Adaptive optimization methods, systems, and storage media for automotive fault detection systems

Abstract

This application relates to the field of automotive fault detection technology, and discloses an adaptive optimization method, system, and storage medium for an automotive fault detection system. The method includes: real-time detection of attitude data based on a preset tilt angle threshold and steering signal delay, generating angular velocity sensing data and vibration interference identification results; feature decomposition of the attitude data stream, identifying abnormal fluctuations and determining the presence of potential fault signals, updating the calibration frequency of the onboard sensors, and obtaining adjusted parameter configuration information; analyzing the fusion results of the attitude data, calculating an evaluation index for deviation correction accuracy, constructing a real-time feedback mechanism based on the evaluation index, and generating a fault signal correction scheme; constructing a continuous detection loop, analyzing the matching effect between the calibration frequency and the steering signal delay, and iteratively processing the attitude data if the deviation correction accuracy does not reach a first preset standard. This application improves the accuracy and response speed of adaptive optimization in automotive fault detection systems.

Description

Technical Field

[0001] This application relates to the field of automotive fault detection technology, and in particular to an adaptive optimization method, system and storage medium for an automotive fault detection system. Background Technology

[0002] With the rapid development of intelligent vehicles and autonomous driving technologies, vehicle safety and stability have become critical issues. Traditional vehicle fault detection methods largely rely on static sensor calibration and periodic fault checks, which cannot respond in real-time to dynamically changing road environments and vehicle conditions. Therefore, how to detect and adaptively adjust sensor configurations in real-time to cope with different driving conditions has become an important topic for improving the accuracy and reliability of vehicle fault detection.

[0003] Existing fault detection systems typically rely on onboard sensors such as accelerometers and gyroscopes to monitor vehicle attitude data, such as pitch and roll angles, and analyze this data to identify potential fault signals. However, these systems have several drawbacks, such as insufficient accuracy in identifying vibration interference or the inability to calibrate sensors in real time under complex road conditions, leading to misjudgments or missed detections of fault signals. Furthermore, due to the complexity of road conditions and driving behavior, existing methods struggle to dynamically adjust sensor operating parameters and response times, thus failing to adapt to changes in different driving environments.

[0004] To address these issues, this invention proposes an adaptive optimization method. This method involves real-time acquisition of vehicle attitude data, analysis of potential fault signals, and dynamic adjustment of sensor parameters and calibration frequency based on the analysis results, thereby improving the accuracy of fault signal correction. This method effectively copes with complex factors such as external interference and signal delays, enhancing the accuracy of fault detection and the system's adaptive capabilities, thus improving vehicle driving safety. Summary of the Invention

[0005] To address the aforementioned technical problems, this application provides an adaptive optimization method, system, and storage medium for an automotive fault detection system, which improves the accuracy and response speed of the adaptive optimization of the automotive fault detection system.

[0006] Firstly, this application provides an adaptive optimization method for an automotive fault detection system, the adaptive optimization method for the automotive fault detection system comprising:

[0007] The vehicle uses onboard sensors to collect attitude data in real time during vehicle operation. Based on a preset tilt angle threshold and steering signal delay, the attitude data is detected in real time to generate angular velocity sensing data and vibration interference identification results.

[0008] The attitude data is preprocessed and calibrated to generate an attitude data stream. Feature decomposition is performed on the attitude data stream to identify abnormal fluctuations and determine whether there are potential fault signals. Based on the potential fault signals, the parameter configuration of the vehicle sensor is adjusted and the calibration frequency of the vehicle sensor is updated to obtain the adjusted parameter configuration information.

[0009] Based on the adjusted parameter configuration information, the fusion results of attitude data are analyzed, the evaluation index of deviation correction accuracy is calculated, a real-time feedback mechanism is constructed based on the evaluation index, the response time is dynamically adjusted, and a fault signal correction scheme is generated.

[0010] Based on the fault signal correction scheme, a continuous detection loop is constructed to analyze the matching effect between the calibration frequency and the steering signal delay. If the deviation correction accuracy does not reach the first preset standard, the attitude data is iteratively processed.

[0011] In conjunction with the first aspect, the generation of angular velocity sensing data and vibration disturbance identification results includes:

[0012] Based on the attitude data, the difference between the actual tilt angle of the monitored vehicle and the preset tilt angle threshold is obtained. Based on the steering signal delay, the delay data between the trigger time of the steering signal and the actual steering action is recorded. Based on the difference in the actual tilt angle and the delay data, the initial angular velocity sensing data is generated.

[0013] The attitude data is preliminarily analyzed to obtain vibration data. The angular velocity sensing data is correlated and compared with the vibration data to generate the vibration interference identification result and determine the signal offset range.

[0014] Based on the signal offset range, abnormal data points in the attitude data are filtered out.

[0015] In conjunction with the first aspect, the generation of attitude data stream includes:

[0016] Based on the signal offset range, the attitude data is denoised using signal noise filtering technology to generate preprocessed attitude data.

[0017] The sampling direction of the preprocessed attitude data is adjusted based on the sensor orientation correction logic to generate the orientation deviation of the vehicle sensor under different driving conditions. The preprocessed attitude data is then calibrated based on the orientation deviation to generate the attitude data stream.

[0018] Time series analysis and interpolation are performed on the attitude data stream to optimize its integrity.

[0019] In conjunction with the first aspect, the identification of abnormal fluctuations and determination of whether there are potential fault signals includes:

[0020] Based on feature decomposition, feature parameters related to vehicle attitude change in the attitude data stream are obtained, and the path deviation trend of vehicle attitude change is obtained based on the feature parameters.

[0021] Vibration interference feature values ​​are extracted from the vibration data in the attitude data stream. Based on the vibration interference feature values, it is determined whether there are abnormal fluctuations in the attitude data stream. If the abnormal fluctuations exceed a preset tilt angle threshold, it is determined that there is a potential fault signal.

[0022] Based on the judgment results, the time points and amplitudes of abnormal fluctuations are recorded, an abnormal fluctuation distribution map is generated, and the periodic characteristics of the abnormal fluctuation distribution map are obtained.

[0023] In conjunction with the first aspect, obtaining the adjusted parameter configuration information includes:

[0024] Based on the priority of the potential fault signals, the safety impact level of the monitored vehicle is generated. Based on the safety impact level, the operating parameters, response time characteristics and calibration frequency of the on-board sensors are adjusted to generate the adjusted parameter configuration information.

[0025] Verify the applicability of the parameter configuration information under different driving scenarios, and record the effect of parameter configuration adjustments.

