation of the hinge point of the parallel mechanism, the offset of the parallel actuator, the geometric parameter deviation of the serial manipulator, and the installation posture error between the coordinate system {P} of the moving platform of the parallel mechanism and the coordinate system {B} of the base of the serial manipulator.

[0007] Preferably, the parallel branch constraint residual is constructed in the following ways: For the first j Group measurement data and the first i Branch chain to measure actuator length With predicted length The difference is used as the residual, and the predicted length takes into account the hinge point deviation and actuator offset, satisfying:

[0008] in r Ai For the first time on the platform i One hinge point A i exist{ P The nominal position vector in} r Bi For the next platform i One hinge point B i exist{ G The nominal position vector in} p j For the first j Under a measurement configuration, the coordinate system of the parallel mechanism moving platform { P}Origin relative to { G The translation vector of}, Δ A i Δ B i The upper and lower platforms are respectively i The positional deviation of each hinge point, Δ l i This represents the initial length deviation of the electric cylinder. R Pj For the first j Under each measurement configuration { P}arrive{ G The rotation matrix of} For the first jGroup measurement data i Error of one branch, For the first j Error vectors for all branches of the measurement data set.

[0009] Preferably, the end-effector pose measurement residual includes position residual and attitude residual; the attitude residual is obtained through a logarithmic mapping on SO(3), satisfying:

[0010] in, and These are the measured and predicted values ​​of the position vector of the system's end effector, respectively. and These are the measured and predicted values ​​of the attitude vector of the system's end effector, respectively. For the first j The group of measurement data includes the position error of the end effector. For the first j Group measurement data: attitude error of the end effector.

[0011] Preferably, the uniform linearization adopts a right-multiplication perturbation model, which introduces right perturbation increments into the rotation of the platform attitude, installation attitude and end attitude respectively, and derives the joint identification Jacobian matrix under the same Lie algebra perturbation framework.

[0012] Preferably, the weighted nonlinear least squares objective function includes:

[0013] in w l , w p , w r These are the weighting coefficients. x The vector of parameters to be estimated contains global error parameters and implicit local pose variables for each measurement configuration. m This represents the total number of observations.

[0014] Preferably, an iterative optimization algorithm is used to jointly solve for the global error parameters and the implicit local pose variables of each measurement configuration; the iterative optimization algorithm includes the Levenberg-Marquardt algorithm, whose iterative update satisfies:

[0015] in J The global Jacobian matrix. E For the global residual vector, k The damping factor, k For the number of iterations, I It is the identity matrix. right This represents the correction amount for the parameter to be calibrated.

[0016] Preferably, in step S3, a soft anchoring constraint is applied to the implicit local pose variables of the first set of measurement configurations, so that the cost term for its deviation from the preset reference pose is added to the least squares objective function to eliminate implicit local pose drift.

[0017] A performance-aware motion planning system for a serial-parallel hybrid redundant robotic arm, provided by the present invention, includes: Module M1: Establishes the ideal kinematic model of the serial-parallel hybrid robot and defines multiple coordinate systems; Module M2: Collects multiple sets of measurement data under a preset excitation trajectory, and introduces the pose of the parallel linkage platform corresponding to each set of measurement data as an implicit local pose variable based on the ideal kinematic model of the serial-parallel hybrid robot, and constructs the parallel branch constraint residual and the end pose measurement residual. Module M3: The residuals are uniformly linearized on SE(3) to obtain the joint identification Jacobian matrix of global error parameters and implicit local pose variables. Then, a weighted nonlinear least squares objective function is constructed to solve the global error parameters and implicit local pose variables of each measurement configuration to obtain the calibration results. Module M4: Updates the ideal kinematic model of the serial-parallel hybrid robot using calibration results and performs error compensation.

[0018] Preferably, the plurality of coordinate systems include the parallel mechanism moving platform coordinate system {P}, the base coordinate system {G}, the serial manipulator base coordinate system {B}, and the end effector coordinate system {F}; the plurality of sets of measurement data include the measured actuator variables of each drive branch of the parallel mechanism, the joint variables of the serial manipulator, and the measured pose of the end effector in the base coordinate system {G}.

