Motion control domain model testing method and system based on big data analysis

Abstract

The application relates to the technical field of data analysis, and particularly discloses a motion control domain model testing method and system based on big data analysis, which comprises a comprehensive evaluation module; the comprehensive evaluation module obtains a comprehensive evaluation coefficient according to running data in a next testing action completion process after an interference signal is added, then obtains an evaluation score after all testing actions in the action set are completed, and outputs an evaluation result according to the evaluation score; the evaluation result includes that a target robot exists a running risk and the test is passed; the application dynamically adjusts the intensity of the interference signal according to the action evaluation coefficient through a disturbance injection module, realizes self-adaptive rigorous testing in the testing process, the comprehensive evaluation module extracts dynamic response indexes such as a rise time, a maximum adjustment amount and a stable time, calculates the comprehensive evaluation coefficient, and judges whether the model passes the test by comparing with a preset threshold value.

Description

Technical Field

[0001] This invention relates to the field of data analysis technology, specifically to a method and system for testing motion control domain models based on big data analysis. Background Technology

[0002] The motion control threshold model is a scheme applied to the joint and limb control of bionic robots. This scheme believes that the nervous system controls the position of the limb by setting an equilibrium point, and the limb will eventually stabilize near this equilibrium point. Thus, by setting an optimal threshold, a human-like motion generation and control framework is provided for the robot, aiming to create robots with more natural movements, stronger adaptability, and safer interaction.

[0003] However, the actual movement of the human body is more flexible than described by this model, and it can quickly correct itself when encountering obstacles. Therefore, it is necessary to test the robot that uses this model. Existing tests are based on results, and the robot is evaluated by the success rate of completing the target task. However, this method is obviously not suitable for evaluating the performance of this type of robot when dealing with obstacles.

[0004] In view of this, the present invention proposes a motion control domain model testing method and system based on big data analysis, which can conduct more accurate testing of robots controlled by motion control domain models through the analysis of a large amount of data. Summary of the Invention

[0005] The purpose of this invention is to provide a method and system for testing motion control domain models based on big data analysis, and to solve the following technical problems: How to conduct more accurate tests on robots controlled by motion control domain models through analysis of large amounts of data.

[0006] The objective of this invention can be achieved through the following technical solutions: A motion control domain model testing system based on big data analytics includes: The action management module has a preset set of actions for testing the target robot. The action set includes several test actions. The target robot is a bionic robot equipped with a motion control domain model. The data acquisition module is used to acquire the operating data of the target robot and perform data preprocessing. The data preprocessing includes data cleaning, data alignment and data standardization. The operating data includes tracking error. The big data analysis module is used to extract feature parameters of the target robot's running data under a preset motion stage, which includes at least an acceleration stage, a constant speed stage, and a deceleration stage. The motion evaluation module evaluates the target robot after it completes the test motions in the motion set and obtains motion evaluation coefficients. The disturbance injection module adjusts the disturbance signal according to the action evaluation coefficient of the previous completed test action before the start of the next test action. The disturbance signal includes friction torque pulse, inertia change and cogging effect simulation signal. The comprehensive evaluation module obtains a comprehensive evaluation coefficient based on the running data during the completion of the next test action after the addition of an interference signal. Then, after all test actions within the action set are completed, an evaluation score is obtained, and an evaluation result is output based on the evaluation score. The evaluation result includes whether the target robot has operational risks or whether the test has passed.

[0007] The above technical solution, as described in this embodiment, utilizes a big data analysis module to statistically analyze historical data and extract the mean and standard deviation of tracking errors for each preset motion stage (acceleration, constant speed, deceleration) as benchmark feature parameters. Based on this, a perturbation injection module dynamically adjusts the intensity of the interference signal according to the motion evaluation coefficient, achieving adaptive and rigorous testing during the testing process. Finally, a comprehensive evaluation module extracts dynamic response indicators such as rise time, maximum adjustment amount, and settling time, calculates a comprehensive evaluation coefficient, and determines whether the model passes the test by comparing it with a preset threshold. This invention upgrades traditional static threshold comparison to dynamic risk assessment based on big data, enabling the discovery of potential performance degradation of the model under complex operating conditions. Furthermore, by integrating tracking error with multiple time-domain dynamic indicators, it achieves a multi-dimensional comprehensive evaluation of the model's response speed, stability, and robustness, resulting in more comprehensive and accurate results. Simultaneously, the introduction of an adaptive perturbation injection mechanism effectively exposes defects such as nonlinearity and parameter perturbation that are difficult to detect in conventional testing. It also features three levels of safety protection, balancing the aggressiveness of the test with equipment safety.