[0026] In conjunction with the first aspect, the evaluation index for the accuracy of the calculation deviation correction includes:

[0027] Based on the adjusted parameter configuration information, the real-time acquired attitude data is recalibrated to generate new data fusion results;

[0028] Based on the new data fusion results, the deviation data of the monitored vehicle on the current driving path is obtained. The deviation data is matched with the historical driving path deviation data, the deviation value between the current attitude data and the historical attitude data is calculated, and an evaluation index of deviation correction accuracy is generated based on the deviation value.

[0029] In conjunction with the first aspect, the fault signal correction scheme includes:

[0030] The evaluation indicators are analyzed from multiple dimensions, various driving conditions are classified, statistical methods are used to calculate the coefficient of variation of the indicators under each driving condition, a dataset is generated, and the calculation method of the evaluation indicators is optimized based on the dataset.

[0031] A real-time feedback model is constructed based on the optimized evaluation index, and the changes in angular velocity sensing data are monitored in real time. The response time of the real-time feedback model is dynamically adjusted, and the feedback results for abnormal data are output to generate the fault signal correction scheme.

[0032] The fault signal correction scheme was initially verified, and its application effect in actual driving was recorded. Based on the application effect, the stability of the fault signal correction scheme was analyzed.

[0033] In conjunction with the first aspect, the construction of a continuous detection loop based on the fault signal correction scheme includes:

[0034] Based on the fault signal correction scheme, trigger conditions are set, attitude data is collected in real time based on the trigger conditions, the ratio between calibration frequency and steering signal delay is calculated based on the attitude data, and the ratio is set as the matching effect.

[0035] If the matching effect does not meet the second preset standard, the deviation data in the current detection loop is recorded, deviation correction suggestions are generated, the parameter settings of the detection loop are adjusted, and the operation logic of the detection loop is optimized based on the deviation correction accuracy before and after the parameter adjustment.

[0036] Obtain the evaluation index of the current deviation correction accuracy. If the evaluation index does not meet the first preset standard, backtrack to the signal and noise filtering stage, reprocess the original attitude data for noise filtering, and generate updated attitude data.

[0037] The calibration process is re-executed on the updated attitude data to generate a new attitude data stream. Anomaly fluctuation analysis is performed on the new attitude data stream to update the determination results of potential fault signals.

[0038] Secondly, this application provides an automotive fault detection system, the automotive fault detection system comprising:

[0039] The acquisition module uses onboard sensors to collect attitude data of the vehicle in real time during driving. Based on a preset tilt angle threshold and steering signal delay, the attitude data is detected in real time to generate angular velocity sensing data and vibration interference identification results.

[0040] The identification module is used to preprocess and calibrate the attitude data, generate an attitude data stream, perform feature decomposition on the attitude data stream, identify abnormal fluctuations and determine whether there are potential fault signals, adjust the parameter configuration of the vehicle sensor based on the potential fault signals, update the calibration frequency of the vehicle sensor, and obtain the adjusted parameter configuration information.

[0041] The feedback module is used to analyze the fusion results of attitude data based on the adjusted parameter configuration information, calculate the evaluation index of deviation correction accuracy, construct a real-time feedback mechanism based on the evaluation index, dynamically adjust the response time, and generate a fault signal correction scheme.

[0042] The optimization module is used to construct a continuous detection loop based on the fault signal correction scheme, analyze the matching effect between the calibration frequency and the steering signal delay, and iteratively process the attitude data if the deviation correction accuracy does not reach the first preset standard.

[0043] A third aspect of this application provides a computer-readable storage medium having a computer program stored thereon, characterized in that, when the computer program is run by a processor, it causes the processor to execute the adaptive optimization method for an automotive fault detection system as described in any one of the preceding descriptions.

[0044] The technical solution provided in this application acquires vehicle sensor attitude data in real time and detects the data by combining tilt angle thresholds and steering signal delays, enabling timely identification of potential fault signals during vehicle operation. A Kalman filter is used to fuse tilt angle differences and steering signal delay data to generate smooth angular velocity sensing data, thereby improving sensitivity to vehicle attitude changes in dynamic environments and effectively enhancing fault detection accuracy and real-time response capabilities. By correlating and comparing angular velocity sensing data with vibration data and calculating the linear correlation between the two based on the Pearson correlation coefficient, vibration interference caused by non-fault factors, such as uneven road surfaces and engine vibration, can be accurately identified. By analyzing the priority of potential fault signals and assessing their impact on vehicle safety, the operating parameters and calibration frequency of the sensors can be dynamically adjusted according to different driving conditions. Based on historical and real-time acquired data, methods such as autoregressive integral moving average models and wavelet transforms are used to dynamically calibrate, denoise, and extract features from the attitude data, enabling adaptive adjustment of the signal offset range, accurate identification of abnormal data points, and improved data quality and reliability through iterative processing. A feedback mechanism based on evaluation metrics is constructed by real-time evaluation of deviation correction accuracy and dynamic adjustment of response time. When the system detects that the deviation correction accuracy does not meet the preset standard, it can automatically backtrack to the noise filtering stage, reprocess the data, and perform calibration. By combining steering signal delay, vibration interference identification, and dynamic adjustment of signal offset range, the system can more accurately capture changes in vehicle attitude and dynamic anomalies, avoiding misjudgments caused by slow sensor response or inaccurate data. Attached Figure Description

[0045] To more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings used in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0046] Figure 1This is a schematic diagram of an embodiment of the adaptive optimization method for an automotive fault detection system in this application.

[0047] Figure 2 This is a schematic diagram of an embodiment of converting the Fourier transform of fault data in the attitude data stream into frequency domain features in this application.

[0048] Figure 3 This is a schematic diagram of an embodiment of abnormal fluctuation detection after threshold comparison in this application.

[0049] Figure 4 This is a schematic diagram of an embodiment of the adaptive optimization system of the vehicle fault detection system in this application. Detailed Implementation

[0050] This application provides an adaptive optimization method, system, and storage medium for an automotive fault detection system. The terms "first," "second," "third," "fourth," etc. (if present) in the specification, claims, and accompanying drawings are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments described herein can be implemented in a sequence other than that illustrated or described herein. Furthermore, the terms "comprising" or "having" and any variations thereof are intended to cover a non-exclusive inclusion; for example, a process, method, system, product, or device that includes a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or devices.

