A robot parameter identification method and device, electronic equipment and storage medium

Abstract

Embodiments of the present application relate to the field of industrial automation, and disclose a robot parameter identification method and device, electronic equipment and a storage medium. In the present application, motion data of each joint of the robot is acquired; the viscous friction and the joint motion speed with a power in a friction force calculation formula of a nonlinear friction force model are made to present a linear relationship; after linearization of the adjusted nonlinear friction force model, the dynamic parameters vector of each joint is identified by least square method according to rigid body dynamics; measurement data is acquired from the measured torque, and the estimated friction force is calculated; and the power of the joint speed in the friction force calculation formula of the nonlinear friction force model is updated by using a pattern search algorithm. The rigid body dynamics parameters and the nonlinear friction force parameters of the robot are identified at the same time, and the identification accuracy of the dynamics parameters of the robot is improved.

Description

Technical Field

[0001] The embodiments of the present invention relate to the field of industrial automation, and in particular to a method, apparatus, electronic device and storage medium for robot parameter identification. Background Technology

[0002] In practical applications, most industrial robots are equipped with speed reducers to achieve deceleration and power amplification, but this also increases the friction at the joint ends of the robotic arm. The friction at the joint ends has a significant impact on the dynamic behavior of the robotic arm; therefore, a friction model needs to be incorporated into the dynamic modeling of the robotic arm to better describe its dynamic behavior. A commonly used friction model is a linear friction model composed of Coulomb friction and viscous friction. Since the friction parameters at the joint ends of the robotic arm can often only be obtained through identification, a common identification method is to combine the linear friction model with the rigid body dynamics model of the robotic arm, linearize the model, and finally identify all parameters using the least squares method. However, recent research shows that the linear friction model cannot well describe the joint friction characteristics at reversal points and low speeds, and that viscous friction has a non-linear relationship with joint motion speed.

[0003] The inventors have discovered at least the following problems in the related technologies: Since the nonlinear friction force model cannot be linearized, it is necessary to identify the nonlinear friction force parameters and rigid body dynamic parameters of each robotic arm in the robot separately. This process is highly repetitive, consumes a lot of computational resources, and also causes cumulative errors, resulting in a decrease in identification accuracy. Summary of the Invention

[0004] The purpose of this invention is to provide a robot parameter identification method, device, electronic device, and storage medium, which enables the simultaneous identification of rigid body dynamic parameters and nonlinear friction force parameters of the robotic arm, and improves the identification accuracy of the robotic arm's dynamic parameters.

[0005] To address the aforementioned technical problems, embodiments of the present invention provide a robot parameter identification method, comprising: acquiring motion data of each joint of a robot; wherein the motion data includes joint motion velocity, joint position, and joint current of each joint of the robot; constructing a nonlinear friction model based on the motion data of each joint of the robot; adjusting the value of the power of the joint motion velocity in the friction calculation formula of the nonlinear friction model to make the viscous friction and the joint motion velocity with the power in the friction calculation formula of the nonlinear friction model linearly related; linearizing the adjusted nonlinear friction model, and identifying the dynamic parameter vector of each joint of the robot in the linearized nonlinear friction model using the least squares method according to rigid body dynamics; acquiring measurement data from the measured torque of the robot, and calculating and estimating the friction force based on the measurement data, inertial parameters, and rigid body dynamics; and updating the value of the power of the joint motion velocity in the friction calculation formula of the nonlinear friction model using a pattern search algorithm based on the linearized nonlinear friction model and the estimated friction force.

[0006] The present invention also provides a robot parameter identification device, comprising: a data acquisition module for acquiring motion data of each joint of a robot; wherein the motion data includes joint motion velocity, joint position, and joint current of each joint of the robot; a model degradation module for constructing a nonlinear friction model based on the motion data of each joint of the robot, adjusting the value of the power of the joint motion velocity in the friction calculation formula of the nonlinear friction model, so that the viscous friction in the friction calculation formula of the nonlinear friction model and the joint motion velocity with the power exhibit a linear relationship; and a dynamic parameter calculation module for... After linearizing the adjusted nonlinear friction model, the dynamic parameter vectors of each joint of the robot in the linearized nonlinear friction model are identified by the least squares method according to rigid body dynamics. The friction estimation module is used to obtain measurement data from the measured torque of the robot and calculate and estimate the friction force according to the measurement data, inertial parameters and rigid body dynamics. The parameter optimization module is used to update the value of the joint motion velocity power in the friction calculation formula of the nonlinear friction model by using a pattern search algorithm based on the linearized nonlinear friction model and the estimated friction force.

