Robot parameter self-tuning method and system

Abstract

The application relates to a robot parameter self-setting method and system, and belongs to the technical field of robot control. The method obtains a speed loop trajectory model of a robot servo system; based on a preset mode search algorithm, parameter optimization is performed on speed loop parameters of the speed loop trajectory model to obtain optimized parameters, and the speed loop trajectory model is adjusted according to the optimized parameters to obtain an adjusted speed loop trajectory model; an evaluation function is established according to a preset time domain index, and the adjusted speed loop trajectory model is evaluated according to the evaluation function to obtain an evaluation score; when the number of optimization times and the evaluation score of the current time meet preset conditions, the optimized parameters obtained at the current time are confirmed as target parameters, the speed loop parameters of the robot servo system are self-set, the difficulty of parameter setting is effectively reduced, the parameter optimization implementation process is simplified, an accurate model structure and parameters are not needed, and high-reliability parameter self-setting is realized.

Description

Technical Field

[0001] This application relates to the field of robot control technology, and in particular to a method and system for self-tuning robot parameters. Background Technology

[0002] Servo systems, as the core component of industrial robot motion control, have a crucial impact on the robot's motion performance and control accuracy. Traditionally, the adjustment and optimization of robot joint servo system control parameters usually rely on manual adjustments. This process is not only time-consuming and labor-intensive but also easily affected by subjective factors, leading to degraded control performance and even unstable system output. To address the performance degradation caused by improper servo system control parameter settings during manual tuning and the low efficiency of parameter optimization, a self-tuning servo system control parameter scheme is proposed.

[0003] Currently, existing parameter self-tuning methods rely heavily on the accurate mathematical model of the controlled object, resulting in high implementation complexity. If the model is inaccurate or there are unmodeled dynamics, the tuning effect will decrease significantly, and may even lead to system performance degradation. Summary of the Invention

[0004] Based on this, and addressing the technical problems existing in the parameter self-tuning methods of the aforementioned existing robot joint servo systems, a robot parameter self-tuning method and system is provided that uses a time-domain index to design a parameter self-tuning strategy, achieving highly reliable parameter self-tuning without requiring precise model structures and parameters.

[0005] In a first aspect, this application provides a method for self-tuning robot parameters, comprising the following steps:

[0006] Obtain the velocity loop trajectory model of the robot servo system;

[0007] Based on the preset pattern search algorithm, the velocity loop parameters of the velocity loop trajectory model are optimized to obtain the optimized parameters. Based on the optimized parameters, the velocity loop trajectory model is adjusted to obtain the adjusted velocity loop trajectory model.

[0008] Based on the preset time-domain indicators, an evaluation function is established, and the adjusted velocity loop trajectory model is evaluated according to the evaluation function to obtain an evaluation score.

[0009] When the number of optimization attempts and the current evaluation score meet the preset conditions, the optimization parameters obtained in the current attempt are confirmed as the target parameters.

[0010] In one embodiment, the preset time-domain indicators include motor-side trajectory accuracy, motor-side end-effector vibration, motor-side end-effector convergence time, and motor-side corner overshoot.

[0011] The steps for establishing an evaluation function based on preset time-domain indicators include:

[0012] The evaluation function is obtained based on any one of the following: motor-side trajectory accuracy, motor-side end vibration, motor-side end convergence time, and motor-side corner overshoot.

[0013] Alternatively, an evaluation function can be obtained by weighting any combination of motor-side trajectory accuracy, motor-side end vibration, motor-side end convergence time, and motor-side corner overshoot.

[0014] In one embodiment, when the preset time-domain index is the motor-side trajectory accuracy, the evaluation function is:

[0015] ,in, This represents the maximum accuracy of each trajectory segment before optimization. To achieve the maximum accuracy of each trajectory segment after optimization, The evaluation score corresponds to the accuracy of the motor-side trajectory. This is the preset expected score.

[0016] In one embodiment, when the preset time-domain index is the vibration at the motor end, the evaluation function is:

[0017] ,in, To optimize the amplitude of the corresponding frequency component before optimization, To optimize the amplitude of the corresponding frequency component, This represents the maximum value of the amplitude ratio before and after optimization of the corresponding process vibration. This represents the maximum value of the amplitude ratio before and after optimization for the corresponding residual vibration. This is the evaluation score for the vibration at the motor end.

[0018] In one embodiment, when the preset time-domain index is motor-side corner overshoot, the evaluation function is:

[0019] ,in, To address the corner overshoot at each inflection point before optimization, To optimize the corner overshoot at each inflection point, This is the evaluation score for the motor side corner overshoot.

