Multi-axis synchronous error compensation system and method for six-degree-of-freedom platform

Abstract

The invention belongs to the technical field of six-degree-of-freedom platform control, and particularly relates to a multi-axis synchronous error compensation system and method for a six-degree-of-freedom platform, and the method comprises the steps: firstly, sequentially carrying out the comprehensive collection, preprocessing and credibility marking of motion parameters, environmental parameters and load parameters of each axis of the six-degree-of-freedom platform; standard data with credibility marks are output; then, based on the standardized data with the credibility marks, establishing a multi-dimensional synchronization error model so as to accurately calculate synchronization errors, and outputting error component data; then, on the basis of the error component data, a dynamic compensation instruction is generated in combination with a compensation algorithm, and a qualified compensation instruction set is output after inspection; on the basis, by building a preset algorithm library containing a plurality of algorithms, configuring an adaptive compensation algorithm unit and combining working condition parameters such as load and temperature which are collected in real time, dynamic adaptive selection of the compensation algorithm is achieved.

Description

Technical Field

[0001] This invention relates to the field of six-degree-of-freedom platform control technology, specifically to a multi-axis synchronous error compensation system and method for a six-degree-of-freedom platform. Background Technology

[0002] Six-degree-of-freedom (6DOF) platforms, with their ability to achieve multi-directional spatial motion, are widely used in high-end fields such as aerospace simulation, industrial robot debugging, and precision machining. One of the core performance indicators of such platforms is the synchronization accuracy of multi-axis motion, as the synchronization error directly determines the realism of the simulated environment, the pass rate of processed products, and the stability of equipment operation.

[0003] As application scenarios demand increasingly higher precision, the challenge of multi-axis synchronous control in six-degree-of-freedom platforms is gradually increasing. Current technologies often employ a "centralized measurement + empirical compensation" approach for multi-axis synchronous error measurement and compensation. This involves acquiring motion data from a subset of axes using a single measurement module and combining it with a pre-defined compensation algorithm to correct errors. However, this approach has several limitations in practical applications and struggles to meet the demands of high-precision scenarios, as detailed below:

[0004] 1. Poor adaptability of compensation algorithms and lack of dynamic correction mechanism: Existing compensation algorithms are mostly based on preset fixed formulas, calibrated only for specific working conditions (such as fixed load and fixed motion speed), and cannot dynamically adjust the compensation strategy according to the real-time operating status of the platform. For example, in precision machining scenarios, when a six-degree-of-freedom platform drives a tool to process workpieces of different thicknesses (the load changes from 5kg to 10kg), the original compensation algorithm, because it does not incorporate real-time load data, causes the multi-axis synchronization error to increase from 0.02mm to 0.08mm, and the surface roughness of the machined workpiece exceeds the acceptable range. Summary of the Invention

[0005] To address the technical problems of poor adaptability and lack of dynamic correction mechanism in the aforementioned compensation algorithms, this invention provides the following technical solution:

[0006] A multi-axis synchronization error compensation system for a six-degree-of-freedom platform, comprising:

[0007] The multi-dimensional data acquisition module comprehensively collects, preprocesses, and marks the motion parameters, environmental parameters, and load parameters of each axis of the six-degree-of-freedom platform, and outputs standardized data with credibility marking.

[0008] The error analysis and modeling module, based on the standardized data with confidence labels output by the multi-dimensional data acquisition module, establishes a multi-dimensional synchronization error model to accurately calculate the synchronization error and output error component data.

[0009] The long-term performance degradation prediction and pre-compensation module, based on the historical standardized data of the multi-dimensional data acquisition module and the historical error component data of the error analysis and modeling module, performs historical data mining and trend prediction to generate pre-compensation parameters. It then uses a weighted fusion algorithm to combine the parameters with the real-time error component data to output fused error data.

[0010] The dynamic compensation strategy generation module, based on the fused error data output from the long-term performance degradation prediction and pre-compensation modules, first uses a compensation priority sorting unit to prioritize each error component by combining the application scenario weights of the six-degree-of-freedom platform, outputting an error compensation priority list. Then, the adaptive compensation algorithm unit, based on the error compensation priority list and standardized data with confidence markers, adaptively selects a matching compensation algorithm from a preset algorithm library and outputs compensation algorithm parameters adapted to the current working condition. Subsequently, the compensation instruction generation unit, based on the compensation algorithm parameters and the fused error data, generates specific compensation instructions for each axis according to the error compensation priority list and outputs a compensation instruction set with timing markers. Finally, the compensation instruction feasibility verification unit performs dual verification on the compensation instruction set based on standardized data with confidence markers and preset platform physical constraint parameters, outputting a qualified compensation instruction set.

[0011] The real-time compensation execution module accurately executes the compensation instructions based on the qualified compensation instruction set output by the dynamic compensation strategy generation module, and provides real-time feedback and outputs the compensation execution results.

[0012] The closed-loop optimization and fault diagnosis module optimizes system parameters based on the compensation execution results output by the real-time compensation execution module, while simultaneously achieving accurate fault diagnosis and fault-tolerant handling.

[0013] As a preferred embodiment of the multi-axis synchronization error compensation system for a six-degree-of-freedom platform described in this invention, the multi-dimensional data acquisition module includes:

[0014] The multi-source sensor unit uses a laser displacement sensor, angle encoder, force sensor, and temperature sensor to form a multi-source sensor array, which collects real-time displacement and rotation data, real-time load data, and ambient temperature data of each axis, and outputs raw data.

[0015] The data preprocessing unit, based on the raw data output by the multi-source sensor group unit, removes noise interference from the raw data using the Kalman filter algorithm, and standardizes the output data of different types of sensors to output standardized data.

[0016] The data credibility labeling unit calculates the data fluctuation coefficient based on the standardized data output by the data preprocessing unit, and classifies each group of data into high credibility and low credibility levels based on preset thresholds, and outputs standardized data with credibility labels.

[0017] As a preferred embodiment of the multi-axis synchronous error compensation system for a six-degree-of-freedom platform described in this invention, the error analysis and modeling module includes:

[0018] The multi-axis coupling coefficient calculation unit is based on standardized data with confidence labels output by the multi-dimensional data acquisition module. It selects high-confidence data as the basis for calculation and calculates the coupling coefficient between each axis under different motion conditions through the finite element analysis method, and outputs a dynamic coupling coefficient matrix.

[0019] The multi-axis synchronization error modeling unit, based on the dynamic coupling coefficient matrix output by the multi-axis coupling coefficient calculation unit and combined with the standardized data with confidence markers output by the multi-dimensional data acquisition module, establishes a multi-dimensional synchronization error model of displacement error, rotation angle error and coupling error, and inputs motion parameters, load parameters and temperature parameters of each axis, and outputs preliminary synchronization error values.

[0020] The error decomposition unit decomposes the preliminary synchronization error value output by the multi-axis synchronization error modeling unit into systematic error, random error and coupling error, and outputs the decomposed error component data.