[0007] Embodiments of the present invention also provide an electronic device, comprising: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor, the instructions being executed by the at least one processor to enable the at least one processor to perform the robot parameter identification method described above.

[0008] Embodiments of the present invention also provide a computer-readable storage medium storing a computer program, which, when executed by a processor, implements the robot parameter identification method described above.

[0009] In this embodiment of the invention, motion data of each joint of the robot is acquired; wherein, the motion data includes the joint motion velocity, joint position, and joint current of each joint of the robot; a nonlinear friction model is constructed based on the motion data of each joint of the robot; the value of the power of the joint motion velocity in the friction calculation formula of the nonlinear friction model is adjusted so that the viscous friction and the joint motion velocity with the power in the friction calculation formula of the nonlinear friction model have a linear relationship; after linearizing the adjusted nonlinear friction model, the dynamic parameter vector of each joint of the robot in the linearized nonlinear friction model is identified by least squares method according to rigid body dynamics; measurement data is obtained from the measured torque of the robot, and the friction force is estimated based on the measurement data, inertial parameters, and rigid body dynamics; based on the linearized nonlinear friction model and the estimated friction force, a pattern search algorithm is used to update the value of the power of the joint motion velocity in the friction calculation formula of the nonlinear friction model. This invention addresses the issues of high operational repetition and excessive computational resource consumption caused by the need to separately identify the nonlinear friction parameters and rigid body dynamic parameters of each robotic arm in a nonlinear friction model. It achieves simultaneous identification of the rigid body dynamic parameters and nonlinear friction parameters of the robotic arm, while improving the identification accuracy of the robotic arm's dynamic parameters.

[0010] Furthermore, after linearizing the adjusted nonlinear friction model, the dynamic parameter vectors of each robot joint in the linearized nonlinear friction model are identified using the least squares method based on rigid body dynamics. This includes: combining the rigid body dynamics model of the robotic arm generated based on rigid body dynamics with the linearized nonlinear friction model and describing it as a linear regression form; calculating the observation matrix and observation response matrix based on the dynamic model described as a linear regression form; and identifying the dynamic parameter vectors of each robot joint using the least squares method with the observation matrix and observation response matrix. This more accurately obtains all the dynamic parameters of each robot joint in a single identification, further avoiding the cumulative error problem that is easily caused by step-by-step identification.

[0011] Furthermore, the calculation of the observation matrix and observation response matrix based on the dynamic model described in linear regression form includes: using QR decomposition to eliminate irrelevant variables in the dynamic model described in linear regression form, and recombining the remaining parameters to form a minimal set of inertial parameters; describing the minimal set of inertial parameters in linear form; and calculating the linear regression observation matrix and observation response torque vector of the dynamic model described in linear regression form based on the minimal set of inertial parameters in linear form. Further optimization is made to address the occasional case where the observation matrix cannot reach full rank, eliminating parameters that have no impact on the dynamic model, improving operational efficiency while also enhancing the accuracy of dynamic parameter identification for the robotic arm.

[0012] In addition, the dynamic parameter vector is composed of the inertial parameters, gravity parameters, equivalent rotational parameters of the joint transmission chain, and friction parameters of each joint of the robot.

[0013] Furthermore, the step of updating the power of the joint motion velocity in the friction calculation formula of the nonlinear friction model using a pattern search algorithm based on the linearized nonlinear friction model and the estimated friction force includes: using the pattern search algorithm, taking the minimum value of the vector L2 norm of the friction calculation formula in the linearized nonlinear friction model and the estimated friction force as the criterion, and using the friction parameters in the dynamic parameter vector as the initial solution of the pattern search algorithm, optimizing the friction parameters in the dynamic parameter vector to obtain the global optimal solution of the pattern search algorithm; and updating the power of the joint motion velocity in the friction calculation formula of the nonlinear friction model based on the global optimal solution of the pattern search algorithm. This quickly locks the new power and updates the dynamic model, further improving the accuracy of the friction parameters in the nonlinear friction model.

