Fault diagnosis method and device, electronic equipment and storage medium

Abstract

The invention relates to the technical field of vehicles, and discloses a fault diagnosis method and device, electronic equipment and a storage medium, the fault diagnosis method comprises the following steps: after a vehicle triggers a fault event, obtaining abnormal detection data and driving scene data associated with the fault event in the operation process of the vehicle; analyzing the correlation between the abnormal detection data and the driving scene data to obtain a correlation analysis result; fault diagnosis is conducted on the vehicle based on the correlation analysis result and historical associated data, a fault diagnosis result corresponding to the fault event is obtained, and the historical associated data is determined based on historical operation data of the vehicle and / or abnormal data of other vehicles when the fault event occurs. And potential safety hazards and maintenance cost of the vehicle are reduced.

Description

Technical Field

[0001] This invention relates to the field of vehicle technology, and more specifically to fault diagnosis methods, devices, electronic equipment, and storage media. Background Technology

[0002] Vehicle fault diagnosis is crucial in vehicle maintenance, especially given the increasing level of vehicle electronics. However, current vehicle fault diagnosis technologies often rely on human experience or single sensor data, resulting in low diagnostic reliability and increasing vehicle safety risks and maintenance costs. Summary of the Invention

[0003] In view of this, embodiments of the present invention provide a fault diagnosis method, apparatus, electronic device, and storage medium.

[0004] In a first aspect, embodiments of the present invention provide a fault diagnosis method, comprising: after a vehicle triggers a fault event, acquiring abnormal detection data and driving scenario data associated with the fault event during vehicle operation; analyzing the correlation between the abnormal detection data and the driving scenario data to obtain a correlation analysis result; and performing fault diagnosis on the vehicle based on the correlation analysis result and historical associated data to obtain a fault diagnosis result corresponding to the fault event, wherein the historical associated data is determined based on the vehicle's historical operating data and / or abnormal data of other vehicles when a fault event occurs.

[0005] The fault diagnosis method provided in this invention obtains abnormal detection data and driving scenario data associated with the fault event during vehicle operation after the vehicle triggers a fault event, analyzes the correlation between the abnormal detection data and driving scenario data, and performs multi-dimensional diagnosis by combining historical associated data. This effectively avoids the limitations of a single data source, thereby significantly improving the accuracy and reliability of fault diagnosis, reducing the risk of misdiagnosis and missed diagnosis, and reducing vehicle safety hazards and maintenance costs.

[0006] In conjunction with the first aspect, in one implementation, acquiring abnormal detection data and driving scenario data associated with fault events during vehicle operation includes: acquiring the vehicle type and the diagnostic scenario for diagnosing the vehicle; acquiring driving data and driving scenario data of the vehicle within a target time period based on the diagnostic scenario and vehicle type; and extracting features from the driving data based on fault events to obtain abnormal detection data of the vehicle within the target time period.

[0007] The fault diagnosis method provided in this invention obtains the vehicle type and diagnosis scenario to collect targeted driving data and driving scenario data that are highly correlated with the fault event. This effectively avoids interference from irrelevant information and improves the accuracy and efficiency of data acquisition. Through a feature extraction process based on the fault event, it ensures that the abnormal detection data has high relevance and representativeness, thereby further enhancing the reliability of subsequent correlation analysis, reducing the impact of noise in the diagnosis process, shortening the fault diagnosis response time, and providing more comprehensive data support for fault prediction under different vehicle models and scenarios.

[0008] In conjunction with the first aspect, in one implementation, the correlation between anomaly detection data and driving scenario data is analyzed to obtain correlation analysis results, including: acquiring spatiotemporal data and environmental data of the vehicle's location based on driving scenario data; calculating a first correlation coefficient between anomaly detection data and spatiotemporal data, and a second correlation coefficient between anomaly detection data and environmental data; and analyzing the correlation between anomaly detection data and driving scenario data based on the first correlation coefficient and / or the second correlation coefficient to obtain correlation analysis results.

[0009] The fault diagnosis method provided in this invention obtains spatiotemporal data and environmental data based on driving scenario data, and calculates the first correlation coefficient and the second correlation coefficient. This can accurately quantify the correlation between abnormal detection data and external factors, thereby further improving the accuracy and reliability of correlation analysis. This helps to identify fault modes in specific scenarios, reduce noise interference in the diagnosis process, optimize resource allocation, and shorten fault response time.

[0010] In conjunction with the first aspect, in one implementation, vehicle fault diagnosis is performed based on correlation analysis results and historical correlation data to obtain fault diagnosis results corresponding to fault events, including: if the correlation analysis results indicate that the correlation between abnormal detection data and driving scenario data is less than a threshold, then vehicle fault diagnosis is performed based on the statistical characteristics of the abnormal detection data to obtain a preliminary diagnosis result; the preliminary diagnosis result is matched with historical diagnosis results in historical correlation data to obtain a diagnosis matching result; if the diagnosis matching result indicates that the preliminary diagnosis result is the same as the historical diagnosis result, then the preliminary diagnosis result is updated based on the maintenance records corresponding to the historical diagnosis results to obtain a fault diagnosis result.

