A robot dynamics identification method considering static separation of load and unified prediction of dynamic friction

Abstract

The application belongs to the technical field of robots, and discloses a robot dynamics identification method considering unified prediction of load static separation and dynamic friction, the system comprising: nonlinear friction pre-identification; a load-related friction static separation method (LFSS); dynamic friction modeling based on HTFO; PRF Fourier series excitation trajectory design; model verification. The application not only constructs a complete robot dynamics identification technology system, but also provides convenient tools and platforms for the popularization and use of the application in practical application, greatly promoting the development and application of robot technology under complex working conditions.

Description

TECHNICAL FIELD

[0001] The present application belongs to the field of robot technology, and in particular relates to a robot dynamics identification method considering unified prediction of static separation of load and dynamic friction. BACKGROUND

[0002] In the process of robot operation, significant changes in load, friction and disturbance pose high requirements for its dynamic adaptability. However, the prior art has many challenges in determining the dynamics parameters of the robot:

[0003] 1. In the dynamic identification of the traditional model-based experimental system, it is usually difficult to accurately consider the influence of load force and joint friction on the system characteristics because the mechanism is unknown, which leads to difficulty in accurately determining the dynamics parameters of the robot.

[0004] 2. The commonly used linear friction model (such as Coulomb-viscous friction model) cannot accurately describe the joint friction in the actual drive system due to the nonlinear characteristics of friction, especially in low speed conditions, which will cause large identification errors.

[0005] 3. Although there are some nonlinear friction models, these models are difficult to integrate into the dynamic identification system, and there is often a large deviation in the zero speed range. In addition, there is a strong correlation between the load and friction of the robot joint, and the external load will cause the friction and inertia force to be coupled, resulting in inaccurate dynamic parameter identification, and the existing method for separating nonlinear friction usually relies on the initial identified inertia parameters or multiple iterations, which is difficult to effectively decouple the inertia and friction parameters. SUMMARY

[0006] In view of the problems existing in the prior art, the present application provides a robot dynamics identification method considering unified prediction of static separation of load and dynamic friction.

[0007] The present application is implemented as follows: a robot dynamics identification method considering unified prediction of static separation of load and dynamic friction, the system comprises:

[0008] S1: nonlinear friction pre-identification;

[0009] S2: load-dependent friction static separation method (LFSS);

[0010] S3: dynamic friction modeling based on HTFO;

[0011] S4: PRF Fourier series excitation trajectory design;

[0012] S5: model verification.

[0013] Further, the S1 specifically comprises:

[0014] In view of the fact that the friction characteristics have a key influence on the performance of industrial robots in actual operation, especially at low speeds, and that traditional linear friction models, such as the Coulomb-viscous friction model, cannot accurately depict the actual friction behavior, the original linear friction model is modified:

[0015] The original linear friction model is:

[0016]

[0017] wherein, is the Coulomb friction, represents the static friction force, which plays an important role at the moment of starting the robot and during low-speed operation; this model only simply considers the linear relationship between the Coulomb friction and the viscous friction and the speed, and does not cover complex friction phenomena such as the Stribeck effect; the revised model is:

[0018]

[0019] wherein, is the Stribeck speed, which determines the key turning point of the change in friction with the speed; b is the damping attenuation coefficient, reflecting the attenuation characteristics of the nonlinear viscous force at different speeds; m is an empirical constant (m = 1) corresponding to the Tustin model, (m = 2) corresponding to the Gauss model), and different m values correspond to different friction speed variation laws;

[0020] In order to accurately identify the friction parameters, special joint trajectories are designed; in these trajectories, the joint acceleration is strictly guaranteed to be zero, and the friction is induced by changing the direction of the joint speed; high-precision sensors are used to collect the friction force data of the joint at different speeds, and the particle swarm optimization (PSO) method is used for parameter pre-identification; for example, for the HS-CO610 robot, after a large number of experiments and complex calculations, the detailed nonlinear friction parameters of each joint are obtained; taking Joint1 as an example, the Coulomb friction coefficient indicates the characteristics of the joint in terms of Coulomb friction; reflects the difference between the static friction force and the Coulomb friction force, and these parameters provide an important basis for subsequent accurate friction modeling.

[0021] Further, the S2 specifically comprises:

[0022] (1) Construction of a static separation method: in-depth research into the characteristics of robots in high-load and high-dynamic operating environments shows that the Coulomb friction coefficient The friction model and inertial parameters change significantly with the transmission force and external force. In accurate robot dynamics analysis, it is necessary to accurately identify the friction model and inertial parameters separately, and subtract the joint friction from the joint torque to determine the inertial torque. However, since the force / torque generated by inertial motion and joint friction are coupled with each other, the inertial parameter model deviates significantly from the actual value, which greatly affects the identification accuracy.