[0019] Preferably, the module M2 includes constructing an error parameter set; the error parameter set includes the position deviation of the hinge point of the parallel mechanism, the offset of the parallel actuator, the geometric parameter deviation of the serial manipulator, and the installation posture error between the coordinate system {P} of the moving platform of the parallel mechanism and the coordinate system {B} of the base of the serial manipulator.

[0020] Preferably, the parallel branch constraint residual is constructed in the following ways: For the first j Group measurement data and the first i Branch chain to measure actuator length With predicted length The difference is used as the residual, and the predicted length takes into account the hinge point deviation and actuator offset, satisfying:

[0021] in rAi For the first time on the platform i One hinge point A i exist{ P The nominal position vector in} r Bi For the next platform i One hinge point B i exist{ G The nominal position vector in} p j For the first j Under a measurement configuration, the coordinate system of the parallel mechanism moving platform { P}Origin relative to { G The translation vector of}, Δ A i Δ B i The upper and lower platforms are respectively i The positional deviation of each hinge point, Δ l i This represents the initial length deviation of the electric cylinder. R Pj For the first j Under each measurement configuration { P}arrive{ G The rotation matrix of} For the first j Group measurement data i Error of one branch, For the first j Error vectors for all branches of the measurement data set.

[0022] Preferably, the end-effector pose measurement residual includes position residual and attitude residual; the attitude residual is obtained through a logarithmic mapping on SO(3), satisfying:

[0023] in, and These are the measured and predicted values ​​of the position vector of the system's end effector, respectively. and These are the measured and predicted values ​​of the attitude vector of the system's end effector, respectively. For the first j The group of measurement data includes the position error of the end effector. For the first j Group measurement data: attitude error of the end effector.

[0024] Preferably, the uniform linearization adopts a right-multiplication perturbation model, which introduces right perturbation increments into the rotation of the platform attitude, installation attitude and end attitude respectively, and derives the joint identification Jacobian matrix under the same Lie algebra perturbation framework.

[0025] Preferably, the weighted nonlinear least squares objective function includes:

[0026] in w l , w p , w r These are the weighting coefficients. x The vector of parameters to be estimated contains global error parameters and implicit local pose variables for each measurement configuration. m This represents the total number of observations.

[0027] Preferably, an iterative optimization algorithm is used to jointly solve for the global error parameters and the implicit local pose variables of each measurement configuration; the iterative optimization algorithm includes the Levenberg-Marquardt algorithm, whose iterative update satisfies:

[0028] in J The global Jacobian matrix. E For the global residual vector, k The damping factor, k For the number of iterations, I It is the identity matrix. right This represents the correction amount for the parameter to be calibrated.

[0029] Preferably, in module M3, a soft anchoring constraint is applied to the implicit local pose variables of the first set of measurement configurations, so that the cost term for its deviation from the preset reference pose is added to the least squares objective function to eliminate implicit local pose drift.

[0030] Compared with the prior art, the present invention has the following beneficial effects: 1. This invention does not require explicit parallel forward kinematics. Instead, it absorbs the unknown pose of the parallel platform through implicit local pose variables, thereby reducing the complexity of each configuration forward solution iteration.

[0031] 2. The rotational linearization of the present invention is consistent. Based on the uniform processing of rotational quantities by right perturbation of Lie group, geometrically consistent Jacobi is obtained, which improves convergence and stability.

[0032] 3. The present invention has strong joint calibration capability, and can simultaneously identify parallel, series and installation errors, avoiding the error propagation of step-by-step calibration.

[0033] 4. The present invention has strong applicability and is applicable to serial-parallel structures such as Stewart platform + robotic arm and other parallel / serial hybrid systems, and can be used with a variety of external measuring devices. Attached Figure Description

[0034] Other features, objects, and advantages of the present invention will become more apparent from the following detailed description of non-limiting embodiments with reference to the accompanying drawings: Figure 1 This is a schematic diagram of the structure and coordinate system definition of the serial and parallel robot system of the present invention.