[0011] The fault diagnosis method provided in this invention can effectively isolate noise interference and avoid misjudgment caused by changes in the external environment by performing preliminary diagnosis based on the statistical characteristics of abnormal detection data when the correlation analysis results show that the influence of external driving scenario factors is small. By combining historical correlation data to match the diagnosis results, the accuracy of fault attribution can be improved. By updating the preliminary diagnosis results based on historical maintenance records after successful matching, the practicality and reliability of the diagnosis results can be enhanced, and the false alarm rate of faults can be reduced.

[0012] In conjunction with the first aspect, in one implementation, vehicle fault diagnosis is performed based on the statistical characteristics of abnormal detection data to obtain preliminary diagnostic results, including: determining candidate faulty components of the vehicle based on fault events; obtaining the dependencies and fault priorities between candidate faulty components; if there are no dependencies between candidate faulty components, analyzing the abnormal data corresponding to the candidate faulty components according to the fault priorities to obtain first diagnostic information of the candidate faulty components and the confidence level corresponding to the first diagnostic information; identifying candidate faulty components whose confidence level corresponding to the first diagnostic information is higher than a first preset threshold as faulty components, and generating preliminary diagnostic results based on the faulty components and the corresponding first diagnostic information.

[0013] The fault diagnosis method provided in this invention locates candidate faulty components based on fault events and analyzes their dependencies. When no dependencies exist, it filters high-reliability diagnostic information based on fault priority and confidence threshold, effectively filtering low-confidence noise interference and avoiding misjudgments of non-critical components. Simultaneously, it prioritizes high-priority faults, improving diagnostic accuracy and efficiency. Furthermore, the confidence mechanism reduces redundant data processing, optimizes computational resource allocation, and shortens overall response time. Dependency analysis prevents misdiagnosis caused by component coupling, enhancing the practicality of results and reducing the false negative rate.

[0014] In conjunction with the first aspect, in one implementation, the method of diagnosing vehicle faults based on the statistical characteristics of abnormal detection data to obtain preliminary diagnostic results further includes: if there is a dependency relationship between candidate faulty components, extracting the temporal correlation features of the abnormal detection data of the first candidate faulty component and the abnormal detection data of the second candidate faulty component; performing causal correlation analysis on the abnormal data of the first candidate faulty component and the second candidate faulty component based on the temporal correlation features to obtain second diagnostic information and the confidence level corresponding to the second diagnostic information; identifying candidate faulty components whose confidence level corresponding to the second diagnostic information is higher than a second preset threshold as faulty components, and generating preliminary diagnostic results based on the faulty components and the corresponding second diagnostic information.

[0015] The fault diagnosis method provided in this invention can effectively identify concurrent faults or causal chain faults by capturing the temporal characteristics between faulty components with dependencies. When there is a strong correlation between abnormal data of multiple components, causal correlation analysis can distinguish between fundamental faults and derived faults, avoiding misjudgments caused by isolated diagnoses. At the same time, the extraction of temporal correlation features enhances the diagnostic process's ability to analyze complex fault modes. Under the premise of ensuring diagnostic reliability, it can significantly reduce the diagnostic complexity of highly coupled complex systems, thereby reducing missed reports caused by unclear component interaction relationships, improving the comprehensiveness of fault coverage, and providing a more complete basis for subsequent maintenance decisions.

[0016] In conjunction with the first aspect, in one implementation, the vehicle is diagnosed based on the correlation analysis results and historical correlation data to obtain the fault diagnosis result corresponding to the fault event. The method further includes: if the correlation analysis results indicate that the correlation between the abnormal detection data and the driving scenario data is greater than or equal to a threshold, then the temporal features of the driving scenario data are extracted and causal correlation analysis is performed with the abnormal detection data to obtain third diagnostic information; the confidence weight of the abnormal detection data is adjusted according to the confidence of the third diagnostic information, and the vehicle is diagnosed based on the historical correlation data to generate a fault diagnosis result.

[0017] The fault diagnosis method provided in this invention effectively distinguishes between abnormal data caused by real component failures and transient pseudo-anomalies caused by specific operating conditions (such as extreme road surfaces and severe weather) through dynamic analysis of driving scenario data. When a strong correlation is detected between abnormal data and the current driving scenario, causal correlation analysis is performed using time-series features to accurately isolate the influence of environmental interference factors on the diagnostic conclusions, significantly reducing the false alarm rate. At the same time, the confidence weight of abnormal detection data is dynamically adjusted based on the scenario correlation analysis results, enabling the diagnostic logic to have environmental adaptability, improving the diagnostic robustness of sensitive devices under complex operating conditions such as autonomous driving systems, and further enhancing the system's ability to identify fault modes in rare or complex driving scenarios, ensuring the stability and reliability of diagnostic results in changing environments.