[0021] As a preferred embodiment of the multi-axis synchronous error compensation system for a six-degree-of-freedom platform described in this invention, the real-time compensation execution module includes:

[0022] The compensation instruction parsing unit, based on the qualified compensation instruction set output by the dynamic compensation strategy generation module, parses the axis number, compensation parameters and execution timing corresponding to each compensation instruction, and outputs the parsing results;

[0023] The drive control unit, based on the analysis results output by the compensation command analysis unit, precisely controls the drive motors of each axis of the six-degree-of-freedom platform, executes compensation actions in sequence, and simultaneously collects and outputs the operating parameters of the drive motors in real time.

[0024] The compensation execution feedback unit calculates the actual error correction amount of the compensation action based on the operating parameters of the drive motor output by the drive control unit and the standardized data with confidence markers output by the multi-dimensional data acquisition module, and outputs the compensation execution result.

[0025] As a preferred embodiment of the multi-axis synchronous error compensation system for a six-degree-of-freedom platform described in this invention, the closed-loop optimization and fault diagnosis module includes:

[0026] The compensation effect evaluation unit calculates the multi-axis synchronization error value after compensation based on the compensation execution result output by the real-time compensation execution module, compares it with the preset accuracy threshold, and outputs three categories of evaluation compensation effect: qualified, needs optimization, and unqualified.

[0027] When the evaluation result of the compensation effect evaluation unit is that optimization is needed or unqualified, the system parameter optimization unit adjusts the multi-dimensional synchronous error model parameters of the error analysis and modeling module and the compensation algorithm parameters of the dynamic compensation strategy generation module in reverse based on the compensation execution result.

[0028] The fault diagnosis unit, based on the standardized data with confidence labels output by the multi-dimensional data acquisition module and the compensation execution results output by the real-time compensation execution module, will locate the fault location and output a fault diagnosis report when multiple consecutive sets of data are low confidence or the compensation execution results are consistently unqualified, by using a fault feature matching algorithm.

[0029] The fault-tolerant processing unit takes corresponding fault-tolerant measures based on the fault diagnosis report output by the fault diagnosis unit.

[0030] As a preferred embodiment of the multi-axis synchronization error compensation system for a six-degree-of-freedom platform described in this invention, the long-term performance degradation prediction and pre-compensation module includes:

[0031] The historical data storage and feature extraction unit, based on the historical standardized data from the multi-dimensional data acquisition module and the historical error component data from the error analysis and modeling module, first classifies and stores them, then extracts attenuation features through feature engineering, and outputs a structured attenuation feature dataset.

[0032] The attenuation trend prediction unit, based on the attenuation feature dataset output by the historical data storage and feature extraction unit, adopts an LSTM-based prediction model and combines the platform's cumulative runtime and the design life parameters of key components to establish a performance attenuation prediction model for each axis, so as to output the attenuation error prediction value of each axis within a future preset period, and mark the prediction confidence level.

[0033] The pre-compensation parameter generation unit generates pre-compensation parameters based on the attenuation error prediction value and prediction confidence output by the attenuation trend prediction unit, combined with a preset attenuation threshold.

[0034] The pre-compensation and real-time error fusion unit, based on the pre-compensation parameters output by the pre-compensation parameter generation unit and the current error component data output by the error analysis and modeling module, uses a weighted fusion algorithm to fuse the pre-compensation parameters and the real-time error component data, and outputs the fused error data to the dynamic compensation strategy generation module.

[0035] A method for multi-axis synchronization error compensation of a six-degree-of-freedom platform includes the following specific steps:

[0036] S1, Multi-dimensional Data Acquisition:

[0037] S11, Multi-source sensor array: It adopts a multi-source sensor array composed of laser displacement sensor, angle encoder, force sensor and temperature sensor to collect real-time displacement and rotation data, real-time load data and ambient temperature data of each axis, and output raw data.

[0038] S12, Data Preprocessing: Based on the raw data output from the multi-source sensor group steps, noise interference in the raw data is removed by Kalman filtering algorithm, and the output data of different types of sensors are standardized to output standardized data.

[0039] S13, Data confidence labeling: Based on the standardized data output from the data preprocessing step, calculate the data fluctuation coefficient, and combine it with a preset threshold to classify each group of data into high confidence and low confidence levels, and output standardized data with confidence labels.

[0040] S2, Error Analysis and Modeling:

[0041] S21, Multi-axis coupling coefficient calculation: Based on the standardized data with confidence labels output from the multi-dimensional data acquisition steps, high-confidence data is selected as the basis for calculation. The coupling coefficient between each axis under different motion conditions is calculated by the finite element analysis method, and the dynamic coupling coefficient matrix is ​​output.

[0042] S22, Multi-axis synchronization error modeling: Based on the dynamic coupling coefficient matrix output by the multi-axis coupling coefficient calculation step, combined with the standardized data with confidence markers output by the multi-dimensional data acquisition step, a multi-dimensional synchronization error model of displacement error, rotation angle error and coupling error is established, and the motion parameters, load parameters and temperature parameters of each axis are input, and the preliminary synchronization error value is output.

[0043] S23, Error Decomposition: Based on the preliminary synchronization error value output from the multi-axis synchronization error modeling step, decompose it into systematic error, random error and coupling error, and output the decomposed error component data;

[0044] S3, Long-term performance degradation prediction and pre-compensation:

[0045] S31, Historical Data Storage and Feature Extraction: Based on the historical standardized data from the multi-dimensional data acquisition steps and the historical error component data from the error analysis and modeling steps, the data is first classified and stored, and then the decay features are extracted through feature engineering, and a structured decay feature dataset is output.

[0046] S32, Attenuation Trend Prediction: Based on the attenuation feature dataset output from the historical data storage and feature extraction steps, an LSTM-based prediction model is used, combined with the platform's cumulative runtime and the design life parameters of key components, to establish a performance attenuation prediction model for each axis, so as to output the attenuation error prediction value of each axis within the future preset period, and mark the prediction confidence level.

[0047] S33, Pre-compensation parameter generation: Based on the attenuation error prediction value and prediction confidence output by the attenuation trend prediction step, and combined with the preset attenuation threshold, pre-compensation parameters are generated.

[0048] S34, Pre-compensation and real-time error fusion: Based on the pre-compensation parameters output by the pre-compensation parameter generation step and the current error component data output by the error analysis and modeling step, a weighted fusion algorithm is used to fuse the pre-compensation parameters and the real-time error component data, and the fused error data is output to the dynamic compensation strategy generation step.

[0049] S4, Dynamic Compensation Strategy Generation: Based on the fused error data output from the long-term performance degradation prediction and pre-compensation steps, the error components are first prioritized by a compensation priority sorting step, combined with the application scenario weights of the six-degree-of-freedom platform, and an error compensation priority list is output. Then, an adaptive compensation algorithm step adaptively selects a matching compensation algorithm from a preset algorithm library based on the error compensation priority list and standardized data with confidence markers, and outputs compensation algorithm parameters adapted to the current working condition. Subsequently, a compensation instruction generation step generates specific compensation instructions for each axis according to the error compensation priority list based on the compensation algorithm parameters and the fused error data, and outputs a compensation instruction set with time sequence markers. Finally, a compensation instruction feasibility verification step performs double verification on the compensation instruction set based on standardized data with confidence markers and preset platform physical constraint parameters, and outputs a qualified compensation instruction set.