[0008] As a further technical solution of the present invention: the process of extracting feature parameters of the target robot's running data under a preset motion phase includes: Obtain the tracking error from historical operational data under undisturbed conditions; Through formula and Calculate the mean and standard deviation of the error for each preset motion stage, where N is the total number of sampling points within a preset motion stage. It is the tracking error value of the i-th sampling point. Obviously, the mean and standard deviation of the error are one of the characteristic parameters of the target robot's running data.

[0009] As a further technical solution of the present invention: the process of obtaining motion evaluation coefficients includes: Record action set One test action; Real-time acquisition of tracking error sequence under current test conditions ; Through formula Calculate the motion evaluation coefficient ; in The mean of the tracking error after completing the current test action. This is the theoretical error curve of the target robot under standard conditions. It is a preset error integration threshold. and These are the preset first weight value and the second weight value, respectively.

[0010] As a further technical solution of the present invention: the process of adjusting the interference signal includes: Set the initial strength according to the type of interference signal added. ; Through the formula: Get adjustment multiplier The adjustment amount is ; in and For different ranges of gain coefficients and , and The risk coefficient threshold is determined based on historical data. It is the maximum magnification allowed under the initial strength.

[0011] The above technical solution provides a method for dynamically applying disturbances during the testing process. The interference signal of the present invention is dynamically adjusted based on the performance of the testing process. Therefore, it can better test the anti-interference limit of the target robot and provide more diverse data to facilitate big data screening and statistics, thus realizing a virtuous cycle of big data analysis and testing.

[0012] As a further technical solution of the present invention: the process of obtaining the comprehensive evaluation coefficient based on the running data during the completion of the next test action after adding the interference signal includes: Obtain the application time of the interference signal and the dynamic response index after application; Through formula Calculate the comprehensive evaluation coefficient ; in It refers to the rise time, which is the time it takes for the steady-state value of the response to increase from 10% to 90% in the dynamic response metric. That is the maximum adjustment amount. It is the curve equation for the standard rise time that varies with the adjustment amount. The standard rise time curve equation is plotted based on empirical data. This refers to the allowable deviation in rise time. This is a preset influence coefficient, set according to the preset motion phase in which the interference signal is applied. For example, if the interference signal is applied during the acceleration phase, then... =0.8, if it is in the uniform speed range then =1, , and These are the preset indicator weight values, which are constant values ​​set according to the indicator type, usually between 0 and 1.

[0013] As a further technical solution of the present invention: the process of outputting evaluation results based on evaluation scores includes: Obtain the overall evaluation coefficient for each action after completing a set of m test actions. And calculate the sum of the comprehensive evaluation coefficients except for the first action. ; Sum value With preset threshold Compare; like ≤ If so, it is determined that the target robot has operational risks; like > If the test is passed, the target robot is deemed to have passed.

[0014] As a further technical solution of the present invention: maximum adjustment amount The acquisition process includes: Obtain the response curve showing the steady-state value of the dynamic response index as a function of time after completing a test action, and locate the peak position in the curve. ; Through formula Calculate the maximum adjustment amount ; in It is the final value of the response curve after completing a test action. It is the initial value of the response curve of a test action of the product.

[0015] As a further technical solution of the present invention: the sampling rate of the steady-state response value is at least 10 times the closed-loop bandwidth of the system.

[0016] As a further technical solution of the present invention: the friction torque pulse, inertia mutation and cogging effect simulation signal in the interference signal are synchronously applied to the target robot test process and the same adjustment ratio is used.

[0017] This invention also provides a method for testing motion control domain models based on big data analysis, comprising the following steps: Extract the feature parameters of the target robot's running data under a preset motion phase, which includes at least an acceleration phase, a constant speed phase, and a deceleration phase. The data acquisition module is used to acquire the target robot's operational data and perform data preprocessing. The operational data includes tracking errors. The robot is evaluated and its action evaluation coefficient is obtained after it completes the test actions in the action set. Before the next test action begins, the interference signal is adjusted based on the action evaluation coefficient of the previous completed test action; The comprehensive evaluation coefficient is obtained based on the running data during the completion of the next test action after the interference signal is added. Then, the evaluation score is obtained after all test actions in the action set are completed, and the evaluation result is output based on the evaluation score. The evaluation result includes whether the target robot has operational risks and whether the test is passed.