[0023] To address this, an LFSS method is proposed; this method will... Modeled as a polynomial function of four-quadrant dynamics and external forces, the key expression is derived through analysis of a large amount of experimental data and theoretical derivation:

[0024]

[0025] in, These are coefficients related to joint i, reflecting the linear relationship between joint friction and load torque. For the rotary joints of a serial robot, extensive experiments and theoretical analyses have revealed that only load torque aligned with the joint's rotation axis affects joint friction. This primarily includes the torque generated by gravity and the external load torque applied to the robot. When the robot is in static equilibrium, the angular velocity can be approximated as zero, i.e. ,at this time and Thus, a static model of the robot considering load-related static friction was obtained:

[0026]

[0027] Where K is the current constant and I is the motor input current. Based on this, by designing an experiment, it is only necessary to obtain the data on the external load change and current change of the robot in the static position, so that the parameters of the load-related friction model can be efficiently identified by the linear regression method.

[0028] (2) Implementation of the recognition process: First, based on the robot's structure and motion characteristics, the static loading position is carefully designed. In this process, the impact of coordinate system changes on experimental results is fully considered, using the coordinate transformation formula:

[0029]

[0030] and

[0031]

[0032] Accurately calculate the representation of external forces in joint space; among which, Represents external force in a general sense, derived from force and torque composition, The robot's Jacobian matrix is ​​obtained relative to the base coordinates G using the vector product method.

[0033] Next, using the above formula, the change in external force is obtained by transforming the force coordinate from T to G. With joint torque changes Precise relationship:

[0034]

[0035] Based on this relationship, a series of static load experiments were designed, and the conditions were optimized:

[0036]

[0037] The Particle Swarm Optimization (PSO) algorithm was used to determine the optimal posture of the robot in the static loading experiment, so as to maximize the loading effect of the robot and obtain sufficiently accurate experimental data.

[0038] Then, based on the experimentally collected data, the parameters of the load-related friction model were accurately fitted using the least squares (LS) method. Through detailed analysis and complex calculations of a large amount of data, the parameter fitting results of each joint were obtained. The identification results were evaluated by calculating the normalized root mean square error. The results show that the method can accurately identify the friction load correlation coefficient with an overall error of about 10%, effectively avoiding the coupling effect of inertia and friction parameters, and providing a reliable foundation for subsequent dynamic modeling.

[0039] Furthermore, S3 specifically includes:

[0040] In-depth research into existing dynamic friction models reveals that while the LuGre model can describe the dynamic characteristics of friction to some extent, it still has some shortcomings in robotic applications. To better adapt to the actual operating conditions of robots, the LuGre model has been simplified and improved. The original LuGre model is as follows:

[0041]

[0042] in, Let be the coefficient of dynamic friction and z be an internal variable. This model presents certain complexity and computational difficulties in describing robot joint friction, and its performance in critical regions such as zero-velocity intersections is not ideal. By introducing the Laplacian operator s and the hyperbolic tangent function, it is transformed into a dynamic friction model based on the first-order hyperbolic tangent HTFO:

[0043]

[0044] After such improvement, the model has significant advantages in describing the dynamic friction behavior of robot joints, especially in the zero speed cross range, can more accurately reflect the continuous characteristics of the actual friction, effectively reducing the complexity of dynamic parameter identification; through in-depth analysis and derivation of the robot motion equation, combined with the improved friction model above, the robot dynamic model is established, the effective decoupling of friction and inertia is realized, and the key to be identified parameters are determined, which provides a theoretical basis for subsequent parameter identification.

[0045] Further, the S4 specifically includes:

[0046] In view of the limitations of the traditional time domain excitation method in identifying dynamic friction, a pseudo-random finite PRF Fourier series excitation trajectory with time-frequency constraint is proposed; on the basis of the traditional finite Fourier series synthesis of periodic excitation reference, such as:

[0047]

[0048] In-depth consideration of the dynamic characteristics of robot joints, design the excitation current expression:

[0049]

[0050] Among them, For realizing the limited Fourier excitation trajectory, the basic shape and parameters of the trajectory are determined by the calculation method thereof; the PRBS pseudo-random binary sequence is used to effectively stimulate the friction, which generates a series of pulse signals to stimulate the robot joint to produce different friction responses through a specific coding rule; through in-depth analysis of the frequency response of the robot joint, the expression of the linearized HTFO friction model in the frequency domain is used:

[0051]

[0052] By estimating the frequency response function, the accurate identification of the dynamic friction parameters is realized; on the HS-CO610 robot experimental platform, through careful experimental design and data analysis, it is verified that the excitation design can effectively capture the joint friction dynamics in the standard dynamic parameter excitation and identification process.