[0014] Furthermore, adjusting the value of the power of the joint motion velocity in the friction calculation formula of the nonlinear friction model includes: adjusting the value of the power of the joint motion velocity in the friction calculation formula of the nonlinear friction model to 1. This reduces resource waste caused by the confirmation process.

[0015] In addition, acquiring motion data for each joint of the robot includes: controlling the robot's operating trajectory; collecting the actual position, velocity, and current information of the robot's motors; and converting the actual position, velocity, and current information of the motors into motion data for each joint of the robot after unit conversion and data filtering based on rigid body dynamics. The calculated joint information is more ideal and saves the cost of deploying acquisition devices at multiple joints, while also avoiding the impact of acquisition devices on the robot's joints. Attached Figure Description

[0016] One or more embodiments are illustrated by way of example with reference numerals in the accompanying drawings. These illustrations do not constitute a limitation on the embodiments. Elements with the same reference numerals in the drawings are denoted as similar elements. Unless otherwise stated, the figures in the drawings are not to be limited by scale.

[0017] Figure 1 This is a flowchart of a robot parameter identification method according to an embodiment of the present invention; Figure 2 This is a schematic diagram of the structure of a robot parameter identification device according to another embodiment of the present invention; Figure 3 This is a schematic diagram of the structure of an electronic device according to another embodiment of the present invention. Detailed Implementation

[0018] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the various embodiments of the present invention will be described in detail below with reference to the accompanying drawings. However, those skilled in the art will understand that many technical details are presented in the various embodiments of the present invention to facilitate a better understanding of this application. However, the technical solutions claimed in this application can be implemented even without these technical details and various changes and modifications based on the following embodiments. The division of the various embodiments below is for ease of description and should not constitute any limitation on the specific implementation of the present invention. The various embodiments can be combined with and referenced by each other without contradiction.

[0019] One embodiment of the present invention relates to a robot parameter identification method, which can be applied to robot control equipment, such as electronic devices like computers or robot-specific operation and control equipment. In this embodiment, motion data of each joint of the robot is acquired; wherein, the motion data includes the joint motion velocity, joint position, and joint current of each joint of the robot; a nonlinear friction model is constructed based on the motion data of each joint of the robot; the value of the power of the joint motion velocity in the friction calculation formula of the nonlinear friction model is adjusted so that the viscous friction and the joint motion velocity with the power in the friction calculation formula of the nonlinear friction model have a linear relationship; after linearizing the adjusted nonlinear friction model, the dynamic parameter vector of each joint of the robot in the linearized nonlinear friction model is identified by least squares method according to rigid body dynamics; measurement data is obtained from the measured torque of the robot, and the friction force is estimated based on the measurement data, inertial parameters, and rigid body dynamics; based on the linearized nonlinear friction model and the estimated friction force, a pattern search algorithm is used to update the value of the power of the joint motion velocity in the friction calculation formula of the nonlinear friction model. This invention addresses the problems of high operational repetition and excessive computational resource consumption caused by the need to separately identify the nonlinear friction parameters and rigid body dynamic parameters of each robotic arm in a nonlinear friction model. It achieves simultaneous identification of both the rigid body dynamic parameters and nonlinear friction parameters of the robotic arm, while improving the accuracy of the robotic arm's dynamic parameter identification. The implementation details of this robot parameter identification method are described below. These details are provided for ease of understanding and are not essential for implementing this solution.

[0020] like Figure 1As shown, in step 101, the robot control device acquires motion data of each joint of the robot. This motion data includes not only the joint motion velocity of each joint, but also more detailed data such as the actual position of the joint ends (joint position), joint current, and acceleration information. Motion data can be acquired directly by installing acquisition devices at the joints, or by using a traditional robot kinematics model to analyze the robot's excitation trajectory based on data acquired from the motor ends. Acquiring motion data of each joint using a rigid body kinematics model avoids the resource consumption caused by additional acquisition devices at the joints and the frictional errors caused by the weight of the acquisition devices themselves. In one example, acquiring motion data of each joint of the robot specifically involves: the robot control device controlling the robot to run the excitation trajectory; acquiring the actual position information, velocity information, and current information of the robot's motor ends; and converting the actual position information, velocity information, and current information of the motor ends into motion data of each joint of the robot after unit conversion and data filtering according to rigid body dynamics.