[0018] Secondly, embodiments of the present invention provide a fault diagnosis device, comprising: an acquisition module, configured to acquire abnormal detection data and driving scenario data associated with the fault event during vehicle operation after a fault event is triggered; an analysis module, configured to analyze the correlation between the abnormal detection data and the driving scenario data to obtain a correlation analysis result; and a diagnosis module, configured to perform fault diagnosis on the vehicle based on the correlation analysis result and historical correlation data, and generate a fault diagnosis result, wherein the historical correlation data is determined based on the vehicle's historical operating data and / or abnormal data of other vehicles when the fault event occurs.

[0019] Thirdly, embodiments of the present invention provide an electronic device, including: a memory and a processor, the memory and the processor being communicatively connected to each other, the memory storing computer instructions, and the processor executing the computer instructions to perform the fault diagnosis method of the first aspect or any corresponding embodiment described above.

[0020] Fourthly, embodiments of the present invention provide a computer-readable storage medium storing computer instructions, which are used to cause a computer to perform the fault diagnosis method of the first aspect or any corresponding embodiment described above.

[0021] Fifthly, embodiments of the present invention provide a computer program product, including computer instructions, which are used to cause a computer to execute the in-vehicle environment control method of the first aspect or any corresponding embodiment described above. Attached Figure Description

[0022] To more clearly illustrate the specific embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the specific embodiments or the prior art 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 from these drawings without creative effort.

[0023] Figure 1 This is a flowchart illustrating a fault diagnosis method according to some embodiments of the present invention; Figure 2 This is a flowchart illustrating another fault diagnosis method according to some embodiments of the present invention; Figure 3 This is a structural block diagram of a fault diagnosis device according to an embodiment of the present invention; Figure 4 This is a schematic diagram of the hardware structure of an electronic device according to an embodiment of the present invention. Detailed Implementation

[0024] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0025] According to an embodiment of the present invention, a fault diagnosis method embodiment is provided. It should be noted that the steps shown in the flowchart in the accompanying drawings can be executed in a computer system such as a set of computer-executable instructions. Furthermore, although a logical order is shown in the flowchart, in some cases, the steps shown or described may be executed in a different order than that shown here.

[0026] This embodiment provides a fault diagnosis method. Figure 1 This is a flowchart of a fault diagnosis method according to an embodiment of the present invention, such as... Figure 1 As shown, the process includes the following steps: Step S101: After the vehicle triggers a fault event, acquire the abnormal detection data and driving scenario data associated with the fault event during vehicle operation.

[0027] Vehicle-triggered fault events refer to events in which the vehicle detects preset fault indicators exceeding the threshold range or the system generates fault alarm signals during operation. These events include, but are not limited to, engine abnormalities, brake system failures, battery voltage abnormalities, sensor failures, control unit errors, or abnormal driving behavior. These events are monitored in real time by the on-board diagnostic system (hereinafter referred to as the system) and trigger fault event notifications.

[0028] Furthermore, the anomaly detection data is extracted from driving data during vehicle operation. This driving data includes basic operational data, state evolution data, and hardware health data that reflect the real-time operating conditions of the vehicle's core components. Driving scenario data is determined based on spatiotemporal data and environmental data related to the vehicle's location.

[0029] Based on the basic operating data, it can be collected through the vehicle's CAN bus sensor interface. The collected data includes, but is not limited to, engine speed, vehicle speed, engine oil temperature, coolant pressure, brake pedal travel, and battery voltage. The collection frequency can be 100Hz. Its core function is to reflect the real-time operating conditions of the vehicle's core components and monitor the operating status of the power system, transmission system, and braking system.

[0030] Status evolution data can be collected through a reconfigurable I / O interface module (supporting dynamic switching of multiple communication protocols such as UART, CAN, FlexRay, and Ethernet, and adaptable to sensor interface types of different vehicle models). The collected content includes, but is not limited to, the duration of abnormal sensor data, the amplitude of abnormal value fluctuations, and the trend of coordinated changes of related components. The collection frequency can be 100Hz. Its core function is to capture the dynamic evolution process of faults from "budding" to "manifestation" and avoid misjudgment caused by instantaneous interference.

[0031] Hardware health data can be collected through built-in monitoring sensors in the ECU. The collected data includes, but is not limited to, diagnosing the internal temperature of the ECU, the stability of the power supply voltage (range: 9V–16V, monitoring fluctuations), and the signal strength of the communication module. The collection frequency can be 50Hz. Its core function is to eliminate false diagnoses caused by hardware abnormalities of the ECU itself and to ensure the reliability of fault diagnosis.

[0032] Spatiotemporal data can be collected via a GPS module (dual-mode Beidou and GPS) + RTC (real-time clock). The collected data includes, but is not limited to, vehicle latitude and longitude (accuracy ≤ 10m) and diagnostic trigger time (accuracy ≤ 11ms). The collection frequency can be 10Hz for location data and 1ms for time data. Its core function is to correlate the specific scenario in which the fault occurs, such as recurring faults in a specific road segment or time period.