[0035] Figure 2 This is a schematic diagram illustrating the kinematic model and geometric parameter definitions of the series-parallel hybrid system of the present invention.

[0036] Figure 3 This is a flowchart of the calibration process of the present invention.

[0037] Figure 4 This is a schematic diagram illustrating the sensitivity analysis of the excitation signal involved and the parameters to be calibrated in this invention. Detailed Implementation

[0038] The present invention will now be described in detail with reference to specific embodiments. These embodiments will help those skilled in the art to further understand the present invention, but do not limit the invention in any way. It should be noted that those skilled in the art can make several changes and improvements without departing from the concept of the present invention. These all fall within the protection scope of the present invention.

[0039] like Figure 3 As shown, an implicit local pose calibration method for a serial-parallel hybrid robot is disclosed, wherein the robot includes a parallel mechanism and a serial robotic arm mounted on the moving platform of the parallel mechanism, comprising the following steps: a) Establish an ideal kinematic model of the serial-parallel hybrid robot, and define the coordinate system of the moving platform of the parallel mechanism { P} and base coordinate system { G}, and the serial robotic arm base coordinate system { B} and the end coordinate system { F}; b) Given an excitation trajectory, collect multiple sets of measurement data, including at least: measured actuator variables of each drive chain of the parallel mechanism, joint variables of the serial robotic arm, and end effector data in the global coordinate system. G The end effector's measured pose is obtained through an external measurement system, which includes, but is not limited to, laser tracking measurement systems and optical motion capture systems. Theoretically, the higher the measurement accuracy of the external measurement system, the more accurate the calibration result.

[0040] c) Construct an error parameter set, which includes at least: the hinge point position deviation of the parallel mechanism, the offset of the parallel actuator, the geometric parameter deviation of the serial robotic arm, and { P}and{ B Installation pose error between}; d) Introduce the platform pose corresponding to each set of measurement data as an implicit local pose variable, and construct the parallel branch constraint residual and the end-effector pose measurement residual; the implicit local pose variable includes the first... j Translation vector of the moving platform under a measurement configuration p j With attitude vector Pj .

[0041] Parallel branch constraint residuals are constructed as follows: For the first j Group measurement data and the first i Branch chain to measure actuator length With predicted length The difference is used as the residual, and the predicted length takes into account the hinge point deviation and actuator offset, satisfying:

[0042] in r Ai For the first time on the platform i One hinge point A i exist{ P The nominal position vector in} r Bi For the next platform i One hinge point B i exist{ G The nominal position vector in} p j No. j Under a measurement configuration, the coordinate system of the moving platform { P}Origin relative to { G The translation vector of}, Δ A i Δ B i The upper and lower platforms are respectively i The positional deviation of each hinge point, Δ l i This represents the initial length deviation of the electric cylinder. R Pj No. j Under each measurement configuration { P}arrive{ G The rotation matrix of} For the first j Group measurement datai Error of one branch, For the first j Error vectors for all branches of the measurement data set.

[0043] The end-effector pose measurement residuals include position residuals and attitude residuals, where the attitude residuals are obtained through a logarithmic mapping over SO(3), satisfying:

[0044] in, and These are the measured and predicted values ​​of the position vector of the system's end effector, respectively. and These are the measured and predicted values ​​of the attitude vector of the system's end effector, respectively. For the first j The group of measurement data includes the position error of the end effector. For the first j Group measurement data: attitude error of the end effector.

[0045] e) Perform uniform linearization on the residuals on SE(3) to obtain the joint identification Jacobian matrix for global error parameters and implicit local pose variables; the uniform linearization adopts the right multiplication perturbation model, introduces right perturbation increments for the rotation of platform attitude, installation attitude and end attitude respectively, and derives the Jacobian matrix under the same Lie algebra perturbation framework.

[0046] f) Construct a weighted nonlinear least squares objective function, and use an iterative optimization algorithm to jointly solve for the global error parameters and the implicit local pose variables of each measurement configuration to obtain the calibration results; the weighted nonlinear least squares objective function is:

[0047] in w l , w p , w r These are the weighting coefficients. x The vector of parameters to be estimated contains global error parameters and implicit local pose variables for each measurement configuration. m This represents the total number of observations.