[0021] In step 102, a nonlinear friction model is constructed based on the motion data of each joint of the robot. The robot control device adjusts the value of the joint motion velocity to the power of the friction calculation formula in the nonlinear friction model. This process essentially uses the nonlinear friction model to fit the friction force to be approximated. For the i-th joint of the robot, the friction calculation formula of the nonlinear friction model is as follows:

[0022] in, Here are the parameters for Coulomb friction. These are parameters related to viscous friction. To account for the asymmetric bias of the positive and negative friction forces, The difference between the proposed nonlinear friction model and the traditional friction model lies in the fact that its viscous friction component is not linearly related to the joint motion velocity, but rather to some unknown factor related to the velocity. The exponentiation is linearly related. Therefore, it can be adjusted... The value of causes the nonlinear friction model to degenerate into a linear friction model. In this case, the viscous friction and the joint motion velocity with a power in the friction calculation formula of the nonlinear friction model exhibit a linear relationship. For the sake of simplifying the model, without needing to try multiple times, the value of the power of the joint motion velocity in the friction calculation formula of the nonlinear friction model can also be directly adjusted to 1, which will also make the model exhibit a linear relationship.

[0023] Since the model has degenerated into a linear model, in step 103, the robot control device can directly linearize the adjusted nonlinear friction model and then identify the dynamic parameter vectors of each joint of the robot in the linearized nonlinear friction model using the least squares method based on rigid body dynamics.

[0024] For solving linearized friction models, there is no single solution method. For linear models derived from nonlinear models, a relatively simple approach is to derive the model by calculating the observation matrix and the observation response matrix. Specifically: The rigid body dynamics model of the robotic arm generated based on rigid body dynamics is combined with the linearized nonlinear friction force model and described in a linear regression form:

[0025] In the formula For the number of joints, For linear regression observation matrix, This is an unknown dynamic parameter vector, representing a combination of all parameters in the robot's kinematic model. In this example, the dynamic parameter vector is chosen to consist of the inertial parameters, gravity parameters, equivalent rotational parameters of the joint transmission chains, and friction parameters of each joint. The parameters in the text represent the first... The inertial parameters, gravity parameters, equivalent rotational parameters of the joint transmission chain, and friction parameters of the joint.

[0026] in, , , , It is the friction force parameter matrix.

[0027] After describing the model in regression form, the observation matrix and observation response matrix are calculated based on the dynamic model described in linear regression form. In this example, the observation response matrix is ​​represented as an observation response torque vector. However, since some columns in the observation matrix may sometimes have linear relationships, this can lead to incomplete column rank, resulting in not all parameters being identifiable. To address these special cases, some optional improvements can be made to the calculation method. For example, this can be improved by eliminating parameters that have no impact on the dynamic model and recombining the remaining parameters to form a minimal set of inertial parameters, as detailed below: QR decomposition is used to eliminate irrelevant variables in the dynamic model described in linear regression form, and the remaining parameters are recombine to form a minimal set of inertial parameters. This minimal set of inertial parameters is then described in linear form. The linear form of the minimal inertial combination obtained by QR decomposition is as follows:

[0028] In the formula, To minimize the number of inertial parameters, This is a dynamic parameter vector, containing the minimum set of inertial parameters for the robotic arm and the linear friction parameters of the joints. This is the observation matrix.

[0029] Subsequently, based on the minimum inertia parameter set in the linear form, the linear regression observation matrix and the observed response torque vector of the dynamic model described in linear regression form are calculated, specifically: Assuming there are m data points, the linear form of the minimum inertia combination can be obtained by using QR decomposition. The observation matrix and observation response can be calculated, i.e.:

[0030]

[0031] In the formula It is the observation matrix. It is the observed response torque vector.

[0032] The dynamic parameter vectors of each joint of the robot are obtained by identifying the observation matrix and the observation response matrix using the least squares method. In this example, the linear regression observation matrix and the observation response torque vector are obtained by identifying them using the least squares method.