[0018] The beneficial effects of this invention are: (1) The disturbance injection module of the present invention dynamically adjusts the intensity of the interference signal according to the action evaluation coefficient to realize adaptive and rigorous testing during the testing process. Finally, the comprehensive evaluation module extracts dynamic response indicators such as rise time, maximum adjustment amount, and settling time, calculates the comprehensive evaluation coefficient, and determines whether the model passes the test by comparing it with the preset threshold. The present invention upgrades the traditional static threshold comparison to dynamic risk assessment based on big data, which can discover the potential performance degradation of the model under complex working conditions; in addition, it integrates tracking error and multiple time-domain dynamic indicators to realize a multi-dimensional comprehensive evaluation of the model's response speed, stability and robustness, and the results are more comprehensive and accurate; at the same time, it introduces an adaptive disturbance injection mechanism to effectively stimulate defects such as nonlinearity and parameter perturbation that are difficult to expose by conventional testing, and is equipped with three-level safety protection to balance the aggressiveness of testing and equipment safety.

[0019] (2) The interference signal of the present invention is dynamically adjusted based on the performance of the test process. Therefore, it can better test the anti-interference limit of the target robot during the test process and provide more diverse data to facilitate the screening and statistics of big data, so as to realize a virtuous cycle of big data analysis and testing process. Attached Figure Description

[0020] The invention will now be further described with reference to the accompanying drawings.

[0021] Figure 1 This is a schematic diagram of the module composition of the testing system of the present invention; Figure 2 This is a flowchart of the testing method steps of the present invention. Detailed Implementation

[0022] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0023] Please see Figure 1 As shown, in one embodiment, a motion control domain model testing system based on big data analysis is provided, including: The action management module has a preset set of actions for testing the target robot. The action set includes several test actions. The target robot is a bionic robot equipped with a motion control domain model. The data acquisition module is used to acquire the operating data of the target robot and perform data preprocessing. The data preprocessing includes data cleaning, data alignment and data standardization. The operating data includes tracking error. The big data analysis module is used to extract feature parameters of the target robot's running data under a preset motion stage, which includes at least an acceleration stage, a constant speed stage, and a deceleration stage. The motion evaluation module evaluates the target robot after it completes the test motions in the motion set and obtains motion evaluation coefficients. The disturbance injection module adjusts the disturbance signal according to the action evaluation coefficient of the previous completed test action before the start of the next test action. The disturbance signal includes friction torque pulse, inertia change and cogging effect simulation signal. The comprehensive evaluation module obtains a comprehensive evaluation coefficient based on the running data during the completion of the next test action after the addition of an interference signal. Then, after all test actions within the action set are completed, an evaluation score is obtained, and an evaluation result is output based on the evaluation score. The evaluation result includes whether the target robot has operational risks or whether the test has passed.

[0024] This embodiment utilizes a big data analysis module to statistically analyze historical data, extracting the mean and standard deviation of tracking errors for each preset motion stage (acceleration, constant speed, deceleration) as benchmark feature parameters. Based on this, a perturbation injection module dynamically adjusts the intensity of the interference signal according to the motion evaluation coefficient, achieving adaptive and rigorous testing during the testing process. Finally, a comprehensive evaluation module extracts dynamic response indicators such as rise time, maximum adjustment amount, and settling time, calculates a comprehensive evaluation coefficient, and determines whether the model passes the test by comparing it with a preset threshold. This invention upgrades traditional static threshold comparison to dynamic risk assessment based on big data, enabling the discovery of potential performance degradation of the model under complex operating conditions. Furthermore, by integrating tracking error with multiple time-domain dynamic indicators, it achieves a multi-dimensional comprehensive evaluation of the model's response speed, stability, and robustness, resulting in more comprehensive and accurate results. Simultaneously, the introduction of an adaptive perturbation injection mechanism effectively exposes defects such as nonlinearity and parameter perturbations that are difficult to detect in conventional testing. It also features three levels of safety protection, balancing the aggressiveness of the test with equipment safety.

[0025] The process of extracting feature parameters from the target robot's operational data under a preset motion phase includes: Obtain the tracking error from historical operational data under undisturbed conditions; Through formula and Calculate the mean and standard deviation of the error for each preset motion stage, where N is the total number of sampling points within a preset motion stage. It is the tracking error value of the i-th sampling point. Obviously, the mean and standard deviation of the error are one of the characteristic parameters of the target robot's running data.