[0050] S5, Real-time Compensation Execution:

[0051] S51, Compensation Instruction Parsing: Based on the dynamic compensation strategy, the qualified compensation instruction set output by the generation step is parsed, the axis number, compensation parameters and execution timing corresponding to each compensation instruction are parsed, and the parsing results are output.

[0052] S52, Drive Control: Based on the analysis results output from the compensation instruction analysis steps, the drive motors of each axis of the six-degree-of-freedom platform are precisely controlled, and compensation actions are executed in sequence. At the same time, the operating parameters of the drive motors are collected in real time and output.

[0053] S53, Compensation Execution Feedback: Based on the operating parameters of the drive motor output by the drive control step, combined with the standardized data with confidence markers output by the multi-dimensional data acquisition step, calculate the actual error correction amount of the compensation action, and output the compensation execution result.

[0054] S6, Closed-loop optimization and fault diagnosis:

[0055] S61, Compensation effect evaluation: Based on the compensation execution results output by the real-time compensation execution steps, calculate the multi-axis synchronization error value after compensation, compare it with the preset accuracy threshold, and output three categories of evaluation compensation effect: qualified, needs optimization, and unqualified.

[0056] S62, System parameter optimization: When the evaluation result of the compensation effect evaluation step is that it needs to be optimized or is unqualified, the multi-dimensional synchronous error model parameters of the error analysis and modeling step and the compensation algorithm parameters of the dynamic compensation strategy generation step are adjusted in reverse based on the compensation execution result.

[0057] S63, Fault Diagnosis: Based on the standardized data with confidence labels output from the multi-dimensional data acquisition steps and the compensation execution results output from the real-time compensation execution steps, when multiple consecutive sets of data are detected as low confidence or the compensation execution results are continuously unqualified, the fault location is located through the fault feature matching algorithm, and a fault diagnosis report is output.

[0058] S64, Fault Tolerance: Based on the fault diagnosis report output by the fault diagnosis steps, take corresponding fault tolerance measures.

[0059] Compared with existing technologies:

[0060] 1. By building a preset algorithm library containing multiple algorithms and equipping it with an adaptive compensation algorithm unit, combined with real-time collected load, temperature and other operating parameters, the system can dynamically adapt and select compensation algorithms. The system can automatically match the most suitable compensation algorithm and dynamically adjust the algorithm parameters according to changes in the platform's operating status. It can adapt to different operating conditions without manual intervention, completely solving the problem of poor adaptability of fixed algorithms, ensuring the accuracy and applicability of compensation strategies under various operating conditions, and greatly improving the system's adaptability to complex and changing operating conditions.

[0061] 2. By integrating data credibility marking and compensation execution result analysis, and combining fault feature matching and fault tolerance processing mechanisms, it can achieve accurate location and rapid response of faulty units. After a fault is detected, fault tolerance measures can be activated in a timely manner to avoid the expansion of errors caused by faulty data and ensure the continuous and stable operation of the system. Attached Figure Description

[0062] Figure 1 This is a schematic diagram of the overall framework of the present invention;

[0063] Figure 2 This is a schematic diagram of the multi-dimensional data acquisition module framework of the present invention;

[0064] Figure 3 This is a schematic diagram of the error analysis and modeling module framework of the present invention;

[0065] Figure 4 This is a schematic diagram of the dynamic compensation strategy generation module framework of the present invention;

[0066] Figure 5 This is a schematic diagram of the real-time compensation execution module framework of the present invention;

[0067] Figure 6 This is a schematic diagram of the closed-loop optimization and fault diagnosis module framework of the present invention;

[0068] Figure 7 This is a schematic diagram of the long-term performance degradation prediction and pre-compensation module framework of the present invention. Detailed Implementation

[0069] To make the objectives, technical solutions, and advantages of the present invention clearer, the embodiments of the present invention will be described in further detail below with reference to the accompanying drawings.

[0070] This invention provides a multi-axis synchronization error compensation system for a six-degree-of-freedom platform. Please refer to [link / reference]. Figure 1 ,include:

[0071] The multi-dimensional data acquisition module comprehensively collects, preprocesses, and marks the motion parameters, environmental parameters, and load parameters of each axis of the six-degree-of-freedom platform, and outputs standardized data with credibility marking.

[0072] The error analysis and modeling module, based on the standardized data with confidence labels output by the multi-dimensional data acquisition module, establishes a multi-dimensional synchronization error model to accurately calculate the synchronization error and output error component data.

[0073] The long-term performance degradation prediction and pre-compensation module, based on the historical standardized data of the multi-dimensional data acquisition module and the historical error component data of the error analysis and modeling module, performs historical data mining and trend prediction to generate pre-compensation parameters. It then uses a weighted fusion algorithm to combine the parameters with the real-time error component data to output fused error data.

[0074] The dynamic compensation strategy generation module generates dynamic compensation instructions based on the fused error data from the long-term performance degradation prediction and pre-compensation module, combined with the compensation algorithm. After verification, it outputs a qualified set of compensation instructions.

[0075] The real-time compensation execution module accurately executes the compensation instructions based on the qualified compensation instruction set output by the dynamic compensation strategy generation module, and provides real-time feedback and outputs the compensation execution results.

[0076] The closed-loop optimization and fault diagnosis module optimizes system parameters based on the compensation execution results output by the real-time compensation execution module, while simultaneously achieving accurate fault diagnosis and fault-tolerant processing.

[0077] The long-term performance degradation prediction and pre-compensation module, based on the historical standardized data from the multi-dimensional data acquisition module and the historical error component data from the error analysis and modeling module, performs historical data mining and trend prediction to generate pre-compensation parameters, which are then combined with real-time error component data and output to the dynamic compensation strategy generation module.

[0078] Please see Figure 2 The multi-dimensional data acquisition module includes:

[0079] The multi-source sensor unit consists of a laser displacement sensor, an angle encoder, a force sensor, and a temperature sensor, forming a multi-source sensor array. It collects real-time displacement and rotation data, real-time load data, and ambient temperature data for each axis and outputs the raw data. Specifically, the laser displacement sensor and angle encoder are installed on the platform's six motion axes (X, Y, Z, A, B, and C axes) to collect real-time displacement and rotation data for each axis; the force sensor is installed on the platform's worktable to collect real-time load data; and the temperature sensor is installed at the drive motor of each axis to collect ambient temperature data.

[0080] The data preprocessing unit, based on the raw data output by the multi-source sensor group unit, removes noise interference (such as instantaneous fluctuation data of sensors) from the raw data through the Kalman filter algorithm, and performs standardization processing on the output data of different types of sensors (unifying the data format and units) to output standardized data.

[0081] The data credibility labeling unit calculates the data fluctuation coefficient (such as the standard deviation of displacement data) based on the standardized data output by the data preprocessing unit, and classifies each group of data into high credibility and low credibility levels based on preset thresholds, and outputs standardized data with credibility labels.