[0033] In the formula This is a vector of dynamic parameters for each joint of the robot, including the minimum set of dynamic inertial parameters of the robotic arm and the linear friction parameters of the joints.

[0034] In step 105, the robot control device acquires measurement data from the robot's measured torque, and calculates and estimates the frictional force based on the measurement data, inertial parameters, and rigid body dynamics, specifically as follows: The robot control device estimates the magnitude of friction by subtracting forces unrelated to friction (i.e., forces calculated from rigid body dynamics using inertial parameters) from the measured torque.

[0035] In the formula The frictional force estimated from the measured torque. This is the set of minimum inertial parameters identified by the least squares method.

[0036] In step 106, the robot control device, based on the linearized nonlinear friction model and the estimated friction, employs a pattern search algorithm to update the value of the power of the joint motion velocity in the friction calculation formula of the nonlinear friction model. In other words, in order to estimate... A pattern search algorithm was used to optimize the friction parameters. In selecting the pattern search algorithm, considering the regularity of friction parameters, the minimum value of the vector L2 norm of the difference between the proposed linearized nonlinear friction model and the actual estimated friction force is chosen as the criterion for optimizing the friction parameters. That is, for joints... :

[0037] In the formula The friction force calculation formula for the nonlinear friction force model is now known data, as the dynamic parameter vector has already been obtained. To measure the estimated frictional force in the torque, the frictional force parameters in the dynamic parameter vector are chosen as the initial solution. Optimizing these parameters ensures that the solution to the above problem eventually converges to the global optimum. After finding the global optimum, a new solution is obtained based on the global optimum obtained by the pattern search algorithm. And replace it in the dynamic model parameters.

[0038] In one example, to verify the effectiveness of the identification method, a six-axis robotic arm was run to verify the trajectory and collect the actual position, velocity, and current information of the motor ends. The collected motor end data was then converted to units and filtered to obtain the actual position, velocity, acceleration, and torque information of the joint ends. The identified dynamic parameters were then added to obtain the predicted torque value. The predicted torque value was then converted into the motor end current value. After repeated testing, the curves of the predicted current value and the actual current value showed a high degree of fit, proving that the robot parameter identification method proposed in this invention identifies the dynamic parameters with very high accuracy.

[0039] In this embodiment, motion data of each joint of the robot is acquired; wherein, the motion data includes the joint motion velocity, joint position, and joint current of each joint of the robot; a nonlinear friction model is constructed based on the motion data of each joint of the robot; the value of the power of the joint motion velocity in the friction calculation formula of the nonlinear friction model is adjusted so that the viscous friction and the joint motion velocity with the power in the friction calculation formula of the nonlinear friction model have a linear relationship; after linearizing the adjusted nonlinear friction model, the dynamic parameter vector of each joint of the robot in the linearized nonlinear friction model is identified by least squares method according to rigid body dynamics; measurement data is obtained from the measured torque of the robot, and the friction force is estimated based on the measurement data, inertial parameters, and rigid body dynamics; based on the linearized nonlinear friction model and the estimated friction force, a pattern search algorithm is used to update the value of the power of the joint motion velocity in the friction calculation formula of the nonlinear friction model. This invention addresses the issues of high operational repetition and excessive computational resource consumption caused by the need to separately identify the nonlinear friction parameters and rigid body dynamic parameters of each robotic arm in a nonlinear friction model. It achieves simultaneous identification of the rigid body dynamic parameters and nonlinear friction parameters of the robotic arm, while improving the identification accuracy of the robotic arm's dynamic parameters.

[0040] The steps described above are for clarity only. In practice, they can be combined into one step or some steps can be broken down into multiple steps. As long as they include the same logical relationship, they are all within the scope of protection of this patent. Adding insignificant modifications or introducing insignificant designs to the algorithm or process, but without changing the core design of the algorithm and process, are also within the scope of protection of this patent.