[0033] Environmental data can be collected through environmental sensors (such as integrated temperature, pressure and altitude sensors). The collected data includes, but is not limited to, the ambient temperature range (-40℃ to 85℃), the atmospheric pressure range (50kPa to 110kPa), and the altitude range (-500m to 5000m). The collection frequency can be 10Hz. Its core function is to troubleshoot faults caused by environmental factors, such as abnormal engine power in high-altitude areas.

[0034] When extracting anomaly detection data, the isolated forest algorithm (tree depth ≤ 15 layers, detection time < 1ms) can be used first to identify anomalies such as data jumps and constant values ​​in driving data. Then, linear interpolation (for continuous data) and pattern matching (for discrete data) are used to fill in missing values, with an accuracy of ≥ 95%. Kalman filtering (for dynamic data such as vehicle speed) and moving average filtering (for static data such as ambient temperature) are used for noise filtering, improving the signal-to-noise ratio by 25dB. Finally, the time-domain data is converted into frequency-domain features through Fast Fourier Transform (FFT, 256 sampling points, transformation time < 2ms) to extract the periodic fluctuations of the data (such as abnormal engine vibration frequency). At the same time, the statistical characteristics of the abnormal data (mean, variance, and peak value of each sensor) are calculated to obtain the anomaly detection data of the vehicle during driving.

[0035] Step S102: Analyze the correlation between anomaly detection data and driving scenario data to obtain correlation analysis results.

[0036] The correlation analysis results include correlations that are less than or greater than or equal to a threshold.

[0037] Step S103: Based on the correlation analysis results and historical associated data, perform fault diagnosis on the vehicle to obtain the fault diagnosis results corresponding to the fault events.

[0038] Historical correlation data is determined based on the vehicle's historical operating data and / or abnormal data from other vehicles during malfunction events. This data can be collected synchronously with the cloud platform via a local storage module. The collected data includes, but is not limited to, records of similar sensor malfunctions within the past 30 days, the time and content of the most recent maintenance / repair, and the matching degree of common malfunction cases for the same vehicle model. The collection frequency can be accessed on demand (when diagnostics are triggered). Its core function is to improve the accuracy of malfunction attribution by combining historical data; for example, if a recently replaced sensor malfunctions again, it may be due to an installation problem.

[0039] Specifically, by comparing the correlation analysis results with historical data, a pre-defined fault diagnosis algorithm (such as one based on a rule engine or machine learning model) is used to calculate the probability distribution and root cause analysis of fault events. For example, when the correlation is greater than or equal to a threshold, common faults of the same vehicle model in historical cases (such as sensor signal drift or wiring aging) are prioritized, and the most recent maintenance record is used to determine whether it is due to human error. Conversely, if the correlation is less than the threshold, abnormal data from other vehicles in the cloud platform are used for horizontal comparison to identify environmental factors (such as insufficient power due to high altitude and low air pressure) or occasional hardware failures. The diagnostic results include fault codes, suggested repair measures, and confidence scores, thereby assisting repair personnel in quickly locating the source of the problem. For example, for sensor faults caused by installation problems, replacement suggestions or recalibration instructions are output.

[0040] In addition, the data transmission interval for anomaly detection data, driving scenario data, and historical associated data can be determined based on the corresponding data priority, and device fingerprint authentication can be performed to ensure the security of data transmission.

[0041] The fault diagnosis method provided in this invention obtains abnormal detection data and driving scenario data associated with the fault event during vehicle operation after the vehicle triggers a fault event, analyzes the correlation between the abnormal detection data and driving scenario data, and performs multi-dimensional diagnosis by combining historical associated data. This effectively avoids the limitations of a single data source, thereby significantly improving the accuracy and reliability of fault diagnosis, reducing the risk of misdiagnosis and missed diagnosis, and reducing vehicle safety hazards and maintenance costs.

[0042] In one possible implementation, when acquiring anomaly detection data and driving scenario data associated with fault events during vehicle operation, the vehicle type and the diagnostic scenario for vehicle diagnosis can be obtained first. Then, based on the diagnostic scenario and vehicle type, the vehicle's driving data and driving scenario data within the target time period can be acquired. Finally, feature extraction is performed on the driving data based on the fault event to obtain the vehicle's anomaly detection data within the target time period. The vehicle type includes, but is not limited to, gasoline vehicles, electric vehicles, and hybrid vehicles, and the diagnostic scenario includes, but is not limited to, routine inspections and fault repairs. Furthermore, the data collection frequency in fault repair and diagnostic scenarios can be increased, and the feature extraction weights for driving data corresponding to different vehicle types after a fault event is triggered also differ.