[0051] For ease of understanding, the specific process of the embodiments of this application is described below. Please refer to [link / reference]. Figure 1 One embodiment of the adaptive optimization method for the vehicle fault detection system in this application includes:

[0052] Step S101: Use on-board sensors to collect attitude data of the vehicle in real time during driving. Based on the preset tilt angle threshold and steering signal delay, the attitude data is detected in real time to generate angular velocity sensing data and vibration interference identification results.

[0053] The generation of angular velocity sensing data and vibration disturbance identification results includes:

[0054] Based on attitude data, the difference between the actual tilt angle of the monitored vehicle and the preset tilt angle threshold is obtained. Based on the steering signal delay, the delay data between the trigger time of the steering signal and the actual steering action is recorded. Based on the difference in the actual tilt angle and the delay data, initial angular velocity sensing data is generated. The attitude data is preliminarily analyzed to obtain vibration data. The angular velocity sensing data is correlated and compared with the vibration data to generate vibration interference identification results and determine the signal offset range. Based on the signal offset range, abnormal data points in the attitude data are screened out.

[0055] It is understood that the executing entity of this application can be an adaptive optimization device of an automotive fault detection system, or it can be a terminal or a server; no specific limitation is made here. This application's embodiment uses a server as the executing entity for illustration.

[0056] Specifically, attitude data refers to the set of data collected in real time by various onboard sensors such as gyroscopes, accelerometers, magnetometers, and inertial measurement units (IMUs) to describe the vehicle's spatial motion state. This data includes, but is not limited to, parameters such as pitch angle, roll angle, yaw angle, angular velocity, and linear acceleration. Attitude data can be used to determine whether the vehicle is experiencing abnormal motion states such as tilting, rolling, sharp turns, or bumps. Angular velocity sensing data refers to data extracted from the raw attitude data, reflecting the characteristics of the vehicle's angular velocity changes around each axis. Angular velocity sensing data not only includes the original angular velocity measurements but also incorporates information on the deviation between the vehicle's actual tilt angle and a first preset threshold, as well as the impact of steering signal delay on dynamic response. It is crucial intermediate data for subsequent fault diagnosis and interference identification. Vibration interference identification results refer to the interference signals caused by non-fault factors, such as uneven road surfaces, engine vibration, and tire imbalance, identified by the automotive fault detection system through analysis of high-frequency vibration components in the attitude data and correlation comparison with angular velocity sensing data. These interference signals can be used to distinguish between real fault signals and environmental noise, improving detection accuracy. Tilt angle threshold refers to a pre-set critical angle value used to determine whether a vehicle is in an abnormal tilt state. The tilt angle threshold can be dynamically adjusted according to vehicle model, load, and driving conditions, and can be a fixed value or an adaptive threshold dynamically optimized based on a machine learning model. Steering signal delay refers to the time delay between the driver turning the steering wheel and the vehicle actually producing a measurable steering action, such as changes in yaw rate or front wheel angle. Steering signal delay is affected by factors such as the mechanical characteristics of the steering system, the response speed of the electronic power steering system, vehicle speed, and tire grip. Signal offset range refers to the statistically permissible fluctuation range of attitude data or angular velocity sensing data under normal driving conditions. Data points exceeding this range are considered abnormal data points, potentially corresponding to potential faults or strong interference events. Abnormal data points are sampling points in the attitude data sequence that significantly deviate from the normal pattern or exceed the signal offset range, possibly caused by sensor malfunction, vehicle component abnormalities, external impacts, or strong electromagnetic interference.

[0057] The system analyzes real-time attitude data to extract the vehicle's actual tilt angles, such as roll and pitch angles, at the current moment. These angles are compared to a preset tilt angle threshold, and the angle difference is calculated. The tilt angle threshold serves as the reference attitude for the vehicle when stationary or traveling at a constant speed on a level surface and can be dynamically calibrated using a self-learning algorithm. Simultaneously, the system monitors the time difference between the steering signal input (e.g., steering wheel angle sensor signal) and the vehicle's actual yaw rate response, denoted as the steering signal delay. The raw angular velocity signals, such as roll and pitch angular velocities, are weighted and corrected so that the angular velocity sensing data not only reflects the physical rotation rate but also includes predictive information about potential loss of control risks, improving the system's sensitivity to slowly accumulating faults.

[0058] High-frequency components in the attitude data are bandpass filtered to extract vibration data. Wavelet transform or short-time Fourier transform (STFT) can be used for time-frequency analysis to identify the main vibration frequency components and further calculate the vibration intensity index. If the vibration intensity index exceeds a preset vibration threshold, a strong vibration event is preliminarily determined. Angular velocity sensing data and vibration data are time-aligned and correlated, and the Pearson correlation coefficient is calculated. If the correlation coefficient is less than a second preset threshold, it indicates that the angular velocity change and vibration are not significantly correlated. In this case, the angular velocity change is more likely caused by abnormal vehicle attitude rather than road surface excitation. Conversely, if the correlation coefficient is greater, it is determined that the vibration significantly interferes with the angular velocity signal, thus generating a vibration interference identification result. If it is determined to be "high interference," the data for the current period is marked as "disturbed," reducing its weight in fault diagnosis; if it is "low interference," the data is allowed to participate in subsequent analysis normally.

[0059] Simultaneously, based on historical normal driving data, such as the past 10 minutes when the vehicle had no steering or bumpy road sections, a dynamic baseline model for each attitude parameter is established. The signal offset range is determined using the 3σ principle or based on kernel density estimation (KDE). For example, the offset range = [μ−3σ, μ+3σ], where μ and σ are the mean and standard deviation of the data within the sliding time window, respectively. This range is adaptively updated according to driving conditions.

[0060] The current attitude data is compared point by point with the aforementioned signal offset range. If any parameter, such as roll angle, angular velocity, or vibration intensity, exceeds its corresponding range, the sampling point is marked as an abnormal data point. Furthermore, consecutive abnormal points can be clustered. If the anomaly continues to exceed a third preset threshold, a fault warning process is triggered.

[0061] Step S102: Preprocess and calibrate the attitude data to generate an attitude data stream. Perform feature decomposition on the attitude data stream to identify abnormal fluctuations and determine whether there are potential fault signals. Adjust the parameter configuration of the vehicle sensor based on the potential fault signals and update the calibration frequency of the vehicle sensor to obtain the adjusted parameter configuration information.