[0041] Another embodiment of the present invention relates to a robot parameter identification device, such as... Figure 2As shown, it includes: a data acquisition module 201, used to acquire motion data of each joint of the robot; wherein, the motion data includes joint motion velocity, joint position, and joint current of each joint of the robot; a model degradation module 202, used to construct a nonlinear friction model based on the motion data of each joint of the robot, and adjust the value of the power of the joint motion velocity in the friction calculation formula of the nonlinear friction model so that the viscous friction and the joint motion velocity with the power in the friction calculation formula of the nonlinear friction model present a linear relationship; and a dynamic parameter calculation module 203, used to convert the adjusted nonlinear friction model into a linear relationship. After linearizing the nonlinear friction model, the dynamic parameter vectors of each joint of the robot in the linearized nonlinear friction model are identified by the least squares method according to rigid body dynamics; the friction estimation module 204 is used to obtain measurement data from the measured torque of the robot, and calculate and estimate the friction force according to the measurement data, inertial parameters and rigid body dynamics; the parameter optimization module 205 is used to update the value of the joint motion velocity to the power of the friction force calculation formula of the nonlinear friction model according to the linearized nonlinear friction model and the estimated friction force, using a pattern search algorithm.

[0042] In one example, after linearizing the adjusted nonlinear friction model, the dynamic parameter vectors of each joint of the robot in the linearized nonlinear friction model are identified by the least squares method according to rigid body dynamics. This includes: combining the rigid body dynamics model of the robot arm generated according to rigid body dynamics with the linearized nonlinear friction model and describing it as a linear regression form; calculating the observation matrix and the observation response matrix according to the dynamic model described as a linear regression form; and identifying the dynamic parameter vectors of each joint of the robot by the least squares method using the observation matrix and the observation response matrix.

[0043] In one example, calculating the observation matrix and observation response matrix based on the dynamic model described in linear regression form includes: using QR decomposition to eliminate irrelevant variables in the dynamic model described in linear regression form, and recombining the remaining parameters to form a minimal set of inertia parameters; describing the minimal set of inertia in a linear form; and calculating the linear regression observation matrix and observation response torque vector of the dynamic model described in linear regression form based on the minimal set of inertia parameters in the linear form.

[0044] In one example, the dynamic parameter vector consists of the inertial parameters, gravity parameters, equivalent rotational parameters of the joint transmission chain, and friction parameters of each joint of the robot.

[0045] In one example, the step of updating the power of the joint motion velocity in the friction calculation formula of the nonlinear friction model using a pattern search algorithm based on the linearized nonlinear friction model and the estimated friction force includes: using a pattern search algorithm, taking the minimum value of the vector L2 norm of the friction calculation formula in the linearized nonlinear friction model and the estimated friction force as the criterion, and using the friction parameters in the dynamic parameter vector as the initial solution of the pattern search algorithm, optimizing the friction parameters in the dynamic parameter vector to obtain the global optimal solution of the pattern search algorithm; and updating the power of the joint motion velocity in the friction calculation formula of the nonlinear friction model based on the global optimal solution of the pattern search algorithm.

[0046] In one example, the value of the power of the joint motion velocity in the friction calculation formula of the adjusted nonlinear friction model includes: Adjust the value of the power of the joint motion velocity in the friction calculation formula of the nonlinear friction model to 1.

[0047] In one example, the data acquisition module 201 is specifically used to: control the robot's running excitation trajectory; collect the actual position information, velocity information, and current information of the robot's motor end; and convert the actual position information, velocity information, and current information of the motor end into motion data of each joint of the robot after unit conversion and data filtering according to the rigid body dynamics.

[0048] In this embodiment, motion data of each joint of the robot is acquired; wherein, the motion data includes the joint motion velocity, joint position, and joint current of each joint of the robot; a nonlinear friction model is constructed based on the motion data of each joint of the robot; the value of the power of the joint motion velocity in the friction calculation formula of the nonlinear friction model is adjusted so that the viscous friction and the joint motion velocity with the power in the friction calculation formula of the nonlinear friction model have a linear relationship; after linearizing the adjusted nonlinear friction model, the dynamic parameter vector of each joint of the robot in the linearized nonlinear friction model is identified by least squares method according to rigid body dynamics; measurement data is obtained from the measured torque of the robot, and the friction force is estimated based on the measurement data, inertial parameters, and rigid body dynamics; based on the linearized nonlinear friction model and the estimated friction force, a pattern search algorithm is used to update the value of the power of the joint motion velocity in the friction calculation formula of the nonlinear friction model. This invention addresses the issues of high operational repetition and excessive computational resource consumption caused by the need to separately identify the nonlinear friction parameters and rigid body dynamic parameters of each robotic arm in a nonlinear friction model. It achieves simultaneous identification of the rigid body dynamic parameters and nonlinear friction parameters of the robotic arm, while improving the identification accuracy of the robotic arm's dynamic parameters.