[0026] The process of obtaining action evaluation coefficients includes: Record action set One test action; Real-time acquisition of tracking error sequence under current test conditions ; Through formula Calculate the motion evaluation coefficient ; in The mean of the tracking error after completing the current test action. This is the theoretical error curve of the target robot under standard conditions. It is a preset error integration threshold. and These are the preset first and second weight values, for example... and .

[0027] The process of adjusting interference signals includes: Set the initial strength according to the type of interference signal added. ; Through the formula: Get adjustment multiplier The adjustment amount is ; in and For different ranges of gain coefficients and , and The risk coefficient threshold is determined based on historical data, for example... and , This is the maximum magnification allowed under the initial strength, and it should be set according to the specific situation, for example... .

[0028] This embodiment provides a scheme for dynamically applying disturbances during the testing process. The interference signal of the present invention is dynamically adjusted based on the performance during the testing process. Therefore, it can better test the anti-interference limit of the target robot and provide more diverse data to facilitate big data screening and statistics, thereby realizing a virtuous cycle of big data analysis and testing.

[0029] The process of obtaining the comprehensive evaluation coefficient based on the operational data during the completion of the next test action after the addition of interference signals includes: Obtain the application time of the interference signal and the dynamic response index after application; Through formula Calculate the comprehensive evaluation coefficient ; in It refers to the rise time, which is the time it takes for the steady-state value of the response to increase from 10% to 90% in the dynamic response metric. That is the maximum adjustment amount. It is the curve equation for the standard rise time that varies with the adjustment amount. The standard rise time curve equation is plotted based on empirical data. This refers to the allowable deviation in rise time. This is a preset influence coefficient, set according to the preset motion phase in which the interference signal is applied. For example, if the interference signal is applied during the acceleration phase, then... =0.8, if it is in the uniform speed range then =1, , and These are the preset indicator weight values, which are constant values ​​set according to the indicator type, usually between 0 and 1.

[0030] The process of outputting evaluation results based on the evaluation score includes: Obtain the overall evaluation coefficient for each action after completing a set of m test actions. And calculate the sum of the comprehensive evaluation coefficients except for the first action. ; Sum value With preset threshold Compare; like ≤ If so, it is determined that the target robot has operational risks; like > If the test is passed, the target robot is deemed to have passed.

[0031] Maximum adjustment The acquisition process includes: Obtain the response curve showing the steady-state value of the dynamic response index as a function of time after completing a test action, and locate the peak position in the curve. ; Through formula Calculate the maximum adjustment amount ; in It is the final value of the response curve after completing a test action. It is the initial value of the response curve of a test action of the product.

[0032] The sampling rate of the steady-state response should be at least 10 times the closed-loop bandwidth of the system. For example, if the bandwidth is 100Hz, the sampling rate should be ≥1kHz to accurately capture peak values.

[0033] The friction torque pulse, inertia mutation, and tooth cogging effect simulation signals in the interference signal are synchronously applied to the target robot during the testing process and are adjusted at the same rate.

[0034] refer to Figure 2 This embodiment also provides a motion control domain model testing method based on big data analysis, including the following steps: S1. Extract the feature parameters of the target robot's running data under the preset motion stage. The preset motion stage includes at least an acceleration phase, a constant speed phase, and a deceleration phase. S2, the acquisition module is used to acquire the target robot's operating data and perform data preprocessing. The operating data includes tracking error; S3. Evaluate the target robot after it completes the test actions in the action set and obtain the action evaluation coefficient; S4. Before the next test action begins, adjust the interference signal according to the action evaluation coefficient of the previous completed test action; S5. Obtain the comprehensive evaluation coefficient based on the running data during the completion of the next test action after the addition of the interference signal. Then, obtain the evaluation score after all test actions in the action set are completed, and output the evaluation result based on the evaluation score. The evaluation result includes whether the target robot has operational risks and whether the test is passed.

[0035] The foregoing has provided a detailed description of one embodiment of the present invention, but this description is merely a preferred embodiment and should not be construed as limiting the scope of the invention. All equivalent variations and modifications made within the scope of the claims of this invention should still fall within the patent coverage of this invention.