[0062] The generation of attitude data stream includes:

[0063] Based on the signal offset range, signal noise filtering technology is used to denoise the attitude data and generate preprocessed attitude data. Based on the sensor orientation correction logic, the sampling orientation of the preprocessed attitude data is adjusted to generate the orientation deviation of the vehicle sensor under different driving conditions. Based on the orientation deviation, the preprocessed attitude data is calibrated to generate an attitude data stream. Time series analysis and interpolation processing are performed on the attitude data stream to optimize its integrity.

[0064] Specifically, the attitude data stream refers to highly complete attitude time-series data after time alignment, denoising, orientation correction, and interpolation completion. Noise filtering is applied to the attitude data through a Kalman filter, which is a recursive algorithm used to estimate the system state from measurements containing uncertainties. First, the state vector, including position and velocity, is initialized. Then, the state estimate is updated using a motion model during the prediction phase. In the update phase, new measurement data is incorporated to correct the estimate, thereby reducing the influence of random noise within the signal offset range.

[0065] The sensor orientation correction logic refers to an online algorithm that rotates the IMU physical coordinate system to the vehicle coordinate system. It adjusts the sampling orientation of the preprocessed attitude data, specifically including: calculating the current sensor attitude matrix based on fused data from the gyroscope and accelerometer. This attitude matrix is ​​a three-dimensional rotation matrix represented by Euler angles. The calculation process involves using accelerometer data for gravity orientation correction and integrating the rotation increment using the gyroscope data to update the matrix elements; adjusting the sampling orientation based on the deviation between the attitude matrix and a preset reference orientation using a quaternion rotation algorithm. Quaternions are a rotation representation method that avoids gimbal lock problems, rotating the current orientation vector to the calibration position through multiplication; verifying the consistency of the adjusted orientation, iteratively correcting if the deviation exceeds a threshold to ensure accuracy under high-speed driving or rough road conditions. For example, in urban road driving scenarios, when a vehicle encounters a slope, the sensor may experience orientation deviation due to tilt. Adjusting the sampling orientation using this logic improves the reliability of the attitude data, reduces the risk of subsequent misjudgments, and thus enhances the robustness of the overall monitoring system.

[0066] The system identifies driving conditions, such as straight roads, curves, or slopes, and classifies them using angular velocity data. An angular velocity exceeding 5 degrees per second is considered a curve condition. For curve conditions, a deviation vector is calculated—the vector difference between the current direction and the ideal direction—and corrected using a proportional-integral-derivative (PID) controller. The PID controller is a feedback mechanism: the proportional term responds to the current error, the integral term accumulates past errors, and the derivative term predicts future errors. A weighted summation is used to output the correction. For slope conditions, a tilt angle threshold is incorporated; if it exceeds 10 degrees, additional gravity compensation is applied. The calculation process decomposes acceleration into gravity and motion components, subtracting gravity to correct the deviation. The correction results from multiple conditions are fused to generate a unified deviation correction value for subsequent data stream generation. For example, in mountainous slope driving, sensor orientation deviations can distort attitude data. This correction accurately captures vibration interference, optimizes fault signal detection accuracy, and helps prevent potential vehicle malfunctions.

[0067] Time series analysis employs an autoregressive integral moving average model, a statistical method for modeling time-dependent data. This model trains parameters by minimizing the sum of squared errors, capturing trends and random fluctuations in historical data to predict and interpolate future missing data. First, autoregressive parameters are estimated to capture the influence of past values, moving average parameters handle random shocks, and the integral order addresses non-stationarity. Missing points are predicted based on the model; if more than three consecutive sampling points are missing, they are marked as potential missing points. Linear interpolation is then applied to fill these missing points. Linear interpolation calculates the slope of a straight line from adjacent known points and then interpolates within the missing locations, thus restoring data continuity. For example, in high-vibration off-road driving, data may be lost due to interference; interpolation maintains the integrity of the data stream, preventing information loss during feature extraction and enhancing the accuracy of fault diagnosis.

[0068] The process of identifying abnormal fluctuations and determining whether there are potential fault signals includes:

[0069] Feature parameters related to vehicle attitude changes in the attitude data stream are obtained based on feature decomposition. The path deviation trend of vehicle attitude changes is obtained based on the feature parameters. Vibration interference feature values ​​are extracted from the vibration data in the attitude data stream. The presence of abnormal fluctuations in the attitude data stream is determined based on the vibration interference feature values. If the abnormal fluctuations exceed a preset tilt angle threshold, a potential fault signal is determined. The time point and amplitude of the abnormal fluctuations are recorded based on the determination results, an abnormal fluctuation distribution map is generated, and the periodicity characteristics of the abnormal fluctuation distribution map are obtained.

[0070] Specifically, based on the angular velocity sensing data and steering signal delay in the attitude data stream, the data sequence is decomposed using the Fourier transform method. The Fourier transform is the process of converting a time-domain signal into a frequency-domain signal. By calculating the frequency components of the signal, low-frequency features related to path deviation are separated. For example... Figure 2This diagram illustrates the Fourier transform of fault data in the attitude data stream into frequency domain features. The low-frequency signal (0-5Hz) mainly reflects the slow attitude changes of the vehicle caused by steering and driving on slopes (such as the gradual change of roll angle when driving on a curve). These features are directly related to the deviation of the vehicle's driving trajectory. The high-frequency signal (15-25Hz) mainly originates from non-fault factors such as uneven road surface, engine vibration, and tire dynamic imbalance (such as short-term bumps when driving over speed bumps). These features need to be distinguished from real fault signals. Continuous attitude data points are converted into a spectrogram, from which the amplitude corresponding to the dominant frequency is extracted as a feature parameter. The extracted amplitude value quantifies the degree of path deviation; for example, in urban road scenarios, if the amplitude value exceeds a preset value, it indicates a significant deviation. Based on the extracted feature parameters, a linear regression method is applied to fit the trend line of path deviation. Linear regression calculates the fitted line for the data points using the least squares method, including calculating the mean and variance of the feature parameters, and then solving for the slope and intercept to predict the future direction of the deviation. In high-speed scenarios, if the slope of the trend line is greater than a preset value, it is judged as an accelerating deviation trend, requiring further adjustment. For example, when performing feature decomposition on the attitude data stream, in an urban road scenario, assuming the attitude data stream contains 100 data points, the dominant frequency extracted after Fourier transform is 0.1Hz, with a corresponding amplitude of 0.6, serving as a path deviation feature parameter, which helps identify slow deviations. In high-speed scenarios, the dominant frequency is 0.2Hz, and the amplitude is 0.8, indicating a rapid deviation. This decomposition can accurately capture path changes at different speeds, improving the adaptability of monitoring.