[0082] Please see Figure 3 The error analysis and modeling module includes:

[0083] The multi-axis coupling coefficient calculation unit is based on standardized data with confidence labels output by the multi-dimensional data acquisition module. It selects high-confidence data as the basis for calculation and calculates the coupling coefficients between axes (such as the mechanical coupling coefficient between the X-axis and the A-axis) under different motion conditions (different loads, different speeds) using the finite element analysis method, and outputs the dynamic coupling coefficient matrix.

[0084] The multi-axis synchronization error modeling unit, based on the dynamic coupling coefficient matrix output by the multi-axis coupling coefficient calculation unit, and combined with the standardized data with confidence markers (such as standardized displacement / rotation data) output by the multi-dimensional data acquisition module, establishes a multi-dimensional synchronization error model of displacement error, rotation error, and coupling error. It also inputs the motion parameters, load parameters, and temperature parameters of each axis and outputs the preliminary synchronization error value.

[0085] The error decomposition unit decomposes the preliminary synchronization error value output by the multi-axis synchronization error modeling unit into systematic errors (such as fixed errors caused by machining deviations), random errors (such as temporary errors caused by load fluctuations), and coupling errors (such as interactive errors caused by multi-axis linkage), and outputs the decomposed error component data.

[0086] Please see Figure 4 The dynamic compensation strategy generation module includes:

[0087] The compensation priority sorting unit, based on the fused error data output by the long-term performance degradation prediction and pre-compensation module, and combined with the application scenario weights of the six-degree-of-freedom platform (such as the Z-axis displacement error having a higher priority than other axes in precision machining scenarios), sorts the priority of each error component and outputs an error compensation priority list.

[0088] Specifically: The system integrates error data from the long-term performance degradation prediction and pre-compensation modules, while also reading the system's pre-configured application scenario configuration file (which users can pre-configure according to their actual needs). The sorting criteria include three dimensions: first, error type weight (systematic error weight > coupled error weight > random error weight; because systematic errors are stable, prioritizing their correction can significantly improve overall accuracy); second, error amplitude weight (the larger the error amplitude, the higher the priority; a linear weighting formula is used to calculate the amplitude weight coefficient); and third, axis importance weight (e.g., in precision machining scenarios, the Z-axis (vertical direction) weight is set to 0.3, the X / Y axes to 0.25, and the A / B / C axes to 0.2; in aerospace simulation scenarios, the A / B / C axis weight is increased to 0.3). The Analytic Hierarchy Process (AHP) is used to calculate the comprehensive priority score of each error component, outputting a structured error compensation priority list containing "error type, corresponding axis number, compensation priority, and priority score," providing clear guidance for subsequent algorithm selection.

[0089] The adaptive compensation algorithm unit, based on the error compensation priority list output by the compensation priority sorting unit, and combined with the standardized data with confidence labels (such as real-time parameters such as load and temperature) output by the multi-dimensional data acquisition module, adaptively selects a matching compensation algorithm from the preset algorithm library (including PID compensation algorithm, fuzzy PID compensation algorithm, and neural network compensation algorithm) (such as selecting the neural network compensation algorithm under high load conditions), and outputs compensation algorithm parameters adapted to the current working condition.

[0090] Specifically: Based on the error compensation priority list output by the compensation priority sorting unit, standardized data with confidence markers are collected simultaneously (focusing on extracting real-time operating parameters such as load, temperature, and shaft speed). The preset algorithm library includes three core algorithms and adaptation rules: 1. Basic PID compensation algorithm, suitable for low load (≤10kg), low speed (≤10mm / s), and gradual error changes; 2. Fuzzy PID compensation algorithm, suitable for medium load (10-30kg), medium speed (10-20mm / s), and small error fluctuations; 3. Neural network compensation algorithm (using BP neural network), suitable for high load (>30kg), high speed (>20mm / s), and complex multi-axis strongly coupled operating conditions. The unit has a built-in working condition matching model. First, it performs an initial matching algorithm based on the real-time working condition. Then, it adjusts the algorithm parameters by combining the error compensation priority list (e.g., a smaller PID proportional coefficient corresponds to a high priority error to avoid overshoot; the weight of the coupling term in the neural network is enhanced when the proportion of coupling error is high). Finally, it outputs compensation algorithm parameters that are adapted to the current working condition, including "algorithm type, core parameters, and working condition description", to ensure that the algorithm is accurately matched with the current operating state.

[0091] The compensation instruction generation unit generates specific compensation instructions (including compensation displacement, compensation speed, and compensation timing) for each axis based on the error data fused from the compensation algorithm parameters output by the adaptive compensation algorithm unit and the long-term performance degradation prediction and pre-compensation module, according to the compensation priority list, and outputs a set of compensation instructions with timing marks.

[0092] Specifically: Based on the error data fused from the compensation algorithm parameters output by the adaptive compensation algorithm unit and the long-term performance degradation prediction and pre-compensation module, compensation instructions are generated according to the principle of "high-priority errors are compensated first, and errors of the same priority are compensated synchronously." The instruction dimensions include: 1. Basic parameters (axis number, compensation direction, compensation displacement, compensation speed, compensation acceleration), where the compensation acceleration is dynamically set according to the axis load (the greater the load, the smaller the acceleration, to avoid overloading the drive motor); 2. Timing parameters (instruction start time, compensation duration, synchronous trigger signals for each axis), with timing accuracy controlled within 1ms to ensure multi-axis synchronous compensation; 3. Safety parameters (instruction execution threshold, such as triggering a pause mechanism when the compensation displacement exceeds a preset range). The final output is a compensation instruction set with timing markers.

[0093] The compensation command feasibility verification unit performs dual verification on the compensation command set output by the compensation command generation unit. This verification is based on standardized data with credibility markers (such as real-time motion status data of each axis (displacement, velocity)) output by the multi-dimensional data acquisition module and preset platform physical constraint parameters (such as displacement limits, maximum acceleration, and joint rotation range of each axis). On the one hand, it verifies whether the single-axis compensation command exceeds the physical constraints (such as whether the Z-axis compensation displacement exceeds the platform's lifting limit). On the other hand, it verifies the coordination of multi-axis synchronous compensation commands (such as whether the combination of X-axis and Y-axis compensation velocities causes the platform's center of gravity to deviate beyond the limit). If the verification passes, a qualified compensation command set is directly output to the real-time compensation execution module. If the verification fails, an unqualified command is marked and fed back to the adaptive compensation algorithm unit. At the same time, a constraint conflict prompt is output, and the adaptive compensation algorithm unit re-optimizes the compensation algorithm parameters based on the constraint conditions to ensure that the generated compensation commands are both accurate and safe.