[0053] Further, the S5 specifically includes:

[0054] A complete experimental system is carefully built on the HS-CO610 robot experimental platform, which includes the robot body, adjustable load device, high-performance controller, high-precision sensor and other key components, which can accurately load in the range of 1-4kg, accurately load, high-performance controller, high-precision sensor and other key components; through the coordinated work of these components, accurate control and data acquisition of the robot under different load and motion conditions can be realized;

[0055] During the verification process, multiple evaluation metrics were used to comprehensively analyze the experimental results. For the verification of the load-dependent friction model, the dynamic identification errors with and without load dependence models were mainly compared. Normalized root mean square error Peak error and other indicators were analyzed. Experimental results show that the LFSS method can significantly improve recognition accuracy. For example, under the minimum load VT1, the total value of the LFSS algorithm is reduced by 8.21% compared with the standard system without load separation, and the peak error and total value are reduced by 44.21%. Under the maximum load VT4, improvements of 30.77% and 62.27% are achieved, respectively. Moreover, the LFSS method provides more consistent fitting results for each axis under different loads, especially for joints such as Joint4, where friction and load are significantly correlated. The value decreased by more than 49%, effectively reducing the error;

[0056] In terms of verifying dynamic friction separation, the performance of Coulomb, tanh and HTFO friction models in predicting joint friction torque was compared. Through the analysis and calculation of a large amount of experimental data, the comparison results of the predicted torque and actual torque of each model, as well as the relative error and time cost, were obtained. The results show that the HTFO model has the best overall performance under different working conditions.

[0057] Furthermore, the S5 experimental procedure specifically includes:

[0058] (1) Under different load conditions, the LFSS method was used to accurately obtain parameters; by conducting a detailed analysis of the force and current changes of the robot joints under different loads, and combining the previously established model and algorithm, accurate parameter values ​​were obtained; for example, under a 1kg load, Joint1's , The values ​​are significantly different from those under a 4kg load, and these differences reflect the influence of load on friction parameters.

[0059] (2) Based on the continuous friction model, nonlinear friction is accurately calculated and parameters are obtained by collecting joint trajectory data at different speeds. In this process, high-precision sensors and advanced data acquisition systems were used to record the motion data of the robot joints under different speed and load combinations. Through complex calculations and analysis, the nonlinear friction parameters of each joint were obtained.

[0060] (3) Based on the PRF Fourier series excitation trajectory and frequency domain separation method, a regression matrix is ​​constructed to identify dynamic friction and inertial characteristics. ; Through processing and analysis of a large amount of experimental data, a regression matrix capable of accurately reflecting dynamic characteristics of the robot is obtained by using mathematical methods and algorithm optimization;

[0061] (4) By using the parameters and matrix obtained above, dynamic friction parameters of the HTFO model are accurately identified by a synchronous identification algorithm such as the least square method And inertia parameters So that the final robot dynamics model is obtained .

[0062] Another object of the present application is to provide a computer device, a computer readable storage medium and an information data processing terminal, characterized in that the computer device comprises a memory and a processor, the memory stores a computer program, and the computer program is executed by the processor to make the processor execute the steps of the robot dynamics identification method considering unified prediction of static separation of load and dynamic friction.

[0063] Another object of the present application is to provide a computer readable storage medium storing a computer program, which is executed by a processor to make the processor execute the steps of the robot dynamics identification method considering unified prediction of static separation of load and dynamic friction.

[0064] Another object of the present application is to provide an information data processing terminal integrating the steps of the robot dynamics identification method considering unified prediction of static separation of load and dynamic friction; in the terminal, the robot system can collect various data in the running process in real time, including joint motion data, load data and friction data, and transmit the data to the processing module of the terminal.

[0065] In combination with the above technical solutions and solved technical problems, the technical solution to be protected by the present application has the following advantages and positive effects:

[0066] By providing the computer device and the information data processing terminal, the present application not only constructs a complete robot dynamics identification technical system, but also provides a convenient tool and platform for its popularization and use in practical application, greatly promoting the development and application of robot technology under complex working conditions.

[0067] 1. By the innovative LFSS method, high-precision static separation of load-dependent friction is successfully realized, thereby effectively breaking the dilemma of friction-inertia coupling and greatly improving the accuracy of the robot dynamics regression model.