[0020] In one embodiment, when the preset time-domain index is the motor-side end-convergence time, the evaluation function is:

[0021] ,in, The convergence time before optimization. The optimized convergence time, This is the evaluation score for the convergence time at the motor side.

[0022] In one embodiment, the step of optimizing the velocity loop parameters of the velocity loop trajectory model based on a preset pattern search algorithm to obtain the optimized parameters includes:

[0023] Axial search steps: Based on the preset pattern search algorithm, perform axial search on the velocity loop parameters of the velocity loop trajectory model to obtain the first intermediate parameters;

[0024] Pattern search steps: Based on the first intermediate parameter, perform a pattern search on the velocity loop parameters of the velocity loop trajectory model to obtain the second intermediate parameter;

[0025] The axial search step and the pattern search step are executed iteratively until the number of iterations reaches a preset first threshold, and the optimization parameters are obtained.

[0026] In one embodiment, after iteratively executing the axial search step and the pattern search step until the number of iterations reaches a preset threshold and the optimization parameters are obtained, the following steps are included:

[0027] Obtain the global optimal solution value and the global suboptimal solution value for the corresponding optimization parameters;

[0028] The difference between the global optimal solution value and the global suboptimal solution value is processed to obtain the baseline value;

[0029] Based on the baseline values, a pattern search is performed on the velocity loop parameters of the velocity loop trajectory model to obtain the optimized optimization parameters.

[0030] In one embodiment, the step of confirming the optimization parameters obtained in the current iteration as the target parameters when the number of optimization attempts and the current evaluation score meet preset conditions includes:

[0031] When the number of optimization attempts reaches a preset first threshold and / or the evaluation score reaches a preset second threshold, the optimization parameters obtained in the current attempt are confirmed as target parameters.

[0032] Secondly, this application also provides a robot parameter self-tuning system, including a processing device, which includes a memory and a processor. The memory stores a computer program, and the processor executes the computer program to implement the steps of any of the robot parameter self-tuning methods described above.

[0033] One of the above technical solutions has the following advantages and beneficial effects:

[0034] In the aforementioned robot parameter self-tuning method, the velocity loop trajectory model of the robot servo system is obtained; based on a preset pattern search algorithm, the velocity loop parameters of the velocity loop trajectory model are optimized to obtain optimized parameters; the velocity loop trajectory model is then adjusted according to the optimized parameters to obtain the adjusted velocity loop trajectory model; an evaluation function is established based on a preset time-domain index, and the adjusted velocity loop trajectory model is evaluated according to the evaluation function to obtain an evaluation score; when the number of optimization attempts and the current evaluation score meet preset conditions, the optimized parameters obtained in the current attempt are confirmed as target parameters, thus realizing the self-tuning of the velocity loop parameters of the robot servo system. This application optimizes the velocity loop parameters of the velocity loop trajectory model through a preset pattern search algorithm and evaluates the parameter optimization through a preset time-domain index, effectively reducing the difficulty of parameter tuning and simplifying the parameter optimization implementation process. The parameter self-tuning strategy is designed based on time-domain indexes, eliminating the need for precise model structures and parameters, and achieving highly reliable parameter self-tuning. Attached Figure Description

[0035] Figure 1 This is a schematic diagram illustrating the application environment of the robot parameter self-tuning method in the embodiments of this application;

[0036] Figure 2 This is a flowchart illustrating the robot parameter self-tuning method in the embodiments of this application;

[0037] Figure 3 This is a flowchart illustrating the steps for obtaining optimization parameters in an embodiment of this application.

[0038] Figure 4 This is a schematic diagram of the velocity loop trajectory model of the robot servo system in an embodiment of this application. Detailed Implementation

[0039] To enable those skilled in the art to better understand the present application, the technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present application, and not all embodiments. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative effort should fall within the scope of protection of the present application.

[0040] It should be noted that the terms "first," "second," etc., in the specification, claims, and accompanying drawings of this application are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate for the embodiments of this application described herein. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover non-exclusive inclusion; for example, a process, method, system, product, or apparatus that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or apparatus.

[0041] In addition, the term "multiple" should mean two or more.

[0042] It should be noted that, unless otherwise specified, the embodiments and features described in this application can be combined with each other. This application will now be described in detail with reference to the accompanying drawings and embodiments.