Claims

1. A motion control domain model testing system based on big data analysis, characterized in that, include: The action management module has a preset set of actions for testing the target robot. The action set includes several test actions. The target robot is a bionic robot equipped with a motion control domain model. The data acquisition module is used to acquire the operating data of the target robot and perform data preprocessing. The operating data includes tracking error. The big data analysis module is used to extract feature parameters of the target robot's running data under a preset motion stage, which includes at least an acceleration stage, a constant speed stage, and a deceleration stage. The motion evaluation module evaluates the target robot after it completes the test motions in the motion set and obtains motion evaluation coefficients. The disturbance injection module adjusts the disturbance signal according to the action evaluation coefficient of the previous completed test action before the start of the next test action. The disturbance signal includes friction torque pulse, inertia change and cogging effect simulation signal. The comprehensive evaluation module obtains a comprehensive evaluation coefficient based on the running data during the completion of the next test action after the addition of an interference signal. Then, after all test actions within the action set are completed, an evaluation score is obtained, and an evaluation result is output based on the evaluation score. The evaluation result includes whether the target robot has operational risks or whether the test has passed.

2. The motion control domain model testing system based on big data analysis according to claim 1, characterized in that, The process of extracting feature parameters from the target robot's operational data under a preset motion phase includes: Obtain the tracking error from historical operational data under undisturbed conditions; Through formula and Calculate the mean and standard deviation of the error for each preset motion stage, where N is the total number of sampling points within a preset motion stage. It is the tracking error value of the i-th sampling point.

3. The motion control domain model testing system based on big data analysis according to claim 1, characterized in that, The process of obtaining action evaluation coefficients includes: Record action set One test action; Real-time acquisition of tracking error sequence under current test conditions ; Through formula Calculate the motion evaluation coefficient ; in The mean of the tracking error after completing the current test action. This represents the theoretical error curve of the target robot under standard conditions. It is a preset error integration threshold. and These are the preset first weight value and the second weight value, respectively.

4. The motion control domain model testing system based on big data analysis according to claim 1, characterized in that, The process of adjusting interference signals includes: Set the initial strength according to the type of interference signal added. ; Through the formula: Get adjustment multiplier The adjustment amount is ; in and For different ranges of gain coefficients and , and The risk coefficient threshold is determined based on historical data. It is the maximum magnification allowed under the initial strength.

5. The motion control domain model testing system based on big data analysis according to claim 1, characterized in that, The process of obtaining the comprehensive evaluation coefficient based on the operational data during the completion of the next test action after the addition of interference signals includes: Obtain the application time of the interference signal and the dynamic response index after application; Through formula Calculate the comprehensive evaluation coefficient ; in It's the rising time. That is the maximum adjustment amount. It is the curve equation for the standard rise time that varies with the adjustment amount. This refers to the allowable deviation in rise time. It is a preset influence coefficient, set according to the preset motion phase in which the interference signal is applied. , and These are the preset indicator weight values.

6. The motion control domain model testing system based on big data analysis according to claim 5, characterized in that, The process of outputting evaluation results based on the evaluation score includes: Obtain the overall evaluation coefficient for each action after completing a set of m test actions. And calculate the sum of the comprehensive evaluation coefficients except for the first action. ; Sum value With preset threshold Compare; like ≤ If so, it is determined that the target robot has operational risks; like > If the test is passed, the target robot is deemed to have passed.

7. The motion control domain model testing system based on big data analysis according to claim 5, characterized in that, Maximum adjustment The acquisition process includes: Obtain the response curve showing the steady-state value of the response over time after completing a test action, and locate the peak position in the curve. ; Through formula Calculate the maximum adjustment amount ; in It is the final value of the response curve after completing a test action. It is the initial value of the response curve of a test action of the product.

8. The motion control domain model testing system based on big data analysis according to claim 7, characterized in that, The sampling rate of the steady-state response value is at least 10 times the closed-loop bandwidth of the system.

9. The motion control domain model testing system based on big data analysis according to claim 1, characterized in that, The interference signals, including friction torque pulses, inertia mutations, and simulated cogging effects, are applied synchronously to the target robot during the testing process and are adjusted at the same rate.

10. A motion control domain model testing method based on big data analysis, applied to the motion control domain model testing system based on big data analysis as described in claims 1-9, characterized in that, Includes the following steps: Extract the feature parameters of the target robot's running data under a preset motion phase, which includes at least an acceleration phase, a constant speed phase, and a deceleration phase. The data acquisition module is used to acquire the target robot's operational data and perform data preprocessing. The operational data includes tracking errors. The robot is evaluated and its action evaluation coefficient is obtained after it completes the test actions in the action set. Before the next test action begins, the interference signal is adjusted based on the action evaluation coefficient of the previous completed test action; The comprehensive evaluation coefficient is obtained based on the running data during the completion of the next test action after the interference signal is added. Then, the evaluation score is obtained after all test actions in the action set are completed, and the evaluation result is output based on the evaluation score. The evaluation result includes whether the target robot has operational risks and whether the test is passed.

Images