[0043] Specifically, vehicle type information is collected in real time through the vehicle's onboard diagnostic interface or remote communication module. For example, the vehicle identification number (VIN) is used to automatically query the database to determine whether it is a gasoline-powered vehicle, an electric vehicle, or a hybrid vehicle. Simultaneously, diagnostic scenarios can be set based on the triggering source of fault events, such as user reports or automatic system detection, distinguishing between routine checks (e.g., data scans during regular maintenance) and fault repairs (e.g., handling sudden alarms). Subsequently, by calling the API interface in the cloud platform, driving data within the target time period is filtered according to the diagnostic scenario and vehicle type. This includes real-time parameters such as vehicle speed, engine speed, and battery status, as well as driving scenario data such as ambient temperature, altitude, and road gradient. In the feature extraction stage, for fault events (such as abnormal sensor signals), the driving data is preprocessed and features are extracted, such as calculating moving averages, variances, peak detection, or spectral analysis, to identify abnormal patterns such as drift and noise, thereby generating structured anomaly detection data. This process can also be cross-validated with historical correlation databases to ensure the comprehensiveness and accuracy of data collection.

[0044] As an example, suppose a hybrid vehicle triggers an engine power reduction alarm (fault event) while driving at high speed. The system automatically identifies the vehicle type as hybrid and the diagnostic scenario as fault repair (based on automatic system detection triggering) through the on-board diagnostic interface. Subsequently, it calls the cloud platform API interface to filter driving data within a target time period (e.g., 30 minutes before the fault occurred) based on the diagnostic scenario and vehicle type. This includes real-time parameters such as engine speed, motor output power, and fuel consumption rate, as well as driving scenario data such as road gradient, wind speed, and outside temperature. In the feature extraction stage, for the fault event, the engine speed data is preprocessed, including noise removal and calculation of moving average, variance, and spectral analysis to identify abnormal fluctuation patterns (such as periodic oscillations). Simultaneously, cross-validation is performed using similar fault cases in the historical correlation database to ensure that the detected abnormal patterns (such as power output drift) match known fault features, thereby generating a structured anomaly detection data report, including a fault probability score and potential cause analysis.

[0045] As an example, suppose an electric vehicle triggers a battery overheating alarm (fault event) during fast charging. The system automatically identifies the vehicle type as an electric vehicle and the diagnostic scenario as fault repair (based on user-reported triggering) through the on-board diagnostic interface. Subsequently, it calls the cloud platform API interface to filter driving data within a target time period (e.g., within 15 minutes of charging start) based on the diagnostic scenario and vehicle type. This includes real-time parameters such as battery voltage, current, temperature, and charging efficiency, as well as driving scenario data such as charging station ambient temperature, humidity, and charging pile power output. In the feature extraction stage, the battery temperature data is preprocessed for the fault event, including smoothing and data standardization, and moving average, standard deviation, and spectral analysis are calculated to identify abnormal temperature rise patterns (e.g., non-linear upward trends). Simultaneously, cross-validation is performed using similar electric vehicle fault cases in the historical correlation database to ensure that the detected abnormal patterns (e.g., sudden temperature gradient changes) match known overheating fault characteristics, thereby generating a structured anomaly detection data report, including a fault probability score, potential cause analysis (e.g., cooling system failure or charging overload), and recommended maintenance measures.

[0046] The fault diagnosis method provided in this invention obtains the vehicle type and diagnosis scenario to collect targeted driving data and driving scenario data that are highly correlated with the fault event. This effectively avoids interference from irrelevant information and improves the accuracy and efficiency of data acquisition. Through a feature extraction process based on the fault event, it ensures that the abnormal detection data has high relevance and representativeness, thereby further enhancing the reliability of subsequent correlation analysis, reducing the impact of noise in the diagnosis process, shortening the fault diagnosis response time, and providing more comprehensive data support for fault prediction under different vehicle models and scenarios.

[0047] In one possible implementation, when analyzing the correlation between anomaly detection data and driving scenario data to obtain correlation analysis results, the spatiotemporal data and environmental data of the vehicle's location in time and space can be obtained based on the driving scenario data; a first correlation coefficient between anomaly detection data and spatiotemporal data, and a second correlation coefficient between anomaly detection data and environmental data can be calculated; based on the first correlation coefficient and / or the second correlation coefficient, the correlation between anomaly detection data and driving scenario data can be analyzed to obtain correlation analysis results.

[0048] Specifically, spatiotemporal data can include the vehicle's location coordinates and timestamps, acquired in real time via GPS modules; environmental data can include temperature, humidity, road conditions, and weather conditions, collected through onboard sensors or external meteorological databases. When calculating the first correlation coefficient, for example, the Pearson correlation coefficient algorithm is used to analyze the linear correlation between anomaly detection data and spatiotemporal data; similarly, a second correlation coefficient is calculated. When analyzing the nonlinear correlation between anomaly detection data and environmental data, the Spearman rank correlation coefficient can also be introduced to handle non-normally distributed data. Based on the values ​​of the first and second correlation coefficients, combined with preset thresholds (e.g., above 0.7 indicates a strong correlation), the correlation strength is comprehensively evaluated, generating correlation analysis results, including the weight of environmental factors on the probability of failure, thereby supporting more accurate fault diagnosis and prediction model optimization.