[0049] It is not difficult to see that this embodiment is a device embodiment corresponding to the above method embodiment, and this embodiment can be implemented in conjunction with the above method embodiment. The relevant technical details mentioned in the above method embodiment are still valid in this embodiment, and will not be repeated here to reduce repetition. Accordingly, the relevant technical details mentioned in this embodiment can also be applied to the above method embodiment.

[0050] It is worth mentioning that all modules involved in this embodiment are logical modules. In practical applications, a logical unit can be a physical unit, a part of a physical unit, or a combination of multiple physical units. Furthermore, to highlight the innovative aspects of this invention, this embodiment does not introduce units that are not closely related to solving the technical problem proposed by this invention; however, this does not mean that other units are absent from this embodiment.

[0051] Another embodiment of the present invention relates to a final electronic device, such as Figure 3As shown, it includes at least one processor 301; and a memory 302 communicatively connected to the at least one processor; wherein the memory 302 stores instructions executable by the at least one processor 301, the instructions being executed by the at least one processor 301 to enable the at least one processor 301 to perform the robot parameter identification method as described above.

[0052] The memory 302 and processor 301 are connected via a bus, which can include any number of interconnecting buses and bridges. The bus connects various circuits of one or more processors 301 and memory 302 together. The bus can also connect various other circuits, such as peripheral devices, voltage regulators, and power management circuits, which are well known in the art and therefore will not be described further herein. A bus interface provides an interface between the bus and the transceiver. The transceiver can be a single element or multiple elements, such as multiple receivers and transmitters, providing a unit for communicating with various other devices over a transmission medium. Data processed by processor 301 is transmitted over a wireless medium via an antenna, which further receives data and transmits it to processor 301.

[0053] Processor 301 is responsible for managing the bus and general processing, and can also provide various functions, including timing, peripheral interfaces, voltage regulation, power management, and other control functions. Memory 302 can be used to store data used by processor 301 during operation.

[0054] Another embodiment of the present invention relates to a computer-readable storage medium storing a computer program. When executed by a processor, the computer program implements the method embodiments described above.

[0055] That is, those skilled in the art will understand that all or part of the steps in the methods of the above embodiments can be implemented by a program instructing related hardware. This program is stored in a storage medium and includes several instructions to cause a device (which may be a microcontroller, chip, etc.) or processor to execute all or part of the steps of the methods described in the various embodiments of this application. The aforementioned storage medium includes various media capable of storing program code, such as a USB flash drive, a portable hard drive, a read-only memory (ROM), a random access memory (RAM), a magnetic disk, or an optical disk.

[0056] Those skilled in the art will understand that the above embodiments are specific examples of implementing the present invention, and in practical applications, various changes in form and detail may be made without departing from the spirit and scope of the present invention.

Claims

1. A method for identifying robot parameters, characterized in that, include: Acquire motion data for each joint of the robot; The motion data includes the joint motion speed, joint position, and joint current of each joint of the robot. A nonlinear friction model is constructed based on the motion data of each joint of the robot. The value of the power of the joint motion velocity in the friction calculation formula of the nonlinear friction model is adjusted so that the viscous friction in the friction calculation formula of the nonlinear friction model and the joint motion velocity with the power are linearly related. After linearizing the adjusted nonlinear friction model, the dynamic parameter vectors of each joint of the robot in the linearized nonlinear friction model are identified by the least squares method according to rigid body dynamics. Measurement data is obtained from the measured torque of the robot, and the frictional force is estimated based on the measurement data, inertial parameters, and rigid body dynamics. Based on the linearized nonlinear friction model and the estimated friction force, a pattern search algorithm is used to update the value of the power of the joint motion velocity in the friction calculation formula of the nonlinear friction model.