[0094] Specifically, the core function is to mitigate the problem of compensation commands exceeding the platform's physical limits or causing operational risks. Its data input includes three parts: 1. The compensation command set output by the compensation command generation unit; 2. Real-time motion status data (displacement, velocity, acceleration) of each axis output by the multi-source sensor group unit; 3. The system's built-in platform physical constraint parameter library (including displacement limits, maximum velocity / acceleration, joint rotation angle range, maximum power of the drive motor, platform center of gravity offset threshold, etc., the parameter library supports factory calibration and subsequent calibration updates). Verification is divided into two steps: the first step is single-axis verification, checking whether the displacement, velocity, and other parameters of each axis's compensation command are within the physical constraint range; if they exceed, it is marked as "single-axis constraint violation"; the second step is multi-axis collaborative verification, calculating the platform center of gravity offset and joint torque when multiple axes synchronously execute compensation commands using a dynamic simulation model; if they exceed the safety threshold, it is marked as "collaborative constraint violation". When the verification passes, a qualified set of compensation instructions is directly output to the real-time compensation execution module; when the verification fails, a detailed feedback report containing "violation instruction number, violation type, and specific parameters exceeding the standard" is output to the adaptive compensation algorithm unit. The adaptive compensation algorithm unit adjusts the algorithm parameters based on the reason for the violation (such as reducing the compensation displacement of the exceeding axis and reducing the compensation speed), and regenerates the compensation instructions, forming a small closed loop of "generation-verification-optimization".

[0095] Please see Figure 5 The real-time compensation execution module includes:

[0096] The compensation instruction parsing unit, based on the qualified compensation instruction set output by the dynamic compensation strategy generation module, parses the axis number, compensation parameters and execution timing corresponding to each compensation instruction, and outputs the parsing results;

[0097] The drive control unit, based on the analysis results output by the compensation command analysis unit, precisely controls the drive motors of each axis of the six-degree-of-freedom platform, executes compensation actions in sequence (such as controlling the X-axis motor to rotate an additional preset angle to offset synchronization errors), and simultaneously collects and outputs the operating parameters of the drive motors in real time (such as speed and current).

[0098] The compensation execution feedback unit calculates the actual error correction amount of the compensation action based on the operating parameters of the drive motor output by the drive control unit, combined with the standardized data with confidence markers (such as real-time displacement / rotation angle data) output by the multi-dimensional data acquisition module, and outputs the compensation execution result (including statuses such as "compensation in place" and "compensation deviation").

[0099] Please see Figure 6 The closed-loop optimization and fault diagnosis module includes:

[0100] The compensation effect evaluation unit calculates the multi-axis synchronization error value after compensation based on the compensation execution result output by the real-time compensation execution module, compares it with the preset accuracy threshold, and outputs three categories of evaluation compensation effect: qualified, needs optimization, and unqualified.

[0101] When the evaluation result of the compensation effect evaluation unit indicates that optimization is needed or that the system parameter optimization unit is unqualified, it adjusts the parameters of the multi-dimensional synchronous error model of the error analysis and modeling module and the compensation algorithm parameters of the dynamic compensation strategy generation module in reverse based on the compensation execution result (such as adjusting the weight values ​​in the coupling coefficient matrix).

[0102] The fault diagnosis unit, based on the standardized data with confidence labels output by the multi-dimensional data acquisition module and the compensation execution results output by the real-time compensation execution module, will locate the fault location (such as sensor fault or drive motor fault) by using a fault feature matching algorithm when it detects that multiple consecutive sets of data are of low confidence or the compensation execution results are continuously unqualified, and will output a fault diagnosis report.

[0103] The fault-tolerant processing unit, based on the fault diagnosis report output by the fault diagnosis unit, takes corresponding fault-tolerant measures (such as calling redundant sensor data to replace the faulty sensor when the sensor fails; and temporarily adjusting the compensation strategy to reduce the motion weight of the faulty axis when the drive motor fails) to ensure the continuous and stable operation of the system.

[0104] Please see Figure 7 The long-term performance degradation prediction and pre-compensation module includes:

[0105] The historical data storage and feature extraction unit, based on the historical standardized data (including motion parameters, load, and temperature of each axis) from the multi-dimensional data acquisition module and the historical error component data from the error analysis and modeling module, first stores the data by daily / weekly / monthly dimensions, and then extracts the core features related to attenuation through feature engineering, including: the cumulative change of error amplitude of each axis, the error growth rate, the difference in error deviation under different loads, the correlation coefficient between motor temperature and error, etc., and outputs a structured attenuation feature dataset.

[0106] The attenuation trend prediction unit, based on the attenuation feature dataset output by the historical data storage and feature extraction unit, adopts an LSTM (Long Short-Term Memory)-based prediction model, combined with the platform's cumulative runtime and the design life parameters of key components (such as lead screws and joint bearings), to establish a performance attenuation prediction model for each axis, so as to output the attenuation error prediction value of each axis within a future preset period (such as the next 7 days or 30 days), and mark the prediction confidence (based on the reverse correction of historical prediction accuracy).

[0107] The pre-compensation parameter generation unit generates pre-compensation parameters (including compensation axis number, compensation amount, implementation cycle, and credibility level) based on the attenuation error prediction value and prediction confidence level output by the attenuation trend prediction unit, combined with a preset attenuation threshold (configurable for different application scenarios). If the prediction error does not exceed the threshold, "progressive pre-compensation parameters" are generated (compensation is applied step by step according to the time gradient). If the prediction error exceeds the threshold, "emergency pre-compensation parameters" are generated and an early warning signal is triggered (prompting maintenance personnel to check the components).

[0108] The pre-compensation and real-time error fusion unit, based on the pre-compensation parameters output by the pre-compensation parameter generation unit and the current error component data output by the error analysis and modeling module, uses a weighted fusion algorithm (the pre-compensation weight is dynamically adjusted according to the prediction confidence, and the higher the confidence, the greater the weight) to fuse the pre-compensation parameters with the real-time error component data, and outputs the fused error data to the dynamic compensation strategy generation module to achieve the synergy between pre-compensation and real-time compensation.

[0109] A method for multi-axis synchronization error compensation of a six-degree-of-freedom platform includes the following specific steps:

[0110] S1, Multi-dimensional Data Acquisition:

[0111] S11, Multi-source sensor array: This array comprises a laser displacement sensor, an angle encoder, a force sensor, and a temperature sensor. It collects real-time displacement and rotation data, real-time load data, and ambient temperature data for each axis, and outputs the raw data. The laser displacement sensor and angle encoder are installed on the platform's six motion axes (X, Y, Z, A, B, and C axes) to collect real-time displacement and rotation data. The force sensor is installed on the platform's worktable to collect real-time load data. The temperature sensor is installed at the drive motor of each axis to collect ambient temperature data.

[0112] S12, Data Preprocessing: Based on the raw data output from the multi-source sensor group steps, noise interference (such as instantaneous fluctuation data of the sensors) in the raw data is removed by Kalman filtering algorithm, and the output data of different types of sensors are standardized (unified data format and units) to output standardized data.

[0113] S13, Data Confidence Labeling: Based on the standardized data output from the data preprocessing step, calculate the data fluctuation coefficient (such as the standard deviation of displacement data), and combine it with a preset threshold to classify each group of data as high or low confidence, and output standardized data with confidence labels.