[0071] Wavelet transform is applied to extract vibration data. Wavelet transform decomposes the signal into wavelet coefficients of different scales. By selecting the Daubechies wavelet basis function, the coefficients are calculated, thus obtaining the peak value and frequency of the vibration interference as characteristic values. The peak value of the characteristic value is compared with the historical average. If the peak deviation exceeds 20%, it is judged as abnormal fluctuation. If the abnormal fluctuation causes the angle to exceed the tilt angle threshold, a potential fault signal is confirmed. This method can identify vehicle stability problems early and improve the effectiveness of safety monitoring. Figure 3 This is a schematic diagram of abnormal fluctuation detection after threshold comparison. Based on the vehicle model, load, and driving conditions, a preset tilt angle threshold (such as 0.7°) is set. Fluctuations exceeding this threshold are considered "abnormal fluctuations that may be associated with faults".

[0072] A scatter plot is generated using time points as the horizontal axis and amplitude as the vertical axis. The periodicity of the distribution plot is then calculated using an autocorrelation function. The autocorrelation function identifies the period by calculating the correlation between the signal and its delayed version. For example, calculating the correlation coefficient at delay k, if the peak value appears at a fixed interval, the period is confirmed to be within that interval. In an uneven road surface scenario, a period of 2 seconds indicates repetitive vibration interference; in a mechanical fault scenario, a period of 0.5 seconds points to high-frequency periodic characteristics. This analysis can distinguish different fault types and improve diagnostic accuracy.

[0073] This includes obtaining the adjusted parameter configuration information, including:

[0074] Based on the priority of potential fault signals, the safety impact level of the monitored vehicle is generated. Based on the safety impact level, the operating parameters, response time characteristics, and calibration frequency of the on-board sensors are adjusted to generate the adjusted parameter configuration information. The applicability of the parameter configuration information under different driving scenarios is verified, and the adjustment effect of the parameter configuration is recorded.

[0075] Specifically, the system collects data on the type and intensity of potential fault signals, such as abnormal tilt angles or vibration interference peaks. An impact score is calculated based on a preset safety impact model, which maps signal intensity to a safety risk level (low, medium, high) based on historical accident data. The score is then adjusted by integrating current vehicle speed and road condition information to ensure the assessment considers the real-time environment and completes the priority evaluation of potential fault signals. In this embodiment, the priority of potential fault signals is based on a comprehensive evaluation of multiple factors, including their type, intensity, and current driving scenario. This priority is further used to generate the safety impact level. For example, in urban road driving scenarios, if a potential fault signal is a turn signal with a delay exceeding 2 seconds, the priority evaluation will classify its safety impact level as high, because delays in urban environments may increase the risk of collisions, which is beneficial for timely triggering of alarm mechanisms. For high-speed driving scenarios, the priority evaluation will additionally consider angular velocity perception data. If abnormal fluctuations exceed a preset threshold of 20%, the safety impact level is upgraded to extremely high, thereby optimizing the response speed of subsequent parameter adjustments.

[0076] The degree of safety impact is directly mapped to the sensor's sampling rate and sensitivity adjustment; for example, under high impact conditions, the sampling rate corresponding to the operating parameters is increased from 10Hz to 20Hz. The response curves of the adjusted parameters under simulated environmental changes are analyzed. A Kalman filter algorithm is applied to fuse noise data; this algorithm reduces the impact of environmental interference on attitude data by iteratively estimating state variables. Through optimization, the sensor can maintain stable output under different driving scenarios such as rain or bumpy roads, which helps reduce false alarms. A feedback loop is introduced to calculate the response time characteristics; this loop uses a proportional-integral-derivative control method to adjust the time constant; for different environments such as night or fog, the response time is dynamically shortened to within 0.5 seconds; the adjusted characteristics are monitored to ensure they match the initial signal offset range. Through dynamic adjustment, the system's response efficiency to sudden events can be improved, reducing the probability of accidents. Tilt angle threshold data is fused to further refine the response time characteristics, ensuring the stability of continuous monitoring in continuous curves.

[0077] The performance metrics of the parameter configuration information in urban and highway scenarios were tested on a simulation platform. Applicability scores were calculated, such as a deviation correction accuracy of 98% in rainy weather scenarios. Iterative verification was performed; if the score fell below a threshold, adjustments were made retrospectively. This verification ensured the robustness of the configuration across multiple scenarios, benefiting practical deployment. For example, in nighttime urban driving verification, the configuration information demonstrated high applicability, reducing vibration interference recognition latency by 15%, thus improving nighttime safety. Logs recorded the performance differences before and after adjustments, i.e., the adjustment effects, such as the reduction in response time.

[0078] Step S103: Based on the adjusted parameter configuration information, analyze the fusion results of attitude data, calculate the evaluation index of deviation correction accuracy, construct a real-time feedback mechanism based on the evaluation index, dynamically adjust the response time, and generate a fault signal correction scheme.

[0079] The evaluation metrics for calculating the accuracy of deviation correction include:

[0080] Based on the adjusted parameter configuration information, the real-time acquired attitude data is recalibrated to generate new data fusion results. Based on the new data fusion results, the deviation data of the monitored vehicle on the current driving path is obtained. The deviation data is matched with the historical driving path deviation data, and the deviation value between the current attitude data and the historical attitude data is calculated. Based on the deviation value, an evaluation index of deviation correction accuracy is generated.

[0081] Specifically, the adjusted parameter configuration information includes updated sensor calibration frequencies and signal offset ranges. This information is used to apply calibration algorithms to the raw attitude data collected from vehicle sensors, ensuring data accuracy. Based on the recalibrated data, angular velocity sensing data and vibration disturbance identification results are fused to form a comprehensive attitude data stream. The fusion result is used to calculate the difference between the vehicle's real-time tilt angle and the preset path, identifying deviations such as the degree of deviation during turns. A time-series matching algorithm is employed to compare the current fusion result with stored historical driving path deviation data. This includes: extracting key features from the fusion result, such as tilt angle thresholds and steering signal delays; calculating similarity with corresponding features in historical data; applying a dynamic time warping algorithm to match the time series (a common method for handling sequence similarity, aligning sequences of different lengths by minimizing the distance between them); recording deviation points during the matching process; and using the Euclidean distance formula to calculate the difference between deviation points in the matching result to obtain the deviation value.

[0082] The deviation value is mapped to the accuracy index to generate an evaluation index for the deviation correction accuracy. For example, the index is considered high accuracy when the deviation is less than 5 degrees.