2. The robot parameter identification method according to claim 1, characterized in that, After linearizing the adjusted nonlinear friction model, the dynamic parameter vectors of each joint of the robot in the linearized nonlinear friction model are identified by the least squares method according to rigid body dynamics, including: The rigid body dynamics model of the robotic arm generated based on rigid body dynamics is combined with the linearized nonlinear friction force model and described in a linear regression form. Based on the dynamic model described in linear regression form, the observation matrix and the observation response matrix are calculated. The dynamic parameter vectors of each joint of the robot are obtained by identifying the observation matrix and the observation response matrix using the least squares method.

3. The robot parameter identification method according to claim 2, characterized in that, The calculation of the observation matrix and observation response matrix based on the dynamic model described in linear regression form includes: QR decomposition is used to eliminate irrelevant variables in the dynamic model described in the form of linear regression, and the parameters that are not eliminated are recombined to form a minimal set of inertial parameters. The minimum set of inertia is described in linear form; The linear regression observation matrix and observation response torque vector of the dynamic model described in linear regression form are calculated based on the minimum set of inertial parameters in the linear form.

4. The robot parameter identification method according to claim 3, characterized in that, The dynamic parameter vector consists of the inertial parameters, gravity parameters, equivalent rotational parameters of the joint transmission chain, and friction parameters of each joint of the robot.

5. The robot parameter identification method according to claim 4, characterized in that, The step of updating the power of the joint motion velocity in the friction calculation formula of the nonlinear friction model using a pattern search algorithm based on the linearized nonlinear friction model and the estimated friction force includes: A pattern search algorithm is used, taking the minimum value of the friction calculation formula in the linearized nonlinear friction model and the vector L2 norm of the estimated friction as the criterion, and taking the friction parameters in the dynamic parameter vector as the initial solution of the pattern search algorithm, to optimize the friction parameters in the dynamic parameter vector to obtain the global optimal solution of the pattern search algorithm; The value of the power of the joint motion velocity in the friction calculation formula of the nonlinear friction model is updated based on the global optimal solution of the pattern search algorithm.

6. The robot parameter identification method according to claim 1, characterized in that, The adjustment of the power of the joint motion velocity in the friction calculation formula of the nonlinear friction model includes: Adjust the value of the power of the joint motion velocity in the friction calculation formula of the nonlinear friction model to 1.

7. The robot parameter identification method according to claim 1, characterized in that, The acquisition of motion data for each joint of the robot includes: Control the robot's operating trajectory; Collect the actual position, speed, and current information of the robot's motor. Based on the rigid body dynamics, the actual position, velocity, and current information of the motor end are converted into motion data of each joint of the robot after unit conversion and data filtering.

8. A robot parameter identification device, characterized in that, include: The data acquisition module is used to acquire motion data of each joint of the robot; The motion data includes the joint motion speed, joint position, and joint current of each joint of the robot. The model degradation module is used to construct a nonlinear friction model based on the motion data of each joint of the robot, and adjust the value of the power of the joint motion velocity in the friction calculation formula of the nonlinear friction model so that the viscous friction in the friction calculation formula of the nonlinear friction model and the joint motion velocity with the power present a linear relationship. The dynamic parameter calculation module is used to linearize the adjusted nonlinear friction force model, and then, based on rigid body dynamics, identify the dynamic parameter vectors of each joint of the robot in the linearized nonlinear friction force model using the least squares method. The friction estimation module is used to obtain measurement data from the robot's measured torque and calculate and estimate the friction force based on the measurement data, inertial parameters, and rigid body dynamics. The parameter optimization module is used to update the value of the power of the joint motion velocity in the friction calculation formula of the nonlinear friction model by using a pattern search algorithm based on the linearized nonlinear friction model and the estimated friction force.

9. An electronic device, characterized in that, include: At least one processor; as well as, A memory communicatively connected to the at least one processor; wherein, The memory stores instructions that can be executed by the at least one processor to enable the at least one processor to perform the robot parameter identification method as described in any one of claims 1 to 7.

10. A computer-readable storage medium storing a computer program, characterized in that, When the computer program is executed by the processor, it implements the robot parameter identification method according to any one of claims 1 to 7.

Images