[0114] S2, Error Analysis and Modeling:

[0115] S21, Multi-axis Coupling Coefficient Calculation: Based on the standardized data with confidence labels output from the multi-dimensional data acquisition steps, high-confidence data is selected as the basis for calculation. The coupling coefficients between axes (such as the mechanical coupling coefficient between the X-axis and the A-axis) under different motion conditions (different loads, different speeds) are calculated using the finite element analysis method, and the dynamic coupling coefficient matrix is ​​output.

[0116] S22, Multi-axis synchronization error modeling: Based on the dynamic coupling coefficient matrix output from the multi-axis coupling coefficient calculation step, combined with the standardized data with confidence markers (such as standardized displacement / rotation data) output from the multi-dimensional data acquisition step, a multi-dimensional synchronization error model of displacement error-rotation error-coupling error is established, and the motion parameters, load parameters and temperature parameters of each axis are input, and the preliminary synchronization error value is output.

[0117] S23, Error Decomposition: Based on the preliminary synchronization error value output from the multi-axis synchronization error modeling step, decompose it into systematic errors (such as fixed errors caused by machining deviations), random errors (such as temporary errors caused by load fluctuations), and coupling errors (such as interactive errors caused by multi-axis linkage), and output the decomposed error component data.

[0118] S3, Long-term performance degradation prediction and pre-compensation:

[0119] S31, Historical Data Storage and Feature Extraction: Based on the historical standardized data (including motion parameters, load, and temperature of each axis) from the multi-dimensional data acquisition steps and the historical error component data from the error analysis and modeling steps, the data is first stored according to the daily / weekly / monthly dimensions. Then, through feature engineering, the core features related to attenuation are extracted, including: the cumulative change of error amplitude of each axis, the error growth rate, the difference in error deviation under different loads, the correlation coefficient between motor temperature and error, etc., and a structured attenuation feature dataset is output.

[0120] S32, Attenuation Trend Prediction: Based on the attenuation feature dataset output from the historical data storage and feature extraction steps, an LSTM (Long Short-Term Memory)-based prediction model is adopted. Combined with the platform's cumulative runtime and the design life parameters of key components (such as lead screws and joint bearings), a performance attenuation prediction model for each axis is established to output the attenuation error prediction value for each axis within a future preset period (such as the next 7 days or 30 days). At the same time, the prediction confidence is marked (based on the reverse correction of historical prediction accuracy).

[0121] S33, Pre-compensation parameter generation: Based on the attenuation error prediction value and prediction confidence level output from the attenuation trend prediction step, and combined with the preset attenuation threshold (configurable for different application scenarios), pre-compensation parameters (including compensation axis number, compensation amount, implementation cycle, and confidence level) are generated. If the prediction error does not exceed the threshold, "progressive pre-compensation parameters" are generated (compensation is applied step by step according to the time gradient). If the prediction error exceeds the threshold, "emergency pre-compensation parameters" are generated and an early warning signal is triggered (prompting maintenance personnel to check the components).

[0122] S34, Pre-compensation and real-time error fusion: Based on the pre-compensation parameters output by the pre-compensation parameter generation step and the current error component data output by the error analysis and modeling step, a weighted fusion algorithm (the pre-compensation weight is dynamically adjusted according to the prediction confidence, and the higher the confidence, the greater the weight) is adopted to fuse the pre-compensation parameters and the real-time error component data, and output the fused error data to the dynamic compensation strategy generation step to achieve the synergy between pre-compensation and real-time compensation.

[0123] S4, Dynamic compensation strategy generation:

[0124] S41, Compensation Priority Ranking: Based on the fused error data output from the long-term performance degradation prediction and pre-compensation steps, and combined with the application scenario weights of the six-degree-of-freedom platform (e.g., in precision machining scenarios, the Z-axis displacement error has a higher priority than other axes), the error components are prioritized and an error compensation priority list is output.

[0125] S42, Adaptive Compensation Algorithm: Based on the error compensation priority list output by the compensation priority sorting step, combined with the standardized data with confidence labels output by the multi-dimensional data acquisition step (such as real-time parameters such as load and temperature), adaptively selects a matching compensation algorithm from the preset algorithm library (including PID compensation algorithm, fuzzy PID compensation algorithm, and neural network compensation algorithm) (such as selecting the neural network compensation algorithm under high load conditions), and outputs the compensation algorithm parameters adapted to the current working condition.

[0126] S43, Compensation command generation: Based on the error data fused from the compensation algorithm parameters output by the adaptive compensation algorithm steps and the long-term performance degradation prediction and pre-compensation steps, specific compensation commands (including compensation displacement, compensation speed and compensation timing) for each axis are generated according to the compensation priority list, and a set of compensation commands with timing marks is output.

[0127] S44, Feasibility Verification of Compensation Commands: Based on the standardized data with credibility markers output from the multi-dimensional data acquisition steps (such as real-time motion status data of each axis (displacement, velocity)) and preset platform physical constraint parameters (such as displacement limits, maximum acceleration, and joint rotation range of each axis), the compensation command set output from the compensation command generation step is double-verified. On the one hand, it verifies whether the single-axis compensation command exceeds the physical constraints (such as whether the Z-axis compensation displacement exceeds the platform's lifting limit), and on the other hand, it verifies the synergy of multi-axis synchronous compensation commands (such as whether the combination of X-axis and Y-axis compensation velocities causes the platform's center of gravity to deviate beyond the limit). If the verification passes, a qualified compensation command set is directly output to the real-time compensation execution step. If the verification fails, unqualified commands are marked and fed back to the adaptive compensation algorithm step, while constraint conflict prompts are output. The adaptive compensation algorithm step then re-optimizes the compensation algorithm parameters based on the constraint conditions to ensure that the generated compensation commands are both accurate and safe.

[0128] S5, Real-time Compensation Execution:

[0129] S51, Compensation Instruction Parsing: Based on the dynamic compensation strategy, the qualified compensation instruction set output by the generation step is parsed, the axis number, compensation parameters and execution timing corresponding to each compensation instruction are parsed, and the parsing results are output.

[0130] S52, Drive Control: Based on the analysis results output from the compensation instruction analysis step, the drive motors of each axis of the six-degree-of-freedom platform are precisely controlled, and compensation actions are executed in sequence (such as controlling the X-axis motor to rotate an additional preset angle to offset synchronization errors). At the same time, the operating parameters of the drive motors (such as speed and current) are collected in real time and output.

[0131] S53, Compensation Execution Feedback: Based on the operating parameters of the drive motor output by the drive control step, combined with the standardized data with confidence markers output by the multi-dimensional data acquisition step (such as real-time displacement / rotation angle data), calculate the actual error correction amount of the compensation action, and output the compensation execution result (including statuses such as "compensation in place" and "compensation deviation").

[0132] S6, Closed-loop optimization and fault diagnosis:

[0133] S61, Compensation effect evaluation: Based on the compensation execution results output by the real-time compensation execution steps, calculate the multi-axis synchronization error value after compensation, compare it with the preset accuracy threshold, and output three categories of evaluation compensation effect: qualified, needs optimization, and unqualified.

[0134] S62, System Parameter Optimization: When the evaluation result of the compensation effect evaluation step is that it needs optimization or is unqualified, the parameters of the multi-dimensional synchronous error model in the error analysis and modeling step and the parameters of the compensation algorithm in the dynamic compensation strategy generation step are adjusted in reverse based on the compensation execution result (such as adjusting the weight values ​​in the coupling coefficient matrix).