[0068] 2. The HTFO dynamic friction model, combined with pre-identified parameters, can describe the dynamic friction behavior of robot joints under various working conditions with extremely high accuracy. During robot movement, especially in critical and complex areas such as zero-speed crossings, it can more accurately reflect the continuous characteristics of actual friction, effectively reducing the complexity of dynamic parameter identification and significantly improving the accuracy of robot dynamic parameter identification.

[0069] 3. In the nonlinear friction pre-identification stage, a special joint trajectory was designed to ensure zero joint acceleration and change the direction of joint velocity to measure friction. Then, high-precision sensors were used to collect data, and particle swarm optimization (PSO) was employed for parameter pre-identification. This method can quickly and accurately acquire nonlinear friction parameters, significantly improving identification efficiency compared to traditional methods and saving time and computational resources for subsequent overall dynamic parameter identification. It also overcomes the limitations of traditional algorithms, making the entire identification process more efficient.

[0070] The expected benefits and commercial value of the technical solution of this invention after transformation are as follows:

[0071] This invention effectively improves the accuracy and efficiency of parameter recognition for robots under varying load and friction conditions through innovative methods, thereby enhancing the robot's motion control stability and operational precision. By accurately identifying dynamic parameters, it effectively reduces errors and equipment wear during the production process, significantly improving product yield in high-precision manufacturing. With the increasingly widespread application of robots, their role in enhancing equipment performance and added value will bring continuous economic benefits. Whether through technology transfer or equipment upgrade services, they can occupy an important position in the market and possess extremely high commercial value.

[0072] The technical solution of this invention fills a technological gap in the industry both domestically and internationally:

[0073] In the field of robot dynamics research, despite numerous studies, effective solutions have been lacking for key issues related to load and friction. The unique identification system proposed in this invention, with its innovative features such as the LFSS method for accurately handling friction-inertial coupling, the HTFO model for precisely describing dynamic friction, and the PRF excitation trajectory for efficient identification, successfully fills a long-standing technological gap in this area.

[0074] The technical solution of this invention solves a technical problem that people have long desired to solve but have been unable to achieve:

[0075] For a long time, robots in the actual operation process of the load and friction changes caused by the dynamics performance fluctuations, has been the industry to be solved key technical problems. The present application in-depth analysis of the load and friction mechanism, build innovative model and identification method, in different load (covering 1-4kg and more widely range) and speed conditions can realize high-precision dynamics parameter identification, effectively inhibit the friction torque and joint reverse error, greatly improve the robot motion precision and stability, achieved the industry long-term technical expectations.

[0076] The technical scheme of the present application overcomes the technical bias:

[0077] The past research is limited to a certain extent by the dependence on the traditional linear friction model, and there are difficulties in the integration of dynamic friction model and the overall dynamics system of the robot, forming a technical bias. Many researchers believe that the traditional linear friction model has a certain applicability in the analysis of robot dynamics, and the limitations of the traditional linear friction model in complex working conditions are not well understood; the present application breaks through the shackles of these traditional concepts, through the innovation revision of the traditional friction model and the effective improvement and integration of the dynamic friction model, proves the feasibility and superiority of the new technical path, brings new ideas and methods for the research of robot dynamics, and promotes the innovative development of the industry technology. BRIEF DESCRIPTION OF DRAWINGS

[0078] Figure 1 is the process of the robot dynamics identification system provided by the embodiment of the present application;

[0079] Figure 2 is the schematic diagram of the robot pre-identification method provided by the embodiment of the present application;

[0080] Figure 3 is the joint stress and coordinate system transformation method in the LFSS method provided by the embodiment of the present application;

[0081] Figure 4 is the performance of the HTFO and the traditional friction model under different conditions;

[0082] Figure 5 is the joint frequency response result of the PRF excitation trajectory design provided by the embodiment of the present application;

[0083] Figure 6 is the hardware architecture of the robot system provided by the embodiment of the present application;

[0084] Figure 7 is the comparison of the dynamics parameters before and after the present application under different friction models provided by the embodiment of the present application;

[0085] Figure 8The application embodiment provides a comparison of the parameter identification accuracy before and after the application under different loads. DETAILED DESCRIPTION

[0086] In order to make the purpose, technical scheme and advantages of the application more clear, the application is further described in detail below in combination with embodiments. It should be understood that the specific embodiments described herein are only used to explain the application and do not limit the application.