[0048] The iterative optimization algorithm is the Levenberg-Marquardt algorithm, and its iterative updates satisfy:

[0049] in J The global Jacobian matrix. E For the global residual vector, kThe damping factor, k For the number of iterations, I It is the identity matrix. right This represents the correction amount for the parameter to be calibrated.

[0050] To eliminate implicit local pose drift, soft anchoring constraints are applied to the implicit local pose variables of the first set of measurement configurations in this step, and the cost term for its deviation from the preset reference pose is added to the least squares objective function.

[0051] g) Use the calibration results to update the robot's kinematics model and perform error compensation.

[0052] like Figure 1 and Figure 2 As shown, in one embodiment, the above steps include: Step 1: Data Collection exist j =1… N Parallel drive quantity is collected at each measurement configuration location. Serial joint quantity End position obtained by external measurement system and posture .

[0053] Step 2: Establish the nominal kinematic model and error parameters The parallel section uses hinge point geometry parameters r Ai and r Bi The branch connection point is indicated; the series connection point uses its set of geometric parameters. f In addition, error parameters are also considered, including: the deviation of the upper platform hinge point Δ. A i The deviation of the lower platform hinge point Δ B i driver zero bias Δ l i Installation error Δ ψ =[Δ p B , Δ B ], Series geometric parameter error Δ f .

[0054] Step 3: Introduce implicit local pose variables For each measurement configuration j The parallel platform pose ( p j , Pj ) as a local variable, where:

[0055] This local variable is not obtained by solving the parallel forward kinematics, but is estimated in the optimization through residual constraints.

[0056] Step 4: Residual Construction 1) Parallel branch length residuals:

[0057] in p j For the first j The coordinate system of the moving platform under the measurement configuration { P}Origin relative to { G The translation vector of} R Pj No. j Under each measurement configuration { P}arrive{ G The rotation matrix of}. 2) End-effector pose residual: Obtained from serial positive kinematics The predicted end-effector pose is obtained by combining the installation transformation and the platform transformation. ,structure

[0058] The total residuals are constructed as follows:

[0059] 3) Definition of parameters to be calibrated:

[0060] in, p j and Pj The first j The position vector and attitude vector of the moving platform are measured in this second measurement.

[0061] 4) Weighted least squares objective

[0062] in w l , w p , w r These are the weighting coefficients. x The vector of parameters to be estimated contains global error parameters and implicit local pose variables for each measurement configuration. m This represents the total number of observations.

[0063] 5) Uniform linearization of Lie groups and Jacobi construction A uniform right perturbation is applied to the rotation:

[0064] Based on this, the length residual and pose residual are linearized to the first order, resulting in:

[0065] in J To identify the Jacobian matrix, right It includes global error parameter increments and increments of local variables in each shape. This uniform linearization ensures geometric consistency in the handling of rotational quantities and allows parallel branch constraints and serial end-position pose constraints to be jointly established within the same framework.

[0066] 6) Levenberg-Marquardt iterative solution In the k Solve in the next iteration:

[0067] in J The global Jacobian matrix. E For the global residual vector, k The damping factor, k This represents the number of iterations. Update the parameters until convergence, output the calibration parameters, and update the robot model.

[0068] Example 1 This embodiment uses a typical series-parallel hybrid system as an example for illustration: The parallel section is a six-branch UPS-type Stewart platform, and the serial section is a seven-DOF serial robotic arm mounted on the parallel platform. Several optical markers are installed on the end effector, and the end effector's pose in the global coordinate system is acquired through an external measurement system. To ensure that all identifiable parameters are adequately excited, reduce parameter correlation, and improve the Jacobian condition number, this embodiment applies continuous multi-sine excitations to both the parallel and serial sections simultaneously during the calibration data acquisition phase. This allows for rich variations in platform translation and attitude within the workspace, while each serial joint completes independent excitation within its permissible range. For the first... i A parallel platform electric cylinder, whose length excitation function is:

[0069] in l i0 For the initial length, a ik For the first k Each sinusoidal component amplitude T ik For a period of time, bik For phase.