[0135] S63, Fault Diagnosis: Based on the standardized data with confidence labels output from the multi-dimensional data acquisition step and the compensation execution results output from the real-time compensation execution step, when multiple consecutive sets of data are detected as low confidence or the compensation execution results are continuously unqualified, the fault location (such as sensor fault, drive motor fault) is located through the fault feature matching algorithm, and a fault diagnosis report is output.

[0136] S64, Fault Tolerance: Based on the fault diagnosis report output by the fault diagnosis steps, take corresponding fault tolerance measures (such as calling redundant sensor data to replace the faulty sensor when the sensor fails; temporarily adjusting the compensation strategy and reducing the motion weight of the faulty axis when the drive motor fails) to ensure the continuous and stable operation of the system.

[0137] Although the present invention has been described above with reference to embodiments, various modifications can be made and components can be replaced with equivalents without departing from the scope of the invention. In particular, as long as there is no structural conflict, the features in the disclosed embodiments can be combined with each other in any manner. The lack of an exhaustive description of these combinations in this specification is merely for the sake of brevity and resource conservation. Therefore, the present invention is not limited to the specific embodiments disclosed herein, but includes all technical solutions falling within the scope of the claims.

Claims

1. A multi-axis synchronization error compensation system for a six-degree-of-freedom platform, characterized in that, include: The multi-dimensional data acquisition module comprehensively collects, preprocesses, and marks the motion parameters, environmental parameters, and load parameters of each axis of the six-degree-of-freedom platform, and outputs standardized data with credibility marking. The error analysis and modeling module, based on the standardized data with confidence labels output by the multi-dimensional data acquisition module, establishes a multi-dimensional synchronization error model to accurately calculate the synchronization error and output error component data. The long-term performance degradation prediction and pre-compensation module, based on the historical standardized data of the multi-dimensional data acquisition module and the historical error component data of the error analysis and modeling module, performs historical data mining and trend prediction to generate pre-compensation parameters. It then uses a weighted fusion algorithm to combine the parameters with the real-time error component data to output fused error data. The dynamic compensation strategy generation module, based on the fused error data output from the long-term performance degradation prediction and pre-compensation modules, first uses a compensation priority sorting unit to prioritize each error component by combining the application scenario weights of the six-degree-of-freedom platform, outputting an error compensation priority list. Then, the adaptive compensation algorithm unit, based on the error compensation priority list and standardized data with confidence markers, adaptively selects a matching compensation algorithm from a preset algorithm library and outputs compensation algorithm parameters adapted to the current working condition. Subsequently, the compensation instruction generation unit, based on the compensation algorithm parameters and the fused error data, generates specific compensation instructions for each axis according to the error compensation priority list and outputs a compensation instruction set with timing markers. Finally, the compensation instruction feasibility verification unit performs dual verification on the compensation instruction set based on standardized data with confidence markers and preset platform physical constraint parameters, outputting a qualified compensation instruction set. The real-time compensation execution module accurately executes the compensation instructions based on the qualified compensation instruction set output by the dynamic compensation strategy generation module, and provides real-time feedback and outputs the compensation execution results. The closed-loop optimization and fault diagnosis module optimizes system parameters based on the compensation execution results output by the real-time compensation execution module, while simultaneously achieving accurate fault diagnosis and fault-tolerant handling.

2. The multi-axis synchronization error compensation system for a six-degree-of-freedom platform according to claim 1, characterized in that, The multi-dimensional data acquisition module includes: The multi-source sensor unit uses a laser displacement sensor, angle encoder, force sensor, and temperature sensor to form a multi-source sensor array, which collects real-time displacement and rotation data, real-time load data, and ambient temperature data of each axis, and outputs raw data. The data preprocessing unit, based on the raw data output by the multi-source sensor group unit, removes noise interference from the raw data using the Kalman filter algorithm, and standardizes the output data of different types of sensors to output standardized data. The data credibility labeling unit calculates the data fluctuation coefficient based on the standardized data output by the data preprocessing unit, and classifies each group of data into high credibility and low credibility levels based on preset thresholds, and outputs standardized data with credibility labels.

3. The multi-axis synchronization error compensation system for a six-degree-of-freedom platform according to claim 1, characterized in that, The error analysis and modeling module includes: The multi-axis coupling coefficient calculation unit is based on standardized data with confidence labels output by the multi-dimensional data acquisition module. It selects high-confidence data as the basis for calculation and calculates the coupling coefficient between each axis under different motion conditions through the finite element analysis method, and outputs a dynamic coupling coefficient matrix. The multi-axis synchronization error modeling unit, based on the dynamic coupling coefficient matrix output by the multi-axis coupling coefficient calculation unit and combined with the standardized data with confidence markers output by the multi-dimensional data acquisition module, establishes a multi-dimensional synchronization error model of displacement error, rotation angle error and coupling error, and inputs motion parameters, load parameters and temperature parameters of each axis, and outputs preliminary synchronization error values. The error decomposition unit decomposes the preliminary synchronization error value output by the multi-axis synchronization error modeling unit into systematic error, random error and coupling error, and outputs the decomposed error component data.

4. The multi-axis synchronization error compensation system for a six-degree-of-freedom platform according to claim 1, characterized in that, The real-time compensation execution module includes: The compensation instruction parsing unit, based on the qualified compensation instruction set output by the dynamic compensation strategy generation module, parses the axis number, compensation parameters and execution timing corresponding to each compensation instruction, and outputs the parsing results; The drive control unit, based on the analysis results output by the compensation command analysis unit, precisely controls the drive motors of each axis of the six-degree-of-freedom platform, executes compensation actions in sequence, and simultaneously collects and outputs the operating parameters of the drive motors in real time. The compensation execution feedback unit calculates the actual error correction amount of the compensation action based on the operating parameters of the drive motor output by the drive control unit and the standardized data with confidence markers output by the multi-dimensional data acquisition module, and outputs the compensation execution result.

5. The multi-axis synchronization error compensation system for a six-degree-of-freedom platform according to claim 1, characterized in that, The closed-loop optimization and fault diagnosis module includes: The compensation effect evaluation unit calculates the multi-axis synchronization error value after compensation based on the compensation execution result output by the real-time compensation execution module, compares it with the preset accuracy threshold, and outputs three categories of evaluation compensation effect: qualified, needs optimization, and unqualified. When the evaluation result of the compensation effect evaluation unit is that optimization is needed or unqualified, the system parameter optimization unit adjusts the multi-dimensional synchronous error model parameters of the error analysis and modeling module and the compensation algorithm parameters of the dynamic compensation strategy generation module in reverse based on the compensation execution result. The fault diagnosis unit, based on the standardized data with confidence labels output by the multi-dimensional data acquisition module and the compensation execution results output by the real-time compensation execution module, will locate the fault location and output a fault diagnosis report when multiple consecutive sets of data are low confidence or the compensation execution results are consistently unqualified, by using a fault feature matching algorithm. The fault-tolerant processing unit takes corresponding fault-tolerant measures based on the fault diagnosis report output by the fault diagnosis unit.