[0049] As an example, suppose a hybrid vehicle is traveling on a highway. During continuous acceleration, the onboard diagnostic system detects abnormal fluctuations in engine speed, peaking at 5000 rpm. At this time, the real-time GPS data includes the location coordinates of 40.0°N, 116.5°E, and a timestamp of 10:15 AM on November 15, 2023; simultaneously, the ambient temperature is measured at -5°C by a temperature sensor, the relative humidity is shown at 45% by a humidity sensor, and the road condition database indicates that the current road section is a dry asphalt surface and the weather is sunny.

[0050] The first correlation coefficient between the anomaly detection data (engine speed sequence) and the spatiotemporal data was calculated: using the Pearson correlation coefficient algorithm, the input speed data sequence [4800, 4900, 5050, 4950] and the corresponding timestamp sequence [10:10, 10:12, 10:14, 10:16] were used, and the first correlation coefficient was calculated to be 0.82, which is significantly higher than the preset threshold of 0.7. This indicates that there is a strong linear correlation between speed fluctuation and driving time, which may be due to mechanical fatigue caused by long-term high-speed operation.

[0051] Next, the second correlation coefficient between the anomaly detection data and the environmental data was calculated: for the non-normally distributed temperature data (sequence [-4,-5,-6,-5]), the Spearman rank correlation coefficient was introduced to analyze the nonlinear correlation between the speed sequence and the temperature sequence. The second correlation coefficient was found to be 0.68, which is close to the threshold but slightly lower than 0.7, indicating that the low temperature environment has a weak but not negligible impact on the speed anomaly.

[0052] Based on the first and second correlation coefficients, the correlation strength is comprehensively evaluated: the influence weight of environmental factors (such as low temperature) is 0.3, and the weight of spatiotemporal factors (time variation) is 0.7; the correlation analysis results are generated, indicating that the failure probability is mainly dominated by the time cumulative effect, increasing the sensitivity threshold for monitoring running time, thereby improving the diagnostic accuracy, such as triggering a preventive maintenance alarm after continuous high-speed driving for more than 1 hour.

[0053] The fault diagnosis method provided in this invention obtains spatiotemporal data and environmental data based on driving scenario data, and calculates the first correlation coefficient and the second correlation coefficient. This can accurately quantify the correlation between abnormal detection data and external factors, thereby further improving the accuracy and reliability of correlation analysis. This helps to identify fault modes in specific scenarios, reduce noise interference in the diagnosis process, optimize resource allocation, and shorten fault response time.

[0054] In one possible implementation, when performing fault diagnosis on a vehicle based on correlation analysis results and historical associated data to obtain fault diagnosis results corresponding to fault events, if the correlation analysis results indicate that the correlation between the abnormal detection data and the driving scenario data is less than a threshold, then the vehicle is diagnosed based on the statistical characteristics of the abnormal detection data to obtain a preliminary diagnosis result; the preliminary diagnosis result is matched with the historical diagnosis results in the historical associated data to obtain a diagnosis matching result; if the diagnosis matching result indicates that the preliminary diagnosis result is the same as the historical diagnosis result, then the preliminary diagnosis result is updated based on the maintenance records corresponding to the historical diagnosis results to obtain the fault diagnosis result.

[0055] Specifically, the statistical characteristics of abnormal detection data include, but are not limited to, mean, variance, standard deviation, and distribution characteristics, used to identify potential fault modes. For example, when the mean of abnormal speed data deviates from a preset benchmark value by more than 10%, it is determined to be a preliminary fault event. During the matching process, the fault code, timestamp, and fault type in the preliminary diagnosis result can be compared with historical diagnosis results in historical associated data. If the fault codes are consistent and the time interval is within a preset range (e.g., within 30 days), the diagnostic matching result is confirmed to be valid. If the diagnostic matching results are consistent, the maintenance records corresponding to the historical diagnostic results are further retrieved. These records include component replacement logs, calibration data, and repair history. Based on this, the preliminary diagnosis result is updated. For example, if the historical maintenance record shows that the same fault has been calibrated and adjusted, the diagnostic result is optimized to "recalibration is required instead of component replacement," thereby generating the final fault diagnosis result.

[0056] As an example, suppose that during vehicle operation, the onboard sensors detect abnormal engine speed data with a mean deviation of 12% from a preset baseline value, a significantly increased variance, and a non-normally skewed distribution. Based on these statistical characteristics, this is identified as a preliminary fault event, with fault code D002 and a timestamp of October 15, 2023. During the matching process, the preliminary diagnostic result is compared with records in the historical database. A historical diagnostic result with the same fault code D002 was found 28 days prior, with the same fault type and a time interval within the preset 30-day range. Therefore, the diagnostic matching result is confirmed as valid. Further retrieval of the maintenance record corresponding to this historical diagnostic result reveals that the component replacement log is empty, and the calibration data indicates that the previous fault was resolved through software parameter adjustments. The repair history includes two calibration operations. Based on this maintenance record, the preliminary diagnostic result is updated and optimized to "recalibration is required instead of hardware replacement," thus generating the final fault diagnosis result: "System calibration abnormal, parameter optimization recommended."