[0043] The robot parameter self-tuning method provided in this application can be applied to, for example... Figure 1 In the application environment shown, the robot parameter self-tuning system includes a processing device comprising a memory 104 and a processor 102. The processor 102 is connected to the memory 104, which stores the velocity loop trajectory model, evaluation score, and optimization parameters. The processor 102 acquires the velocity loop trajectory model of the robot servo system. Based on a preset pattern search algorithm, the system optimizes the velocity loop parameters of the velocity loop trajectory model to obtain optimized parameters. Based on these optimized parameters, the system adjusts the velocity loop trajectory model to obtain an adjusted velocity loop trajectory model. An evaluation function is established based on a preset time-domain index, and the adjusted velocity loop trajectory model is evaluated based on this function to obtain an evaluation score. When the number of optimization attempts and the current evaluation score meet preset conditions, the currently obtained optimized parameters are confirmed as target parameters. The processing device may also include a display 106, which is connected to the processor 102 and displays information such as the velocity loop trajectory model, evaluation score, and optimization parameters.

[0044] In one embodiment, such as Figure 2 As shown, a method for self-tuning robot parameters is also provided, including the following steps:

[0045] Step S210: Obtain the velocity loop trajectory model of the robot servo system.

[0046] The robot can be an industrial robot, such as a six-degree-of-freedom serial robot. The robot has a servo system for high-precision closed-loop control of the robot's joint motors. The velocity loop trajectory model can be an offline simulation model of the robot.

[0047] The velocity loop trajectory model of the robot servo system is as follows Figure 4 As shown, where, This is the speed loop proportional coefficient. The integral coefficient of the velocity loop; Let be the current loop transfer function, and s be the Laplace operator; This is the motor torque coefficient. This is the equivalent rotational inertia at the motor end. As a factor of kinetic friction, Let be the velocity loop filter coefficients. Then, the open-loop transfer function of the corresponding velocity loop trajectory model can be expressed as: .

[0048] After merging and omitting higher-order terms, we get: ; ; ; .

[0049] Step S220: Based on the preset pattern search algorithm, optimize the velocity loop parameters of the velocity loop trajectory model to obtain the optimized parameters, and adjust the velocity loop trajectory model according to the optimized parameters to obtain the adjusted velocity loop trajectory model.

[0050] The preset pattern search algorithm is obtained through the pattern-search algorithm, also known as the Hooke-Jeeves algorithm, which is a derivativeless optimization method. The pattern-search algorithm finds the minimum of the objective function by alternating axial probing (exploring local descent directions) and pattern shifting (accelerating the search along favorable directions), and is suitable for optimization problems where the objective function is non-differentiable or unconstrained. The velocity loop parameters of the velocity loop trajectory model can be velocity loop scaling factors. and velocity loop integral coefficient .

[0051] For example, based on a preset pattern search algorithm, the velocity loop proportional coefficient and velocity loop integral coefficient of the velocity loop trajectory model are optimized. The optimized parameters are then used to adjust the velocity loop trajectory model, resulting in the adjusted velocity loop trajectory model. This, combined with the velocity loop structure, achieves the entire parameter self-tuning process. It should be noted that the processor can preset the number of optimization attempts, and then, based on these attempts, use a preset pattern search algorithm to optimize the velocity loop proportional coefficient and velocity loop integral coefficient of the velocity loop trajectory model, thereby improving the accuracy of the final output optimized parameters.

[0052] Step S230: Establish an evaluation function based on the preset time domain index, and evaluate the adjusted velocity loop trajectory model according to the evaluation function to obtain an evaluation score.

[0053] The preset time-domain indicators can be set based on the characteristic parameters of the corresponding motors on the robot. The evaluation function is used to calculate the score of the corresponding time-domain indicator, thereby evaluating the corresponding velocity loop trajectory model. It should be noted that the evaluation function can be established based on the values ​​of the corresponding time-domain indicator before and after optimization.

[0054] Based on the preset time-domain indicators, an evaluation function corresponding to the preset time-domain indicators is established. Based on the evaluation function, the adjusted velocity loop trajectory model is evaluated, and the evaluation score is calculated. Then, based on the evaluation score, it is determined whether the current optimization parameters obtained by the preset mode search algorithm have target parameters.

[0055] Step S240: When the number of optimization attempts and the current evaluation score meet the preset conditions, the optimization parameters obtained in the current attempt are confirmed as the target parameters.

[0056] By comparing the current number of optimization attempts and the current evaluation score with preset conditions, the parameter optimization is terminated when the number of optimization attempts and the current evaluation score meet the preset conditions, and the optimization parameters obtained in the current attempt are confirmed as the target parameters, thus realizing the optimal solution for output parameter optimization.