6. The multi-axis synchronization error compensation system for a six-degree-of-freedom platform according to claim 1, characterized in that, The long-term performance degradation prediction and pre-compensation module includes: The historical data storage and feature extraction unit, based on the historical standardized data from the multi-dimensional data acquisition module and the historical error component data from the error analysis and modeling module, first classifies and stores them, then extracts attenuation features through feature engineering, and outputs a structured attenuation feature dataset. The attenuation trend prediction unit, based on the attenuation feature dataset output by the historical data storage and feature extraction unit, adopts an LSTM-based prediction model and combines the platform's cumulative runtime and the design life parameters of key components to establish a performance attenuation prediction model for each axis, so as to output the attenuation error prediction value of each axis within a future preset period, and mark the prediction confidence level. The pre-compensation parameter generation unit generates pre-compensation parameters based on the attenuation error prediction value and prediction confidence output by the attenuation trend prediction unit, combined with a preset attenuation threshold. The pre-compensation and real-time error fusion unit, based on the pre-compensation parameters output by the pre-compensation parameter generation unit and the current error component data output by the error analysis and modeling module, uses a weighted fusion algorithm to fuse the pre-compensation parameters and the real-time error component data, and outputs the fused error data to the dynamic compensation strategy generation module.

7. A method for multi-axis synchronization error compensation of a six-degree-of-freedom platform, characterized in that, The specific steps are as follows: S1, Multi-dimensional Data Acquisition: S11, Multi-source sensor array: It adopts a multi-source sensor array composed of laser displacement sensor, angle encoder, force sensor and temperature sensor to collect real-time displacement and rotation data, real-time load data and ambient temperature data of each axis, and output raw data. S12, Data Preprocessing: Based on the raw data output from the multi-source sensor group steps, noise interference in the raw data is removed by Kalman filtering algorithm, and the output data of different types of sensors are standardized to output standardized data. S13, Data Confidence Labeling: Based on the standardized data output from the data preprocessing step, calculate the data fluctuation coefficient, and combine it with a preset threshold to classify each group of data into high confidence and low confidence levels, and output standardized data with confidence labels. S2, Error Analysis and Modeling: S21, Multi-axis coupling coefficient calculation: Based on the standardized data with confidence labels output from the multi-dimensional data acquisition steps, high-confidence data is selected as the basis for calculation. The coupling coefficient between each axis under different motion conditions is calculated by the finite element analysis method, and the dynamic coupling coefficient matrix is ​​output. S22, Multi-axis synchronization error modeling: Based on the dynamic coupling coefficient matrix output by the multi-axis coupling coefficient calculation step, combined with the standardized data with confidence markers output by the multi-dimensional data acquisition step, a multi-dimensional synchronization error model of displacement error, rotation angle error and coupling error is established, and the motion parameters, load parameters and temperature parameters of each axis are input, and the preliminary synchronization error value is output. S23, Error Decomposition: Based on the preliminary synchronization error value output from the multi-axis synchronization error modeling step, decompose it into systematic error, random error and coupling error, and output the decomposed error component data; S3, Long-term performance degradation prediction and pre-compensation: S31, Historical Data Storage and Feature Extraction: Based on the historical standardized data from the multi-dimensional data acquisition steps and the historical error component data from the error analysis and modeling steps, the data is first classified and stored, and then the decay features are extracted through feature engineering, and a structured decay feature dataset is output. S32, Attenuation Trend Prediction: Based on the attenuation feature dataset output from the historical data storage and feature extraction steps, an LSTM-based prediction model is used, combined with the platform's cumulative runtime and the design life parameters of key components, to establish a performance attenuation prediction model for each axis, so as to output the attenuation error prediction value of each axis within the future preset period, and mark the prediction confidence level. S33, Pre-compensation parameter generation: Based on the attenuation error prediction value and prediction confidence output by the attenuation trend prediction step, and combined with the preset attenuation threshold, pre-compensation parameters are generated. S34, Pre-compensation and real-time error fusion: Based on the pre-compensation parameters output by the pre-compensation parameter generation step and the current error component data output by the error analysis and modeling step, a weighted fusion algorithm is used to fuse the pre-compensation parameters and the real-time error component data, and the fused error data is output to the dynamic compensation strategy generation step. S4, Dynamic Compensation Strategy Generation: Based on the fused error data output from the long-term performance degradation prediction and pre-compensation steps, the error components are first prioritized by a compensation priority sorting step, combined with the application scenario weights of the six-degree-of-freedom platform, and an error compensation priority list is output. Then, an adaptive compensation algorithm step adaptively selects a matching compensation algorithm from a preset algorithm library based on the error compensation priority list and standardized data with confidence markers, and outputs compensation algorithm parameters adapted to the current working condition. Subsequently, a compensation instruction generation step generates specific compensation instructions for each axis according to the error compensation priority list based on the compensation algorithm parameters and the fused error data, and outputs a compensation instruction set with time sequence markers. Finally, a compensation instruction feasibility verification step performs double verification on the compensation instruction set based on standardized data with confidence markers and preset platform physical constraint parameters, and outputs a qualified compensation instruction set. S5, Real-time Compensation Execution: S51, Compensation Instruction Parsing: Based on the dynamic compensation strategy, the qualified compensation instruction set output by the generation step is parsed, the axis number, compensation parameters and execution timing corresponding to each compensation instruction are parsed, and the parsing results are output. S52, Drive Control: Based on the analysis results output from the compensation instruction analysis steps, the drive motors of each axis of the six-degree-of-freedom platform are precisely controlled, and compensation actions are executed in sequence. At the same time, the operating parameters of the drive motors are collected in real time and output. S53, Compensation Execution Feedback: Based on the operating parameters of the drive motor output by the drive control step, combined with the standardized data with confidence markers output by the multi-dimensional data acquisition step, calculate the actual error correction amount of the compensation action, and output the compensation execution result. S6, Closed-loop optimization and fault diagnosis: S61, Compensation effect evaluation: Based on the compensation execution results output by the real-time compensation execution steps, calculate the multi-axis synchronization error value after compensation, compare it with the preset accuracy threshold, and output three categories of evaluation compensation effect: qualified, needs optimization, and unqualified. S62, System parameter optimization: When the evaluation result of the compensation effect evaluation step is that it needs to be optimized or is unqualified, the multi-dimensional synchronous error model parameters of the error analysis and modeling step and the compensation algorithm parameters of the dynamic compensation strategy generation step are adjusted in reverse based on the compensation execution result. S63, Fault Diagnosis: Based on the standardized data with confidence labels output from the multi-dimensional data acquisition steps and the compensation execution results output from the real-time compensation execution steps, when multiple consecutive sets of data are detected as low confidence or the compensation execution results are continuously unqualified, the fault location is located through the fault feature matching algorithm, and a fault diagnosis report is output. S64, Fault Tolerance: Based on the fault diagnosis report output by the fault diagnosis steps, take corresponding fault tolerance measures.

Images