[0087] As shown in Figure 1 The application embodiment provides a robot dynamics identification method considering unified prediction of static separation and dynamic friction of load, and the system comprises:

[0088] S1: nonlinear friction pre-identification;

[0089] S2: load-dependent friction static separation method (LFSS);

[0090] S3: dynamic friction modeling based on HTFO;

[0091] S4: PRF Fourier series excitation trajectory design;

[0092] S5: model verification.

[0093] Further, the S1 specifically comprises:

[0094] As shown in Figure 2 Nonlinear friction pre-identification is performed. In view of the fact that the friction characteristics have a key influence on the performance of industrial robots in actual operation, especially at low speed, and the traditional linear friction model, such as the Coulomb-viscous friction model, cannot accurately depict the actual friction behavior, the traditional linear friction model is modified:

[0095] The original linear friction model is:

[0096]

[0097] This model only simply considers the linear relationship between Coulomb friction and viscous friction and speed, and does not cover complex friction phenomena such as Stribeck effect; the revised model is:

[0098]

[0099] Among them, represents the static friction force, which plays an important role at the moment of starting the robot and during low-speed operation; Stribeck velocity, which determines the key turning point of friction changing with velocity; b is the damping attenuation coefficient, which reflects the attenuation characteristics of nonlinear viscous force at different speeds; m is an empirical constant (m = 1 corresponds to the Tustin model, and m = 2 corresponds to the Gauss model), and different m values correspond to different friction velocity change laws;

[0100] To accurately identify the parameters in equation 2, special joint trajectories are designed; in these trajectories, the joint acceleration is strictly guaranteed to be zero, and the friction is induced by changing the joint speed direction; high-precision sensors are used to collect joint friction data at different speeds, and then the particle swarm optimization (PSO) method is used for parameter pre-identification; for example, for the HS-CO610 robot, after a large number of experiments and complex calculations, the detailed nonlinear friction parameters of each joint are obtained; taking Joint1 as an example, which shows the characteristics of the joint in terms of Coulomb friction; which reflects the difference between the static friction force and the Coulomb friction force, and these parameters provide an important basis for subsequent accurate friction modeling.

[0101] The S2 specifically includes:

[0102] (1) Construction of static separation method: through in-depth research on the characteristics of robots in high-load and high-dynamic operating environments, it is found that the Coulomb friction coefficient will change significantly with transmission power and external force; in accurate robot dynamics analysis, it is necessary to accurately identify the friction model and inertia parameters respectively, and to subtract the joint friction from the joint torque to determine the inertia torque; however, due to the coupling of inertia motion and joint friction forces / torques, the inertia parameter model deviates significantly from the actual value, greatly affecting the identification accuracy;

[0103] Therefore, as shown in Figure 3 , load-dependent friction static separation is performed. An LFSS method is proposed; this method models as a polynomial function of four-quadrant dynamics and external forces, and through analysis of a large amount of experimental data and theoretical derivation, the key expressions are obtained:

[0104]

[0105] wherein, are coefficients related to joint i, which reflect the linear relationship between joint friction and load torque; for the rotational joints of a serial robot, it is found through a large number of experiments and theoretical analysis that only the load torque aligned with the joint rotation axis will affect the joint friction, which mainly includes the torque generated by gravity and the load torque applied to the robot externally; when the robot is in a static equilibrium state, since the angular velocity can be approximately considered as zero, i.e. at this time and Thus, the static model of the robot considering the load-dependent static friction is obtained:

[0106]

[0107] On this basis, by designing experiments, only the external load change and current change data of the robot under static position are obtained, and the linear regression method is used to efficiently identify the parameters of the load-dependent friction model;

[0108] (2) Implementation of the identification process: first, according to the structure and motion characteristics of the robot, the static loading position is carefully designed ; In this process, the influence of the change of the coordinate system on the experimental results is fully considered, and the coordinate transformation formula is used:

[0109]

[0110] and

[0111]

[0112] The accurate calculation of the external force in the joint space is obtained; wherein, represents the generalized external force, which is composed of force and torque , is the Jacobian matrix of the robot, which is obtained by vector product method relative to the base coordinate G;

[0113] Then, by using the above formula, the accurate relationship between the change of the external force and the change of the joint torque is obtained by converting the force coordinates from T to G:

[0114]

[0115] Based on this relationship, a series of static load experiments are designed, and the optimal conditions are optimized:

[0116]

[0117] And the particle swarm optimization (PSO) algorithm is used to solve, and the optimal posture of the robot in the static loading experiment is determined to maximize the loading effect of the robot and obtain accurate experimental data;

[0118] Then, according to the data collected in the experiment, the parameters of the load-dependent friction model are accurately fitted by using the least squares method (LS); through detailed analysis and complex calculation of a large amount of data, the parameter fitting results of each joint are obtained; the identification results are evaluated by calculating the normalized root mean square error, and the results show that this method can accurately identify the friction load-dependent coefficient, and the overall error is about 10%, which effectively avoids the coupling effect of inertia and friction parameters, and provides a reliable foundation for subsequent dynamic modeling.