[0057] With preset time-domain indicators as constraints, an evaluation function corresponding to the preset time-domain indicators is designed, and a parameter self-tuning strategy based on a preset pattern search algorithm is constructed. Driven by the evaluation function of the time-domain indicators, the parameter self-tuning strategy iteratively searches within the set optimization range to generate a candidate set of control parameters. Time-domain data is collected using simulation or hardware-in-the-loop method to support parameter iterative optimization, thus forming the entire parameter self-tuning process based on time-domain indicators.

[0058] For example, by building an offline simulation model of the robot, designing the corresponding excitation trajectory, initializing the preset mode search algorithm parameters, optimizing the velocity loop proportional coefficient and velocity loop integral coefficient respectively, adjusting the robot servo system according to the optimized parameters, rerunning the trajectory, calculating the evaluation function of the corresponding preset time domain index, and after the number of optimizations and the current evaluation score meet the preset conditions, the optimized parameters obtained in the current iteration are confirmed as the target parameters, thereby achieving efficient and highly reliable self-tuning of robot servo parameters.

[0059] In the above embodiments, a velocity loop trajectory model of the robot servo system is obtained; based on a preset pattern search algorithm, the velocity loop parameters of the velocity loop trajectory model are optimized to obtain optimized parameters; and the velocity loop trajectory model is adjusted according to the optimized parameters to obtain an adjusted velocity loop trajectory model; an evaluation function is established according to a preset time-domain index, and the adjusted velocity loop trajectory model is evaluated according to the evaluation function to obtain an evaluation score; when the number of optimization attempts and the current evaluation score meet preset conditions, the optimized parameters obtained in the current attempt are confirmed as target parameters, thereby realizing the self-tuning of the velocity loop parameters of the robot servo system. This application optimizes the velocity loop parameters of the velocity loop trajectory model through a preset pattern search algorithm and evaluates the parameter optimization through a preset time-domain index, effectively reducing the difficulty of parameter tuning and simplifying the parameter optimization implementation process. The parameter self-tuning strategy is designed based on time-domain indexes, eliminating the need for precise model structures and parameters, and achieving highly reliable parameter self-tuning.

[0060] In one embodiment, the preset time-domain indicators include motor-side trajectory accuracy, motor-side end-effector vibration, motor-side end-effector convergence time, and motor-side corner overshoot.

[0061] Among these, motor-side trajectory accuracy refers to the degree to which the real-time path of the motor's moving parts closely approximates the theoretical path during motion. Motor-side end-effector vibration can be categorized into process vibration and residual vibration based on time. Motor-side end-effector convergence time refers to the time required for the motor control system to reach a stable state. Motor-side corner overshoot refers to the numerical deviation in speed and position that occurs at trajectory corners.

[0062] For example, the steps for establishing an evaluation function based on preset time-domain indicators include:

[0063] The evaluation function is obtained based on any one of the following: motor-side trajectory accuracy, motor-side end vibration, motor-side end convergence time, and motor-side corner overshoot; or, the evaluation function is obtained based on any combination of the following: motor-side trajectory accuracy, motor-side end vibration, motor-side end convergence time, and motor-side corner overshoot, and the combination is weighted.

[0064] Among them, motor-side trajectory accuracy, motor-side end-effector vibration, motor-side end-effector convergence time, and motor-side corner overshoot each correspond to a specific evaluation function. For example, any one of these four time-domain indicators can be selected as a time-domain index, and a corresponding evaluation function can be established based on the selected time-domain index. Alternatively, by using these four time-domain indicators as combined constraints and weighting the evaluation functions of the corresponding time-domain indicators, an evaluation function combining the four time-domain indicators can be obtained. This approach fully considers the performance of the industrial robot in various aspects, eliminates the need for precise model structures and parameters, effectively reduces the difficulty of parameter tuning, and simplifies the parameter optimization process.

[0065] In one embodiment, when the preset time-domain index is the motor-side trajectory accuracy, the evaluation function is:

[0066] ,in, This represents the maximum accuracy of each trajectory segment before optimization. To achieve the maximum accuracy of each trajectory segment after optimization, The evaluation score corresponds to the accuracy of the motor-side trajectory. This is the preset expected score.

[0067] By using the motor-side trajectory accuracy as a time-domain indicator, the trajectory is divided into N parts based on inflection points, and the maximum accuracy of each part is found. Evaluation Score Less than the preset expected score For example, if the evaluation score Compared with the preset expected score If the ratio is less than 1, the current evaluation score of the corresponding motor-side trajectory accuracy is determined to meet the preset conditions; otherwise, it is not.