[0057] The fault diagnosis method provided in this invention can effectively isolate noise interference and avoid misjudgment caused by changes in the external environment by performing preliminary diagnosis based on the statistical characteristics of abnormal detection data when the correlation analysis results show that the influence of external driving scenario factors is small. By combining historical correlation data to match the diagnosis results, the accuracy of fault attribution can be improved. By updating the preliminary diagnosis results based on historical maintenance records after successful matching, the practicality and reliability of the diagnosis results can be enhanced, and the false alarm rate of faults can be reduced.

[0058] In one possible implementation, when performing fault diagnosis on a vehicle based on the statistical characteristics of anomaly detection data and obtaining a preliminary diagnosis result, candidate faulty components of the vehicle can be identified based on fault events; the dependencies and fault priorities between candidate faulty components can be obtained; if there are no dependencies between candidate faulty components, the abnormal data corresponding to the candidate faulty components can be analyzed according to the fault priority to obtain the first diagnostic information of the candidate faulty components and the confidence level corresponding to the first diagnostic information; candidate faulty components with a confidence level corresponding to the first diagnostic information higher than a first preset threshold are identified as faulty components, and a preliminary diagnosis result is generated based on the faulty components and the corresponding first diagnostic information.

[0059] Specifically, if there are dependencies between candidate faulty components—for example, if the failure of one component may trigger a chain reaction leading to abnormalities in other components—then the fault priorities will be reassessed based on preset dependency rules (such as fault tree models or cause-effect graphs), prioritizing high-priority components to eliminate interference. When analyzing abnormal data corresponding to candidate faulty components, statistical feature extraction methods are employed, including calculating mean shift, analysis of variance, or applying clustering algorithms, to identify abnormal patterns and quantify confidence levels. The confidence threshold can be dynamically adjusted based on the fault distribution in the historical correlation database to improve the accuracy of the analysis. Furthermore, if the dependencies are complex, iterative diagnosis can be performed until all components are independently evaluated or dependency conflicts are resolved. After the analysis is completed, the generated preliminary diagnostic results not only include the identification and diagnostic information of the faulty components but also record the confidence values, providing input for subsequent matching against the historical correlation database and ensuring the consistency and traceability of the diagnostic process.

[0060] As an example, when an abnormally increased engine vibration is detected during vehicle operation, candidate faulty components, including engine mounts, crankshaft bearings, or ignition coils, can be identified based on fault events (such as vibration sensor readings exceeding a preset threshold). When obtaining the dependencies between these components, it is discovered that an ignition coil failure may cause abnormal crankshaft bearing vibration (analyzed using a predefined fault tree model). Therefore, the fault priority is reassessed, and the ignition coil is addressed first to eliminate cascading interference. When analyzing abnormal data, a clustering algorithm (such as K-means) is used to group vibration signals, calculating mean offset and variance to quantify the degree of anomaly. The confidence threshold is dynamically adjusted to 90% based on similar fault distributions in the historical database (such as a high failure rate of ignition coils under high temperatures). Through iterative diagnosis, after resolving the dependency conflict between the ignition coil and crankshaft bearing, the confidence level of the ignition coil reaches 95%, identifying it as a faulty component and generating preliminary diagnostic results, including fault identification, diagnostic description (such as spark loss due to coil aging), and confidence value, providing a traceable record for subsequent database matching.

[0061] Furthermore, in other examples, correlation analysis results and historical data can be input into a deep learning model for fault classification and diagnosis. For example, the self-attention mechanism in a deep learning model can be used to capture the causal relationship of "sensor A malfunction → sensor B fluctuation → system failure". For instance, when "water temperature sensor malfunction" and "fan speed sensor malfunction" occur simultaneously, the model can determine that it is a chain reaction caused by "fan control module failure" rather than an independent failure of the two sensors.

[0062] The fault diagnosis method provided in this invention locates candidate faulty components based on fault events and analyzes their dependencies. When no dependencies exist, it filters high-reliability diagnostic information based on fault priority and confidence threshold, effectively filtering low-confidence noise interference and avoiding misjudgments of non-critical components. Simultaneously, it prioritizes high-priority faults, improving diagnostic accuracy and efficiency. Furthermore, the confidence mechanism reduces redundant data processing, optimizes computational resource allocation, and shortens overall response time. Dependency analysis prevents misdiagnosis caused by component coupling, enhancing the practicality of results and reducing the false negative rate.