[0119] The S3 specifically includes:

[0120] Based on the dynamic friction modeling of HTFO (refer to FIG. 4), the existing dynamic friction model is deeply studied, and it is found that although the LuGre model can describe the dynamic characteristics of friction to some extent, it still has some shortcomings in robot application; in order to better adapt to the actual operation of the robot, the LuGre model is simplified and improved; the original LuGre model is:

[0121]

[0122] The model has certain complexity and calculation difficulty in describing the joint friction of the robot, and the performance in the key area of zero speed cross is not ideal; by introducing Laplacian operator s and hyperbolic tangent function, it is converted into a dynamic friction model based on hyperbolic tangent first-order HTFO:

[0123]

[0124] After such improvement, the model has significant advantages in describing the dynamic friction behavior of the robot joint, especially in the zero speed cross range, which can more accurately reflect the continuous characteristics of the actual friction, and effectively reduces the complexity of dynamic parameter identification; through in-depth analysis and derivation of the robot motion equation, combined with the improved friction model above, the dynamic model of the robot is established, the friction-inertia is effectively decoupled, and the key to be identified parameters are determined, which provides a theoretical basis for subsequent parameter identification.

[0125] The S4 specifically includes:

[0126] As shown in Figure 5 , a PRF Fourier series excitation trajectory is designed, and in view of the limitations of the traditional time domain excitation method in identifying dynamic friction, a pseudo-random finite PRF Fourier series excitation trajectory with time-frequency constraint is proposed; on the basis of the traditional finite Fourier series synthesis of periodic excitation reference, such as:

[0127]

[0128] Deep consideration of the dynamics of the robot joint, the excitation current expression is designed:

[0129]

[0130] wherein, For the implementation of the limited Fourier excitation trajectory, the basic shape and parameters of the trajectory are determined by the way of its calculation; the PRBS pseudo-random binary sequence is used to effectively excite the friction, which generates a series of pulse signals to stimulate the robot joint to produce different friction responses through a specific coding rule; through in-depth analysis of the frequency response of the robot joint, the linearized HTFO friction model is used to express the frequency domain:

[0131]

[0132] By estimating the frequency response function, the accurate identification of the dynamic friction parameters is realized; on the HS-CO610 robot experimental platform, through careful experimental design and data analysis, it is verified that the excitation design can effectively capture the joint friction dynamics in the standard dynamic parameter excitation and identification process.

[0133] The S5 specifically includes:

[0134] A complete experimental system is carefully built on the HS-CO610 robot experimental platform, which includes the robot body, adjustable load device, which can accurately load in the range of 1-4kg, high-performance controller, high-precision sensor and other key components; through the coordinated work of these components, accurate control and data acquisition of the robot under different load and motion conditions can be realized;

[0135] In the verification process, a variety of evaluation indexes are used to comprehensively analyze the experimental results; for the verification of the load-dependent friction model, the dynamics identification error is mainly compared with and without the load-dependent model, as Figure 8 shown, the dynamics identification error is calculated and compared with and without the LFSS method, the normalized root mean square error and peak error indexes are used; the experimental results show that the LFSS method can significantly improve the identification accuracy; for example, under the minimum load VT1, the value of the LFSS algorithm is reduced by 8.21% compared with the standard system without load separation, the peak error and the value of the total error are reduced by 44.21%; at the maximum load VT4, there are 30.77% and 62.27% improvements respectively; moreover, the LFSS method has more consistent fitting results for each axis under different loads, especially for joints such as Joint4 with obvious friction and load dependence, the value is reduced by more than 49%, effectively reducing the error;

[0136] Referring to FIG. 7, in terms of verification of dynamic friction separation, the performances of Coulomb, tanh and HTFO friction models in predicting joint friction torque are compared; through analysis and calculation of a large amount of experimental data, the comparison results of predicted torque and actual torque of each model, relative error and time cost are obtained; the results show that the HTFO model has the best comprehensive performance under different working conditions.

[0137] In summary, through the above specific embodiments, the present application can effectively realize the accurate identification of robot dynamics parameters, and greatly improve the performance of robots in complex working conditions. In actual application, the experimental parameters and methods can be appropriately adjusted and optimized according to the characteristics and application scenarios of different robots to further improve the accuracy and reliability of robot dynamics identification.