[0083] The fault signal correction scheme includes:

[0084] The evaluation indicators are analyzed from multiple dimensions, and various driving conditions are classified. Statistical methods are used to calculate the coefficient of variation of the indicators under each driving condition, generating a dataset. Based on the dataset, the calculation method of the evaluation indicators is optimized. A real-time feedback model is built based on the optimized evaluation indicators, and the changes in angular velocity sensing data are monitored in real time. The response time of the real-time feedback model is dynamically adjusted, and feedback results for abnormal data are output, generating a fault signal correction scheme. The fault signal correction scheme is initially verified, and its application effect in actual driving is recorded. Based on the application effect, the stability of the fault signal correction scheme is analyzed.

[0085] Specifically, multi-dimensional analysis includes statistical analysis of indicators under conditions such as speed and gradient. Various driving conditions include, but are not limited to, straight roads and slopes. For example, statistical methods are used to calculate the coefficient of variation (COP) of the indicator under each driving condition. The COP is the ratio of the standard deviation to the mean. For variable-speed driving scenarios, analysis shows that a low COP indicates indicator stability. Performance data is recorded to generate a dataset. For example, in urban congestion scenarios, records show that the indicator's performance under vibration interference improves fault identification efficiency; this is performance data. Based on the performance data recorded in the dataset, the weights in the calculation formula are adjusted. This includes: identifying low-performance conditions in the performance data, such as high vibration; applying the least squares method to optimize the weights (a regression method that fits data by minimizing the sum of squared errors); and verifying the optimization effect by testing the new calculation method with simulated data.

[0086] A Kalman filter is selected as the basic framework. By inputting evaluation metrics into the filter, the state variables of the attitude data are estimated. The Kalman filter is a recursive algorithm used to estimate the state of a dynamic system from noisy measurements. First, the state covariance matrix is ​​initialized. Then, a prior estimate is calculated through a prediction step. Finally, a new measurement data is fused through an update step to correct the estimate, thereby constructing a real-time feedback model that can respond to deviations in real time. For example, by using the deviation correction accuracy as input, the process noise covariance of the filter is adjusted to ensure the model's sensitivity to changes in vehicle attitude. Based on historical data comparison methods, past driving path deviation data is incorporated into the model initialization to obtain a preliminary real-time feedback model. Using the model's prediction function, angular velocity data is sampled once per second to track its continuous trend. In the prediction function, the current state value is predicted based on the vehicle's motion model, such as predicting the vehicle's angular velocity and acceleration. Through prediction, a prior state estimate of the system is obtained, and the covariance of the prediction error is calculated.

[0087] Based on the vibration interference identification results, the fluctuation characteristics of angular velocity sensing data over different time periods are analyzed. The response time is adjusted according to these fluctuation characteristics. For example, in sharp turning scenarios, when the fluctuation characteristics are high, the response time is adjusted to 0.3 seconds, which can more quickly capture changes in angular velocity, improve the accuracy of fault signal correction, and help prevent potential accidents. Using the adjusted response time, the model's update frequency is increased, such as from once per second to twice per second, thereby improving its sensitivity to abnormal data. The optimization results are mapped to correction parameters to generate fault signal correction schemes, such as adjusting the offset range to 80% of the original value.

[0088] In simulated urban road tests, a fault signal correction scheme was applied, and the effect of reducing attitude data deviation from 1 degree to 0.2 degrees was recorded. The verification process included deploying sensors to collect real-time data, comparing path deviation before and after correction, quantifying the application effect as a 20% improvement in stability, and recording the results. The variance of the application effect data was calculated; a variance less than 0.1 was considered stable, i.e., high stability.

[0089] Step S104: Construct a continuous detection loop based on the fault signal correction scheme, analyze the matching effect between the calibration frequency and the steering signal delay, and if the deviation correction accuracy does not reach the first preset standard, iteratively process the attitude data.

[0090] The continuous detection loop, based on the fault signal correction scheme, includes:

[0091] Triggering conditions are set based on the fault signal correction scheme. Attitude data is collected in real time based on the triggering conditions. The ratio between the calibration frequency and the steering signal delay is calculated based on the attitude data, and the ratio is set as the matching effect. If the matching effect does not reach the second preset standard, the deviation data in the current detection cycle is recorded, deviation correction suggestions are generated, the parameter settings of the detection cycle are adjusted, and the operation logic of the detection cycle is optimized based on the deviation correction accuracy before and after the parameter adjustment. The evaluation index of the current deviation correction accuracy is obtained. If the evaluation index does not reach the first preset standard, the process is backtracked to the signal noise filtering stage, and the original attitude data is reprocessed for noise filtering to generate updated attitude data. The calibration process is re-executed on the updated attitude data to generate a new attitude data stream. Abnormal fluctuation analysis is performed on the new attitude data stream to update the judgment result of potential fault signals.

[0092] Specifically, based on the dynamic response time and deviation correction accuracy evaluation indicators included in the final fault signal correction scheme, triggering conditions are defined, such as a vehicle speed exceeding 30 km / h or a tilt angle change rate greater than 5 degrees / second. Vehicle sensors collect attitude data streams, including angular velocity and vibration data, when the triggering conditions are met. The matching effect between the sensor calibration frequency and the steering signal delay is analyzed using the attitude data. For example, in a scenario where the vehicle is traveling on a straight road, the attitude data shows a calibration frequency of 10 Hz and a steering signal delay of 0.1 seconds. The matching degree is calculated by dividing the calibration frequency by the steering signal delay; a higher result indicates a more ideal matching effect, which helps reduce false fault diagnoses and improve monitoring reliability.

[0093] When the matching degree is lower than the second preset standard, the deviation value, such as frequency offset and delay increment, is recorded; the recorded deviation data is classified, for example, frequency offset and delay increment are quantified separately; suggestions are generated based on the classification results, such as increasing the calibration frequency to 12Hz or shortening the delay to 0.05 seconds.

[0094] The application generates deviation correction suggestions to update parameters, such as adjusting calibration frequency values; simulates attitude data acquisition under the adjusted parameters; compares the deviation correction accuracy before and after adjustment, for example, improving accuracy from 85% to 95%; assesses the impact and generates impact analysis results, such as the improved accuracy reducing the omission of potential fault signals.

[0095] Integrate the impact analysis results, adjust the logic rules, such as adding environmental adaptation conditions; try to optimize the effect of the logic under different vibration disturbances; refine the loop iteration mechanism to ensure that the deviation correction accuracy continues to meet the standard, and complete the operation logic of the detection loop.