[0068] In one embodiment, when the preset time-domain index is the vibration at the motor end, the evaluation function is:

[0069] ,in, To optimize the amplitude of the corresponding frequency component before optimization, To optimize the amplitude of the corresponding frequency component, This represents the maximum value of the amplitude ratio before and after optimization of the corresponding process vibration. This represents the maximum value of the amplitude ratio before and after optimization for the corresponding residual vibration. This is the evaluation score for the vibration at the motor end.

[0070] By using the motor-side end-vibration as a time-domain indicator, a Fast Fourier Transform (FFT) analysis is performed on the end-trajectory, which can be divided into process vibration and residual vibration according to time. It should be noted that process vibration corresponds to the vibration occurring during the command execution, while residual vibration corresponds to the vibration occurring after the command ends. The evaluation score for the motor-side end-vibration is equal to the average of the maximum ratio of the amplitude before and after optimization of the process vibration and the maximum ratio of the amplitude before and after optimization of the residual vibration. Less than the preset expected score of the vibration at the corresponding motor end. For example, if the evaluation score Compared with the preset expected score If the ratio is less than 1, the current evaluation score of the vibration at the end of the corresponding motor side is determined to meet the preset condition; otherwise, it is not met.

[0071] In one embodiment, when the preset time-domain index is motor-side corner overshoot, the evaluation function is:

[0072] ,in, To address the corner overshoot at each inflection point before optimization, To optimize the corner overshoot at each inflection point, This is the evaluation score for the motor side corner overshoot.

[0073] By using motor-side corner overshoot as a time-domain indicator, the corner deviation at each command inflection point is calculated. The evaluation score for motor-side corner overshoot is the maximum value of the corner overshoot ratio before and after optimization at each inflection point. The score is less than the preset expected score of the corresponding motor side corner overshoot. For example, if the evaluation score Compared with the preset expected score If the ratio is less than 1, the current evaluation score of the corresponding motor side corner overshoot meets the preset condition; otherwise, it does not.

[0074] In one embodiment, when the preset time-domain metric is the motor-side end-convergence time, the evaluation function is:

[0075] ,in, The convergence time before optimization. The optimized convergence time, This is the evaluation score for the convergence time at the motor side.

[0076] By using the motor-side end-effector convergence time as a time-domain indicator, the time it takes for the trajectory accuracy curve to converge to the desired threshold band after the motion command is completed is calculated, and the corresponding evaluation score for the motor-side end-effector convergence time is obtained. The expected score is less than the corresponding motor-side end convergence time. For example, if the evaluation score Compared with the preset expected score If the ratio is less than 1, the current evaluation score of the corresponding motor-side end convergence time is determined to meet the preset condition; otherwise, it is not met.

[0077] In one example, the evaluation functions for the corresponding motor-side trajectory accuracy, the corresponding motor-side end vibration, the corresponding motor-side end convergence time, and the corresponding motor-side corner overshoot are weighted to obtain the overall evaluation function:

[0078] .

[0079] in, These are the weighting coefficients of the evaluation function corresponding to the accuracy of the motor-side trajectory. These are the weighting coefficients for the evaluation function corresponding to the vibration at the motor end. These are the weighting coefficients of the evaluation function corresponding to the convergence time at the motor side. These are the weighting coefficients of the evaluation function corresponding to the motor side corner overshoot.

[0080] It should be noted that when optimizing based on time-domain metrics, the robot is required to run the work trajectory multiple times until the performance requirements are met or the optimal number of iterations is reached. The optimization process can employ hardware-in-the-loop online optimization or offline optimization using a robot simulation model.

[0081] In the above embodiments, a parameter self-tuning strategy is designed based on four different time-domain indices: motor-side trajectory accuracy, motor-side end-effector vibration, motor-side corner overshoot, and motor-side end-effector convergence time. This strategy is used to tune the speed loop proportional coefficient and speed loop integral coefficient, achieving PID parameter tuning without requiring precise model structure and parameters. Compared to traditional model-based parameter self-tuning methods that rely on precise model structure and parameters to calculate PID parameters, this application uses time-domain indices as the driving force. Parameter optimization can be performed through an optimization algorithm using an offline model or hardware-in-the-loop approach, effectively reducing the difficulty of parameter tuning. Furthermore, by designing appropriate time-domain index weighting functions, the robustness of the optimized parameters is ensured.

[0082] In one embodiment, such as Figure 3 As shown, based on a preset pattern search algorithm, the steps for optimizing the velocity loop parameters of the velocity loop trajectory model to obtain the optimized parameters include:

[0083] Step S310, Axial Search Step: Based on the preset mode search algorithm, perform axial search on the velocity loop parameters of the velocity loop trajectory model to obtain the first intermediate parameters.