[0138] The S5 experimental process specifically includes:

[0139] (1) Under different load conditions, the LFSS method is used to accurately obtain parameters; through detailed analysis of the force and current changes of the robot joints under different loads, combined with the previously established model and algorithm, accurate parameter values are obtained; for example, under the load of 1kg, the value of Joint1 , is obviously different from that under the load of 4kg, and these differences reflect the influence law of load on friction parameters;

[0140] (2) According to the continuous friction model, the joint trajectory data under different speeds are collected to accurately calculate the nonlinear friction and obtain the parameter ; in this process, high-precision sensors and advanced data acquisition systems are used to record the motion data of the robot joints under different speed and load combinations, and through complex calculation and analysis, the nonlinear friction parameters of each joint are obtained;

[0141] (3) Based on the PRF Fourier series excitation trajectory and frequency domain separation method, a regression matrix for identifying dynamic friction and inertia characteristics is constructed ; through processing and analysis of a large amount of experimental data, mathematical methods and algorithm optimization are used to obtain a regression matrix that can accurately reflect the dynamic characteristics of the robot;

[0142] (4) Using the parameters and matrices obtained above, the dynamic friction parameters and inertia parameters of the HTFO model are accurately identified through a synchronous identification algorithm such as the least squares method, so as to obtain the final robot dynamics model .

[0143] The embodiment of the present application provides a computer device, a computer readable storage medium and an information data processing terminal, characterized in that the computer device comprises a memory and a processor, the memory stores a computer program, and the computer program is executed by the processor to make the processor execute the steps of the robot dynamics identification method considering load static separation and dynamic friction unified prediction.

[0144] The embodiment of the present application provides a computer readable storage medium, which stores a computer program, and the computer program is executed by a processor to make the processor execute the steps of the robot dynamics identification method considering load static separation and dynamic friction unified prediction.

[0145] The embodiment of the present application provides an information data processing terminal, which integrates the steps of the robot dynamics identification method considering load static separation and dynamic friction unified prediction; in the terminal, the robot system can collect various data in the running process of itself in real time, including joint motion data, load data and friction data, and transmit the data to the processing module of the terminal.

[0146] The robot dynamics identification system provided by the present application has a wide application field and can significantly improve the performance and working efficiency of related products, and provides key technical support for the intelligent development of many industries.

[0147] (I) polishing processing of complex metal parts

[0148] In the field of precision metal processing, it is difficult for traditional robots to balance processing accuracy and efficiency in polishing processing of complex-shaped parts. The robot system enabled by the robot dynamics identification system of the present application can accurately control the polishing process. First, according to the three-dimensional model of the metal part, the motion path of the robot is planned, and the path is optimized in combination with the dynamics identification result. In the processing process, the load-dependent friction static separation (LFSS) method is used to adjust the driving force of the robot joint in real time to adapt to different polishing forces and changes in workpiece weight, ensuring stable contact between the milling cutter and the workpiece. Based on the hyperbolic tangent first order (HTFO) dynamic friction modeling, the friction influence of the robot in high-speed motion and direction switching is effectively reduced, and the polishing accuracy is improved.

[0149] (III) manufacturing of medical rehabilitation equipment parts

[0150] In the manufacture of medical rehabilitation equipment, the manufacturing precision of some precise joint parts and transmission parts directly affects the performance of the equipment and the use experience of patients. By using the robot dynamics identification system of the application, the robot can be accurately positioned and operated during the manufacturing process. For example, when producing key parts of a knee joint rehabilitation device, the robot motion is made more stable by using a pseudo-random finite (PRF) Fourier series excitation trajectory design, reducing the influence of vibration and impact on machining precision. By using high-precision dynamics parameter identification, the robot can accurately apply appropriate force and motion at different processing stages to ensure the dimensional accuracy and surface quality of the parts.

[0151] It should be noted that the embodiments of the present application can be realized by hardware, software or a combination of software and hardware. The hardware part can be realized by special logic; the software part can be stored in a memory and executed by a suitable instruction execution system, such as a microprocessor or a specially designed hardware. Those skilled in the art can understand that the above-mentioned devices and methods can be realized by computer executable instructions and / or included in processor control code, such as carrier media, such as magnetic disk, CD or DVD-ROM, programmable memory, such as read-only memory (firmware), or data carrier, such as optical or electronic signal carrier. The device and its modules of the present application can be realized by hardware circuit, such as ultra-large scale integrated circuit or gate array, semiconductor, such as logic chip, transistor, etc., or programmable hardware device, such as field programmable gate array, programmable logic device, etc., or by software executed by various types of processors, or by a combination of the above hardware circuit and software, such as firmware.