[0070] For the i A robotic arm joint, whose joint position excitation function is:

[0071] in q i0 The initial joint angle, a ik For the first k Each sinusoidal component amplitude T ik For a period of time, b ik Phase is used. Each joint uses independent frequency and amplitude combinations to improve excitation richness and reduce coupling between parameters.

[0072] The excitation signal involves sensitivity analysis results of the parameters to be calibrated, as follows: Figure 4 As shown, the system is full rank, and all parameters to be calibrated are observable.

[0073] Measurement data is collected and timestamp aligned during the execution of the excitation trajectory. To verify the stability and repeatability of the method, this embodiment can repeat the acquisition multiple times under the same excitation signal to obtain multiple independent datasets, which are then calibrated and solved separately. Statistical analysis is performed on the variance and consistency of the calibration results.

[0074] The present invention also provides a performance-aware motion planning system for a serial-parallel hybrid redundant robotic arm. The performance-aware motion planning system for the serial-parallel hybrid redundant robotic arm can be implemented by executing the process steps of the performance-aware motion planning method for the serial-parallel hybrid redundant robotic arm. That is, those skilled in the art can understand the performance-aware motion planning method for the serial-parallel hybrid redundant robotic arm as a preferred embodiment of the performance-aware motion planning system for the serial-parallel hybrid redundant robotic arm.

[0075] Specifically, a performance-aware motion planning system for a serial-parallel hybrid redundant robotic arm includes: Module M1: Establishes the ideal kinematic model of the serial-parallel hybrid robot and defines multiple coordinate systems; Module M2: Collects multiple sets of measurement data under a preset excitation trajectory, and introduces the pose of the parallel linkage platform corresponding to each set of measurement data as an implicit local pose variable based on the ideal kinematic model of the serial-parallel hybrid robot, and constructs the parallel branch constraint residual and the end pose measurement residual. Module M3: The residuals are uniformly linearized on SE(3) to obtain the joint identification Jacobian matrix of global error parameters and implicit local pose variables. Then, a weighted nonlinear least squares objective function is constructed to solve the global error parameters and implicit local pose variables of each measurement configuration to obtain the calibration results. Module M4: Updates the ideal kinematic model of the serial-parallel hybrid robot using calibration results and performs error compensation.

[0076] The multiple coordinate systems include the parallel mechanism moving platform coordinate system {P}, the base coordinate system {G}, the serial manipulator base coordinate system {B}, and the end effector coordinate system {F}; the multiple sets of measurement data include the measured actuator variables of each drive branch of the parallel mechanism, the joint variables of the serial manipulator, and the measured pose of the end effector in the base coordinate system {G}.

[0077] The module M2 includes constructing an error parameter set; the error parameter set includes the position deviation of the hinge point of the parallel mechanism, the offset of the parallel actuator, the geometric parameter deviation of the serial manipulator, and the installation posture error between the coordinate system {P} of the moving platform of the parallel mechanism and the coordinate system {B} of the base of the serial manipulator.

[0078] The construction methods for the parallel branch constraint residuals include: For the first j Group measurement data and the first i Branch chain to measure actuator length With predicted length The difference is used as the residual, and the predicted length takes into account the hinge point deviation and actuator offset, satisfying:

[0079] in r Ai For the first time on the platform i One hinge point A i exist{ P The nominal position vector in} r Bi For the next platform i One hinge point B i exist{ G The nominal position vector in} p j For the first j Under a measurement configuration, the coordinate system of the parallel mechanism moving platform { P}Origin relative to { G The translation vector of}, Δ A i Δ B i The upper and lower platforms are respectivelyi The positional deviation of each hinge point, Δ l i This represents the initial length deviation of the electric cylinder. R Pj For the first j Under each measurement configuration { P}arrive{ G The rotation matrix of} For the first j Group measurement data i Error of one branch, For the first j Error vectors for all branches of the measurement data set.