[0096] The evaluation metrics are based on matching accuracy calculated using historical data comparison methods. For example, during vehicle operation, continuous monitoring cycles check the deviation correction accuracy at fixed time intervals, such as 10 seconds. If the accuracy is lower than a preset standard, such as 95%, an iteration is triggered. Specifically, this includes: identifying the type of noise interference in the current attitude data, such as vibration interference or signal offset, and preparing for backtracking based on signal noise filtering technology; extracting the initial signal offset range as the basis for backtracking to ensure that the backtracking is based on the original collected data; and adjusting the backtracking depth using environmental adaptation logic. For example, in high-speed driving scenarios, high-frequency vibration noise is filtered first to improve iteration efficiency. The backtracking mechanism effectively handles data inaccuracies under complex road conditions, which is beneficial to improving the overall reliability of fault detection. For example, in vehicles driving on urban roads, if the deviation correction accuracy is 90% lower than the standard, backtracking is performed to the noise filtering stage to reprocess the original data containing steering signal delays.

[0097] The process involves extracting features of driving path deviation and vibration interference from the new attitude data stream, such as abnormal fluctuations exceeding a tilt angle threshold; integrating real-time feedback mechanisms to analyze continuous changes, for example, using Fourier transform to decompose fluctuation frequencies; identifying potential fault signals and generating judgment results. Specifically, this includes: performing frequency domain analysis on the new attitude data stream to extract frequency features related to vehicle attitude changes; converting the time-domain attitude data to the frequency domain using Fast Fourier Transform to identify low-frequency and high-frequency components related to vehicle tilt and steering, such as low frequencies corresponding to slow attitude changes and high frequencies corresponding to vibration interference, thus obtaining a frequency feature vector; calculating the amplitude spectrum of the frequency features; determining instability if the amplitude exceeds a preset threshold, for example, high-frequency peaks in the amplitude spectrum indicate potential instability under high-speed driving conditions; analyzing the source of abnormal fluctuations based on the stability judgment results to complete the abnormal fluctuation analysis; and confirming the existence of potential fault signals and updating the judgment results for potential fault signals if the source of abnormal fluctuations matches a preset fault mode.

[0098] The adaptive optimization method of the vehicle fault detection system in the embodiments of this application has been described above. The adaptive optimization system of the vehicle fault detection system in the embodiments of this application is described below. Please refer to [link / reference]. Figure 4 One embodiment of the adaptive optimization system for the vehicle fault detection system in this application includes:

[0099] The acquisition module 201 uses on-board sensors to collect attitude data of the vehicle in real time during driving. Based on the preset tilt angle threshold and steering signal delay, the attitude data is detected in real time to generate angular velocity sensing data and vibration interference identification results.

[0100] The identification module 202 is used to preprocess and calibrate the attitude data, generate an attitude data stream, perform feature decomposition on the attitude data stream, identify abnormal fluctuations and determine whether there are potential fault signals, adjust the parameter configuration of the vehicle sensor based on the potential fault signals, update the calibration frequency of the vehicle sensor, and obtain the adjusted parameter configuration information.

[0101] Feedback module 203 is used to analyze the fusion results of attitude data based on the adjusted parameter configuration information, calculate the evaluation index of deviation correction accuracy, build a real-time feedback mechanism based on the evaluation index, dynamically adjust the response time, and generate a fault signal correction scheme.

[0102] The optimization module 204 is used to construct a continuous detection loop based on the fault signal correction scheme, analyze the matching effect between the calibration frequency and the steering signal delay, and iteratively process the attitude data if the deviation correction accuracy does not reach the first preset standard.

[0103] Through the collaborative efforts of the aforementioned components, this application enables timely identification of potential fault signals during vehicle operation by real-time acquisition of vehicle sensor attitude data and combining this data with tilt angle thresholds and steering signal delays. A Kalman filter is used to fuse tilt angle differences and steering signal delay data to generate smooth angular velocity sensing data, thereby improving sensitivity to vehicle attitude changes in dynamic environments and effectively enhancing fault detection accuracy and real-time response capabilities. By correlating and comparing angular velocity sensing data with vibration data and calculating the linear correlation between the two based on the Pearson correlation coefficient, vibration interference caused by non-fault factors, such as uneven road surfaces and engine vibration, can be accurately identified. By analyzing the priority of potential fault signals and assessing their impact on vehicle safety, the operating parameters and calibration frequency of the sensors can be dynamically adjusted according to different driving conditions. Based on historical and real-time acquired data, methods such as autoregressive integral moving average models and wavelet transforms are used to dynamically calibrate, denoise, and extract features from the attitude data. This allows for adaptive adjustment of the signal offset range, accurate identification of abnormal data points, and improved data quality and reliability through iterative processing. A feedback mechanism based on evaluation metrics is constructed by real-time evaluation of deviation correction accuracy and dynamic adjustment of response time. When the system detects that the deviation correction accuracy does not meet the preset standard, it can automatically backtrack to the noise filtering stage, reprocess the data, and perform calibration. By combining steering signal delay, vibration interference identification, and dynamic adjustment of signal offset range, the system can more accurately capture changes in vehicle attitude and dynamic anomalies, avoiding misjudgments caused by slow sensor response or inaccurate data.

[0104] This application also provides a computer-readable storage medium, which can be a non-volatile computer-readable storage medium or a volatile computer-readable storage medium, wherein the computer-readable storage medium stores instructions that, when executed on a computer, cause the computer to perform the steps of the adaptive optimization method for an automotive fault detection system.

[0105] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the specific working processes of the systems and units described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here.

[0106] If the integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this application. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0107] The above-described embodiments are only used to illustrate the technical solutions of this application, and are not intended to limit them. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of this application.