[0152] The above is only a specific embodiment of the present application, but the protection scope of the present application is not limited thereto, any modification, equivalent replacement and improvement within the technical range disclosed by the present application, which is made by those skilled in the art within the spirit and principle of the present application, should be covered within the protection scope of the present application.

Claims

1. A robot dynamics identification system considering unified prediction of load static separation and dynamic friction, characterized in that, The system comprises: A nonlinear friction pre-identification unit for identifying nonlinear friction parameters based on joint friction data using a particle swarm optimization algorithm (PSO); A load-dependent friction static separation unit for constructing a load-dependent friction model based on a polynomial function using a load-dependent friction static separation method (LFSS) and identifying parameters through a static load experiment; A dynamic friction modeling unit for establishing a dynamic friction model based on a high-order hyperbolic tangent function (HTFO) friction model combined with a Laplacian operator (s); An excitation trajectory design unit for constructing an excitation current signal using a pseudo-random finite Fourier series (PRF) excitation trajectory design method to separate dynamic friction parameters; A model verification unit for evaluating the accuracy and stability of the identified dynamic model based on a robot experiment platform using normalized root mean square error (NRMSE) and peak error; The load-dependent friction static separation method LFSS will Modelled as a polynomial function of the four-quadrant dynamics and external forces, the expression is as follows: ; wherein, is a coefficient related to joint i, reflecting the linear relationship between joint friction and load torque; for the rotary joint of serial robot, only the load torque aligned with the joint rotation axis will affect the joint friction, including the torque generated by gravity and the load torque applied externally to the robot; when the robot is in a static equilibrium state, the angular velocity is approximately considered as zero, i.e. , at this time and , the static model of the robot considering load-dependent static friction is obtained: ; On this basis, the external load change and current change data of the robot under static parts are obtained, and the parameters of the load-dependent friction model can be efficiently identified using linear regression method.

2. The system of claim 1, wherein, The nonlinear friction pre-identification unit uses high-precision sensors to collect joint friction data at different speeds, and combines particle swarm optimization (PSO) to identify and revise the friction model parameters, including static friction, Stribeck speed, damping attenuation coefficient, and friction speed variation law coefficient.

3. The system of claim 1, wherein, The load-dependent friction static separation unit collects external force and current data through a static load experiment, and uses linear regression method to fit the load-dependent friction model parameters. The external force in the joint space is calculated using coordinate transformation formula, and the optimal loading posture is determined based on optimization conditions.

4. The system of claim 1, wherein, The dynamic friction modeling unit is based on the LuGre friction model, and uses the Laplacian operator (s) and hyperbolic tangent function to improve the friction model to describe the friction change in the zero-speed cross region. A dynamic friction model based on HTFO is constructed, and the friction-inertia decoupling is realized in the robot motion equation.

5. The system of claim 1, wherein, The excitation trajectory design unit uses the pseudo-random finite Fourier series (PRF) excitation method to synthesize periodic excitation signals through Fourier series, and combines pseudo-random binary sequence (PRBS) to excite the robot joint friction characteristics, realize joint frequency response estimation, and identify friction parameters based on HTFO friction model.

6. The system of claim 1, wherein, The model verification unit evaluates the accuracy of the load-dependent friction model on the robot experiment platform using normalized root mean square error (NRMSE) and peak error, and compares the torque prediction error of Coulomb friction, tanh friction and HTFO friction model to verify the dynamic friction identification ability of HTFO model.

7. The system of claim 1, wherein, The model verification unit accurately identifies the dynamic friction parameters and inertia parameters of the HTFO model based on experimental data using least squares (LS) and synchronous identification algorithm, constructs the final robot dynamics model, and evaluates the identification error of the load-dependent friction model under different load conditions.

8. A computer device, a computer readable storage medium and an information data processing terminal, characterized by, The computer device comprises a memory and a processor, the memory stores a computer program, and the computer program is executed by the processor to enable the processor to implement the system according to any one of claims 1-7. 9.A computer readable storage medium, storing a computer program, the computer program is executed by a processor to enable the processor to implement the system according to any one of claims 1-7. 10.An information data processing terminal, which integrates the robot dynamics identification system considering unified prediction of static separation and dynamic friction according to any one of claims 1-7; in the terminal, the robot system can collect various data in the running process in real time, including joint motion data, load data and friction data, and transmit the data to the processing module of the terminal.

Images