[0084] Specifically, a coordinate system is established for the trajectory of the corresponding velocity loop trajectory model, and the axial search starts from the starting point of the corresponding coordinate system. Begin the search, probing along each coordinate direction to obtain the next point. Thus, the first intermediate parameter is obtained.

[0085] Step S320, Pattern Search Step: Based on the first intermediate parameter, perform a pattern search on the velocity loop parameters of the velocity loop trajectory model to obtain the second intermediate parameter.

[0086] After axial search, based on the base point and starting point corresponding to the first intermediate parameter, along the direction of the line connecting adjacent base points (e.g. The pattern search is performed in the direction of ) to obtain a current optimal extreme value, that is, to obtain the second intermediate parameter.

[0087] Step S330: Iterate through the axial search step and the pattern search step until the number of iterations reaches the preset first threshold to obtain the optimization parameters.

[0088] By continuously iterating through the axial and pattern searches of the velocity loop parameters of the velocity loop trajectory model until the termination condition of the search iteration is met, that is, when the number of iterations reaches a preset first threshold, the iterative search step is terminated, thereby obtaining the optimal parameters.

[0089] In the above embodiments, by using a preset pattern search algorithm, the velocity loop parameters of the velocity loop trajectory model are iteratively searched for axial direction and pattern, which effectively reduces the difficulty of parameter tuning and simplifies the parameter optimization process.

[0090] In one embodiment, after iteratively executing the axial search step and the pattern search step until the number of iterations reaches a preset threshold and the optimization parameters are obtained, the following steps are included:

[0091] Obtain the global optimal solution value and the global suboptimal solution value of the corresponding optimization parameters; perform difference processing on the global optimal solution value and the global suboptimal solution value to obtain the baseline value; based on the baseline value, perform pattern search on the velocity loop parameters of the velocity loop trajectory model to obtain the optimized optimization parameters.

[0092] To enhance the local search capability of the preset pattern search algorithm, after completing one iteration and calculating the optimization parameters, the global optimal solution can be used. and global suboptimal solution The difference between the two values ​​is used as a benchmark value to narrow the iteration range. The pattern search is performed with the benchmark value as the center to gradually obtain the global optimal solution, thereby enhancing the local search capability of the algorithm.

[0093] For example, the specific iterative formula is as follows: , .in, It is a uniformly distributed random number; It is iteration The scaling factor at this time, F is the number of iterations.

[0094] By improving the pattern search algorithm, this application guides the search direction by comparing function values, thus placing very low demands on the objective function. Even if the function is discontinuous, non-differentiable, or contains noise, the search parameters can be stably searched. At the same time, the existence of the benchmark value greatly improves the algorithm's local search capability.

[0095] In the above embodiments, the improved pattern search algorithm is used to optimize the velocity loop parameters of the velocity loop trajectory model, which further reduces the difficulty of parameter tuning, simplifies the parameter optimization process, and achieves highly reliable parameter self-tuning without the need for precise model structure and parameters.

[0096] In one embodiment, the step of confirming the optimization parameters obtained in the current iteration as the target parameters when the number of optimization attempts and the current evaluation score meet preset conditions includes:

[0097] When the number of optimization attempts reaches a preset first threshold and / or the evaluation score reaches a preset second threshold, the optimization parameters obtained in the current attempt are confirmed as target parameters.

[0098] For example, by comparing the current number of optimization attempts with a preset first threshold and the current evaluation score with a preset second threshold, and based on the comparison results, when the current number of optimization attempts reaches the preset first threshold or the evaluation score reaches the preset second threshold, the parameter optimization is terminated, and the optimization parameters obtained in the current attempt are confirmed as the target parameters, thus achieving the optimal solution for output parameter optimization.

[0099] For example, by comparing the current number of optimization attempts with a preset first threshold and the current evaluation score with a preset second threshold, and based on the comparison results, when the current number of optimization attempts reaches the preset first threshold and the evaluation score reaches the preset second threshold, the parameter optimization is terminated, and the optimization parameters obtained in the current attempt are confirmed as the target parameters, thus achieving the optimal solution for output parameter optimization.

[0100] It should be understood that, although Figures 2 to 3 The steps in the flowchart are shown sequentially as indicated by the arrows, but these steps are not necessarily executed in the order indicated by the arrows. Unless otherwise specified herein, there is no strict order in which these steps are executed, and they can be performed in other orders. Figures 2 to 3At least some of the steps in the process may include multiple sub-steps or multiple stages. These sub-steps or stages are not necessarily completed at the same time, but can be executed at different times. The execution order of these sub-steps or stages is not necessarily sequential, but can be executed in turn or alternately with other steps or at least some of the sub-steps or stages of other steps.