Claims

1. An adaptive optimization method for an automotive fault detection system, characterized in that, The adaptive optimization method for the vehicle fault detection system includes: The vehicle uses onboard sensors to collect attitude data in real time during vehicle operation. Based on a preset tilt angle threshold and steering signal delay, the attitude data is detected in real time to generate angular velocity sensing data and vibration interference identification results. The generation of angular velocity sensing data and vibration disturbance identification results includes: Based on the attitude data, the difference between the actual tilt angle of the monitored vehicle and a preset tilt angle threshold is obtained. The delay data between the trigger time of the steering signal and the actual steering action is recorded based on the steering signal delay. Initial angular velocity sensing data is generated based on the difference in the actual tilt angle and the delay data. Preliminary analysis of the attitude data is performed to obtain vibration data. The angular velocity sensing data and vibration data are correlated and compared to generate the vibration interference identification result and determine the signal offset range. Based on the signal offset range, abnormal data points in the attitude data are filtered out. The attitude data is preprocessed and calibrated to generate an attitude data stream. Feature decomposition is performed on the attitude data stream to identify abnormal fluctuations and determine whether there are potential fault signals. Based on the potential fault signals, the parameter configuration of the vehicle sensor is adjusted and the calibration frequency of the vehicle sensor is updated to obtain the adjusted parameter configuration information. Based on the adjusted parameter configuration information, the fusion results of attitude data are analyzed, the evaluation index of deviation correction accuracy is calculated, a real-time feedback mechanism is constructed based on the evaluation index, the response time is dynamically adjusted, and a fault signal correction scheme is generated. The evaluation index for the accuracy of the calculation deviation correction includes: Based on the adjusted parameter configuration information, the real-time acquired attitude data is recalibrated. According to the recalibrated data, the angular velocity sensing data and vibration disturbance identification results are fused to form a comprehensive attitude data stream, which is then set as the new data fusion result. Based on the new data fusion results, the deviation data of the monitored vehicle on the current driving path is obtained, the deviation data is matched with the historical driving path deviation data, the deviation value between the current attitude data and the historical attitude data is calculated, and an evaluation index of deviation correction accuracy is generated based on the deviation value. Based on the fault signal correction scheme, a continuous detection loop is constructed to analyze the matching effect between the calibration frequency and the steering signal delay. If the deviation correction accuracy does not reach the first preset standard, the attitude data is iteratively processed. The step of constructing a continuous detection loop based on the fault signal correction scheme includes: Based on the fault signal correction scheme, trigger conditions are set, attitude data is collected in real time based on the trigger conditions, the ratio between calibration frequency and steering signal delay is calculated based on the attitude data, and the ratio is set as the matching effect. If the matching effect does not meet the second preset standard, the deviation data in the current detection loop is recorded, deviation correction suggestions are generated, the parameter settings of the detection loop are adjusted, and the operation logic of the detection loop is optimized based on the deviation correction accuracy before and after the parameter adjustment. Obtain the evaluation index of the current deviation correction accuracy. If the evaluation index does not meet the first preset standard, backtrack to the signal and noise filtering stage, reprocess the original attitude data for noise filtering, and generate updated attitude data. The calibration process is re-executed on the updated attitude data to generate a new attitude data stream. Anomaly fluctuation analysis is performed on the new attitude data stream to update the determination results of potential fault signals.

2. The adaptive optimization method for an automotive fault detection system according to claim 1, characterized in that, The generated attitude data stream includes: Based on the signal offset range, the attitude data is denoised using signal noise filtering technology to generate preprocessed attitude data. The sampling direction of the preprocessed attitude data is adjusted based on the sensor orientation correction logic to generate the orientation deviation of the vehicle sensor under different driving conditions. The preprocessed attitude data is then calibrated based on the orientation deviation to generate the attitude data stream. Time series analysis and interpolation are performed on the attitude data stream to optimize its integrity.

3. The adaptive optimization method for an automotive fault detection system according to claim 1, characterized in that, The process of identifying abnormal fluctuations and determining whether there are potential fault signals includes: Based on feature decomposition, feature parameters related to vehicle attitude change in the attitude data stream are obtained, and the path deviation trend of vehicle attitude change is obtained based on the feature parameters. Vibration interference feature values ​​are extracted from the vibration data in the attitude data stream. Based on the vibration interference feature values, it is determined whether there are abnormal fluctuations in the attitude data stream. If the abnormal fluctuations exceed a preset tilt angle threshold, it is determined that there is a potential fault signal. Based on the judgment results, the time points and amplitudes of abnormal fluctuations are recorded, an abnormal fluctuation distribution map is generated, and the periodic characteristics of the abnormal fluctuation distribution map are obtained.

4. The adaptive optimization method for an automotive fault detection system according to claim 1, characterized in that, The process of obtaining the adjusted parameter configuration information includes: Based on the priority of the potential fault signals, the safety impact level of the monitored vehicle is generated. Based on the safety impact level, the operating parameters, response time characteristics and calibration frequency of the on-board sensors are adjusted to generate the adjusted parameter configuration information. Verify the applicability of the parameter configuration information under different driving scenarios, and record the effect of parameter configuration adjustments.

5. The adaptive optimization method for an automotive fault detection system according to claim 1, characterized in that, The fault signal correction scheme includes: The evaluation indicators are analyzed from multiple dimensions, various driving conditions are classified, statistical methods are used to calculate the coefficient of variation of the indicators under each driving condition, a dataset is generated, and the calculation method of the evaluation indicators is optimized based on the dataset. A real-time feedback model is constructed based on the optimized evaluation index, and the changes in angular velocity sensing data are monitored in real time. The response time of the real-time feedback model is dynamically adjusted, and the feedback results for abnormal data are output to generate the fault signal correction scheme. The fault signal correction scheme was initially verified, and its application effect in actual driving was recorded. Based on the application effect, the stability of the fault signal correction scheme was analyzed.

6. A vehicle fault detection system, characterized in that, An adaptive optimization method for implementing an automotive fault detection system as described in any one of claims 1 to 5, the automotive fault detection system comprising: The acquisition module uses onboard sensors to collect attitude data of the vehicle in real time during driving. Based on a preset tilt angle threshold and steering signal delay, the attitude data is detected in real time to generate angular velocity sensing data and vibration interference identification results. The identification module is used to preprocess and calibrate the attitude data, generate an attitude data stream, perform feature decomposition on the attitude data stream, identify abnormal fluctuations and determine whether there are potential fault signals, adjust the parameter configuration of the vehicle sensor based on the potential fault signals, update the calibration frequency of the vehicle sensor, and obtain the adjusted parameter configuration information. The feedback module is used to analyze the fusion results of attitude data based on the adjusted parameter configuration information, calculate the evaluation index of deviation correction accuracy, construct a real-time feedback mechanism based on the evaluation index, dynamically adjust the response time, and generate a fault signal correction scheme. The optimization module is used to construct a continuous detection loop based on the fault signal correction scheme, analyze the matching effect between the calibration frequency and the steering signal delay, and iteratively process the attitude data if the deviation correction accuracy does not reach the first preset standard.

7. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is run by the processor, it causes the processor to execute the adaptive optimization method for the automotive fault detection system as described in any one of claims 1 to 5.

Images