[0080] The end-effector pose measurement residual includes position residual and attitude residual; the attitude residual is obtained through a logarithmic mapping on SO(3), satisfying:

[0081] in, and These are the measured and predicted values ​​of the position vector of the system's end effector, respectively. and These are the measured and predicted values ​​of the attitude vector of the system's end effector, respectively. For the first j The group of measurement data includes the position error of the end effector. For the first j Group measurement data: attitude error of the end effector.

[0082] Uniform linearization employs a right-multiplication perturbation model, introducing right perturbation increments into the rotation amounts of the platform attitude, installation attitude, and end attitude, and deriving the joint identification Jacobian matrix within the same Lie algebra perturbation framework.

[0083] The weighted nonlinear least squares objective function includes:

[0084] in w l , w p , w r These are the weighting coefficients. x The vector of parameters to be estimated contains global error parameters and implicit local pose variables for each measurement configuration. m This represents the total number of observations.

[0085] An iterative optimization algorithm is used to jointly solve for the global error parameters and the implicit local pose variables of each measurement configuration; the iterative optimization algorithm includes the Levenberg-Marquardt algorithm, whose iterative update satisfies:

[0086] in J The global Jacobian matrix. E For the global residual vector, k The damping factor, k For the number of iterations, I It is the identity matrix. right This represents the correction amount for the parameter to be calibrated.

[0087] In module M3, a soft anchoring constraint is applied to the implicit local pose variables of the first set of measurement configurations, and the cost term for its deviation from the preset reference pose is added to the least squares objective function to eliminate implicit local pose drift.

[0088] Those skilled in the art will understand that, besides implementing the system and its various devices, modules, and units provided by this invention in the form of purely computer-readable program code, the same functions can be achieved entirely through logical programming of the method steps, making the system and its various devices, modules, and units of this invention function in the form of logic gates, switches, application-specific integrated circuits, programmable logic controllers, and embedded microcontrollers. Therefore, the system and its various devices, modules, and units provided by this invention can be considered as a hardware component, and the devices, modules, and units included therein for implementing various functions can also be considered as structures within the hardware component; alternatively, the devices, modules, and units for implementing various functions can be considered as both software modules implementing the method and structures within the hardware component.

[0089] Specific embodiments of the present invention have been described above. It should be understood that the present invention is not limited to the specific embodiments described above, and those skilled in the art can make various changes or modifications within the scope of the claims, which do not affect the essence of the present invention. Unless otherwise specified, the embodiments and features described in this application can be arbitrarily combined with each other.

Claims

1. A performance-aware motion planning method for a serial-parallel hybrid redundant robotic arm, characterized in that, include: Step S1: Establish an ideal kinematic model of the serial-parallel hybrid robot and define multiple coordinate systems; Step S2: Collect multiple sets of measurement data under the preset excitation trajectory, and based on the ideal kinematic model of the serial-parallel hybrid robot, introduce the parallel platform pose corresponding to each set of measurement data as an implicit local pose variable to construct the parallel branch constraint residual and the end pose measurement residual. Step S3: Perform uniform linearization on SE(3) to obtain the joint identification Jacobian matrix of global error parameters and implicit local pose variables, and then construct a weighted nonlinear least squares objective function to solve for global error parameters and implicit local pose variables of each measurement configuration to obtain calibration results. Step S4: Update the ideal kinematic model of the serial-parallel hybrid robot using the calibration results and perform error compensation.

2. The performance-aware motion planning method for a serial-parallel hybrid redundant robotic arm according to claim 1, characterized in that, The multiple coordinate systems include the parallel mechanism moving platform coordinate system {P}, the base coordinate system {G}, the serial manipulator base coordinate system {B}, and the end effector coordinate system {F}; the multiple sets of measurement data include the measured actuator variables of each drive branch of the parallel mechanism, the joint variables of the serial manipulator, and the measured pose of the end effector in the base coordinate system {G}.

3. The performance-aware motion planning method for a serial-parallel hybrid redundant robotic arm according to claim 2, characterized in that, Step S2 includes constructing an error parameter set; the error parameter set includes the position deviation of the hinge point of the parallel mechanism, the offset of the parallel actuator, the geometric parameter deviation of the serial manipulator, and the installation posture error between the coordinate system {P} of the moving platform of the parallel mechanism and the coordinate system {B} of the base of the serial manipulator.