[0101] In one embodiment, a robot parameter self-tuning device is also provided, comprising:

[0102] The model acquisition unit is used to acquire the velocity loop trajectory model of the robot servo system.

[0103] The pattern search unit is used to optimize the velocity loop parameters of the velocity loop trajectory model based on a preset pattern search algorithm, obtain the optimized parameters, and adjust the velocity loop trajectory model according to the optimized parameters to obtain the adjusted velocity loop trajectory model.

[0104] The evaluation unit is used to establish an evaluation function based on preset time-domain indicators, and to evaluate the adjusted velocity loop trajectory model based on the evaluation function to obtain an evaluation score.

[0105] The optimization judgment unit is used to confirm the optimization parameters obtained in the current iteration as the target parameters when the number of optimization iterations and the current evaluation score meet preset conditions.

[0106] Specific limitations regarding the robot parameter self-tuning device can be found in the limitations of the robot parameter self-tuning method described above, and will not be repeated here. Each module in the aforementioned robot parameter self-tuning device can be implemented entirely or partially through software, hardware, or a combination thereof. These modules can be embedded in hardware or independently of the processor in the robot parameter self-tuning system, or stored in software in the memory of the robot parameter self-tuning system, so that the processor can call and execute the corresponding operations of each module.

[0107] In one embodiment, a robot parameter self-tuning system is also provided, including a processing device. The processing device includes a memory and a processor. The memory stores a computer program, and the processor executes the computer program to implement the steps of any of the robot parameter self-tuning methods described above.

[0108] For example, the processor is used to execute the following steps of a robot parameter self-tuning method:

[0109] Obtain the velocity loop trajectory model of the robot servo system; based on a preset pattern search algorithm, optimize the velocity loop parameters of the velocity loop trajectory model to obtain optimized parameters, and adjust the velocity loop trajectory model according to the optimized parameters to obtain the adjusted velocity loop trajectory model; establish an evaluation function according to a preset time domain index, and evaluate the adjusted velocity loop trajectory model according to the evaluation function to obtain an evaluation score; when the number of optimization attempts and the current evaluation score meet preset conditions, the optimized parameters obtained in the current attempt are confirmed as target parameters.

[0110] In the above embodiments, the velocity loop parameters of the velocity loop trajectory model are optimized by a preset pattern search algorithm, and the parameter optimization is evaluated by a preset time domain index. This effectively reduces the difficulty of parameter tuning and simplifies the parameter optimization process. The parameter self-tuning strategy is designed based on the time domain index, which does not require a precise model structure and parameters, and achieves highly reliable parameter self-tuning.

[0111] In one embodiment, a computer storage medium is also provided, on which a computer program is stored, which, when executed by a processor, implements the steps of any of the above-described active battery balancing control methods.

[0112] For example, when a computer program is executed by a processor, it performs the following steps:

[0113] Obtain the velocity loop trajectory model of the robot servo system; based on a preset pattern search algorithm, optimize the velocity loop parameters of the velocity loop trajectory model to obtain optimized parameters, and adjust the velocity loop trajectory model according to the optimized parameters to obtain the adjusted velocity loop trajectory model; establish an evaluation function according to a preset time domain index, and evaluate the adjusted velocity loop trajectory model according to the evaluation function to obtain an evaluation score; when the number of optimization attempts and the current evaluation score meet preset conditions, the optimized parameters obtained in the current attempt are confirmed as target parameters.

[0114] Those skilled in the art will understand that all or part of the processes in the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium. When executed, the computer program can include the processes of the embodiments of the division operations described above. Any references to memory, storage, databases, or other media used in the embodiments provided in this application can include non-volatile and / or volatile memory. Non-volatile memory can include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM), or flash memory. Volatile memory can include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in various forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), dual data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous link DRAM (SLDRAM), direct memory bus RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM), etc.

[0115] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0116] The embodiments described above are merely illustrative of several implementation methods of this application, and while the descriptions are relatively specific and detailed, they should not be construed as limiting the scope of the patent application. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of this application, and these all fall within the protection scope of this application. Therefore, the protection scope of this patent application should be determined by the appended claims.

[0117] Those skilled in the art will understand that all or part of the processes in the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium. When executed, the computer program can include the processes of the embodiments of the division operations described above. Any references to memory, storage, databases, or other media used in the embodiments provided in this application can include non-volatile and / or volatile memory. Non-volatile memory can include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM), or flash memory. Volatile memory can include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in various forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), dual data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous link DRAM (SLDRAM), direct memory bus RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM), etc.

[0118] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0119] The embodiments described above are merely illustrative of several implementation methods of this application, and while the descriptions are relatively specific and detailed, they should not be construed as limiting the scope of the patent application. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of this application, and these all fall within the protection scope of this application. Therefore, the protection scope of this patent application should be determined by the appended claims.

Claims

1. A method for self-tuning robot parameters, characterized in that, Includes the following steps: Obtain the velocity loop trajectory model of the robot servo system; Based on a preset pattern search algorithm, the velocity loop parameters of the velocity loop trajectory model are optimized to obtain optimized parameters. Based on the optimized parameters, the velocity loop trajectory model is adjusted to obtain the adjusted velocity loop trajectory model. An evaluation function is established based on a preset time-domain index, and the adjusted velocity loop trajectory model is evaluated based on the evaluation function to obtain an evaluation score. When the number of optimization attempts and the current evaluation score meet the preset conditions, the optimization parameters obtained in the current attempt are confirmed as the target parameters. The preset time-domain indicators include motor-side trajectory accuracy, motor-side end vibration, motor-side end convergence time, and motor-side corner overshoot. The step of establishing the evaluation function based on the preset time-domain index includes: The evaluation function is obtained by weighting any combination of the motor-side trajectory accuracy, the motor-side end vibration, the motor-side end convergence time, and the motor-side corner overshoot. The step of optimizing the velocity loop parameters of the velocity loop trajectory model based on the preset pattern search algorithm to obtain the optimized parameters includes: Axial search step: Based on the preset mode search algorithm, perform axial search on the velocity loop parameters of the velocity loop trajectory model to obtain the first intermediate parameter; Pattern search step: Based on the first intermediate parameter, perform a pattern search on the velocity loop parameters of the velocity loop trajectory model to obtain the second intermediate parameter; The axial search step and the pattern search step are iteratively executed until the number of iterations reaches a preset first threshold, and the optimization parameters are obtained. The step of iteratively executing the axial search step and the pattern search step until the number of iterations reaches a preset first threshold to obtain the optimization parameters includes: Obtain the global optimal solution value and the global suboptimal solution value corresponding to the optimization parameters; The difference between the global optimal solution value and the global suboptimal solution value is processed to obtain a baseline value; Based on the baseline value, a pattern search is performed on the velocity loop parameters of the velocity loop trajectory model to obtain the optimized optimization parameters.

2. The robot parameter self-tuning method according to claim 1, characterized in that, When the preset time-domain index is the motor-side trajectory accuracy, the evaluation function is: ， in, This represents the maximum accuracy of each trajectory segment before optimization. To achieve the maximum accuracy of each trajectory segment after optimization, This corresponds to the evaluation score for the accuracy of the motor-side trajectory. This is the preset expected score.

3. The robot parameter self-tuning method according to claim 1, characterized in that, When the preset time-domain index is the vibration at the motor end, the evaluation function is: ， in, To optimize the amplitude of the corresponding frequency component before optimization, To optimize the amplitude of the corresponding frequency component, This represents the maximum value of the amplitude ratio before and after optimization of the corresponding process vibration. This represents the maximum value of the amplitude ratio before and after optimization for the corresponding residual vibration. This is the evaluation score for the vibration at the motor end.

4. The robot parameter self-tuning method according to claim 1, characterized in that, When the preset time-domain index is the motor side corner overshoot, the evaluation function is: ， in, To address the corner overshoot at each inflection point before optimization, To optimize the corner overshoot at each inflection point, This is the evaluation score corresponding to the overshoot of the motor side corner.

5. The robot parameter self-tuning method according to claim 1, characterized in that, When the preset time-domain index is the convergence time at the motor end, the evaluation function is: ， in, The convergence time before optimization. The optimized convergence time, This is the evaluation score corresponding to the convergence time at the end of the motor side.

6. The robot parameter self-tuning method according to any one of claims 1 to 5, characterized in that, The step of confirming the optimization parameters obtained in the current iteration as the target parameters when the number of optimization attempts and the current evaluation score meet the preset conditions includes: When the number of optimization attempts reaches a preset first threshold and / or the evaluation score reaches a preset second threshold, the optimization parameters obtained in the current attempt are confirmed as target parameters.

7. A robot parameter self-tuning system, characterized in that, The device includes a processing unit, which includes a memory and a processor. The memory stores a computer program, and the processor executes the computer program to implement the steps of the robot parameter self-tuning method according to any one of claims 1 to 6.

Images