4. The performance-aware motion planning method for a serial-parallel hybrid redundant robotic arm according to claim 2, characterized in that, The construction methods for the parallel branch constraint residuals include: For the first j Group measurement data and the first i Branch chain to measure actuator length With predicted length The difference is used as the residual, and the predicted length takes into account the hinge point deviation and actuator offset, satisfying: in r Ai For the first time on the platform i One hinge point A i exist{ P The nominal position vector in} r Bi For the next platform i One hinge point B i exist{ G The nominal position vector in} p j For the first j Under a measurement configuration, the coordinate system of the parallel mechanism moving platform { P }Origin relative to { G The translation vector of}, Δ A i Δ B i The upper and lower platforms are respectively i The positional deviation of each hinge point, Δ l i This represents the initial length deviation of the electric cylinder. R Pj For the first j Under each measurement configuration { P }arrive{ G The rotation matrix of} For the first j Group measurement data i Error of one branch, For the first j Error vectors for all branches of the measurement data set.

5. The performance-aware motion planning method for a serial-parallel hybrid redundant robotic arm according to claim 4, characterized in that, The end-effector pose measurement residual includes position residual and attitude residual; the attitude residual is obtained through a logarithmic mapping on SO(3), satisfying: in, and These are the measured and predicted values ​​of the position vector of the system's end effector, respectively. and These are the measured and predicted values ​​of the attitude vector of the system's end effector, respectively. For the first j The group of measurement data includes the position error of the end effector. For the first j Group measurement data: attitude error of the end effector.

6. The performance-aware motion planning method for a serial-parallel hybrid redundant robotic arm according to claim 1, characterized in that, Uniform linearization employs a right-multiplication perturbation model, introducing right perturbation increments into the rotation amounts of the platform attitude, installation attitude, and end attitude, and deriving the joint identification Jacobian matrix within the same Lie algebra perturbation framework.

7. The performance-aware motion planning method for a serial-parallel hybrid redundant robotic arm according to claim 5, characterized in that, The weighted nonlinear least squares objective function includes: in w l , w p , w r These are the weighting coefficients. ξ The vector of parameters to be estimated contains global error parameters and implicit local pose variables for each measurement configuration. m This represents the total number of observations.

8. The performance-aware motion planning method for a serial-parallel hybrid redundant robotic arm according to claim 1, characterized in that, An iterative optimization algorithm is used to jointly solve for the global error parameters and the implicit local pose variables of each measurement configuration; the iterative optimization algorithm includes the Levenberg-Marquardt algorithm, whose iterative update satisfies: in J The global Jacobian matrix. E For the global residual vector, κ The damping factor, k For the number of iterations, I It is the identity matrix. δξ This represents the correction amount for the parameter to be calibrated.

9. The performance-aware motion planning method for a serial-parallel hybrid redundant robotic arm according to claim 1, characterized in that, In step S3, a soft anchoring constraint is applied to the implicit local pose variables of the first set of measurement configurations, so that the cost term of its deviation from the preset reference pose is added to the least squares objective function to eliminate implicit local pose drift.

10. A performance-sensing motion planning system for a serial-parallel hybrid redundant robotic arm, characterized in that, include: Module M1: Establishes the ideal kinematic model of the serial-parallel hybrid robot and defines multiple coordinate systems; Module M2: Collects multiple sets of measurement data under a preset excitation trajectory, and introduces the pose of the parallel linkage platform corresponding to each set of measurement data as an implicit local pose variable based on the ideal kinematic model of the serial-parallel hybrid robot, and constructs the parallel branch constraint residual and the end pose measurement residual. Module M3: The residuals are uniformly linearized on SE(3) to obtain the joint identification Jacobian matrix of global error parameters and implicit local pose variables. Then, a weighted nonlinear least squares objective function is constructed to solve the global error parameters and implicit local pose variables of each measurement configuration to obtain the calibration results. Module M4: Updates the ideal kinematic model of the serial-parallel hybrid robot using calibration results and performs error compensation.

Images