[0063] In one possible implementation, when performing vehicle fault diagnosis based on the statistical characteristics of anomaly detection data and obtaining preliminary diagnostic results, if there is a dependency relationship between candidate faulty components, the temporal correlation features of the anomaly detection data of the first candidate faulty component and the anomaly detection data of the second candidate faulty component with the dependency relationship are extracted; causal correlation analysis is performed on the anomaly data of the first and second candidate faulty components based on the temporal correlation features to obtain second diagnostic information and the confidence level corresponding to the second diagnostic information; candidate faulty components with a confidence level corresponding to the second diagnostic information higher than a second preset threshold are identified as faulty components, and preliminary diagnostic results are generated based on the faulty components and the corresponding second diagnostic information.

[0064] Specifically, time-series analysis models (such as autocorrelation functions or long short-term memory networks) are used to extract temporal correlation features between the anomaly detection data of the first candidate faulty component (e.g., water temperature sensor) and the second candidate faulty component (e.g., fan speed sensor), including hysteresis correlation, fluctuation patterns, and synchronous anomalies. Then, based on these temporal correlation features, a causal inference model (such as Granger causality tests or self-attention-based neural networks) is applied for deep correlation analysis to identify causal chains between components (e.g., "abnormal water temperature → fan speed fluctuation → control module failure"), thereby generating second diagnostic information (e.g., "fan control module failure") and its corresponding confidence score. If the confidence score of the second diagnostic information is greater than a second preset threshold (e.g., 0.95), the candidate faulty component is automatically marked as a confirmed faulty component, and relevant second diagnostic information is integrated according to fault priority (e.g., higher priority for core system components), ultimately outputting a structured preliminary diagnostic result, ensuring clear diagnostic logic and efficient resource utilization.

[0065] As an example, consider a fault scenario in the power battery management system of an electric vehicle. Assume the first candidate faulty component is the battery temperature sensor detecting abnormal high-temperature fluctuations, while the second candidate faulty component is the battery cooling fan speed sensor displaying abnormally low-speed fluctuations. Using a time-series analysis model, such as a Long Short-Term Memory (LSTM) network, the temporal correlation features of the abnormal detection data for these two components are extracted: including the lag correlation of the fan speed decreasing after the water temperature rises (time lag approximately 30 seconds), fluctuation patterns (periodic spikes), and synchronous anomalies (simultaneous occurrence of temperature exceeding limits and excessively low speed). Then, based on these temporal correlation features, a causal inference model, such as a self-attention-based neural network, is applied to perform deep correlation analysis: identifying the causal chain "abnormal battery temperature → abnormal cooling fan speed → battery management control module fault," and generating the second diagnostic information "battery management control module fault" with a confidence score of 0.97. Since the confidence level is higher than the second preset threshold of 0.95, the battery management control module is automatically marked as a confirmed faulty component. Based on the fault priority (the battery management control module is a core component and has a high priority), relevant second diagnostic information is integrated to output a structured preliminary diagnostic result: "Faulty component: Battery management control module; Diagnostic information: Control signal failure leads to abnormal heat dissipation; Priority: High". This ensures that the diagnostic logic is clear and resource-efficient, facilitating subsequent maintenance decisions.

[0066] The fault diagnosis method provided in this invention can effectively identify concurrent faults or causal chain faults by capturing the temporal characteristics between faulty components with dependencies. When there is a strong correlation between abnormal data of multiple components, causal correlation analysis can distinguish between fundamental faults and derived faults, avoiding misjudgments caused by isolated diagnoses. At the same time, the extraction of temporal correlation features enhances the diagnostic process's ability to analyze complex fault modes. Under the premise of ensuring diagnostic reliability, it can significantly reduce the diagnostic complexity of highly coupled complex systems, thereby reducing missed reports caused by unclear component interaction relationships, improving the comprehensiveness of fault coverage, and providing a more complete basis for subsequent maintenance decisions.

[0067] This embodiment provides a fault diagnosis method. Figure 2 This is a flowchart of a fault diagnosis method according to an embodiment of the present invention, such as... Figure 2 As shown, the process includes the following steps: Step S201: After a vehicle malfunction event is triggered, acquire anomaly detection data and driving scenario data associated with the malfunction event during vehicle operation. For details, please refer to [link to relevant documentation]. Figure 1 Step S101 of the illustrated embodiment will not be described again here.

[0068] Step S202: Analyze the correlation between anomaly detection data and driving scenario data to obtain the correlation analysis results. For details, please refer to [link to relevant documentation]. Figure 1 Step S102 of the illustrated embodiment will not be described again here.

[0069] Step S203: Based on the correlation analysis results and historical associated data, perform fault diagnosis on the vehicle to obtain the fault diagnosis results corresponding to the fault events.

[0070] Specifically, step S203 includes: Step S2031: If the correlation analysis results indicate that the correlation between the abnormal detection data and the driving scenario data is greater than or equal to the threshold, then the temporal features of the driving scenario data are extracted and causal association analysis is performed with the abnormal detection data to obtain the third diagnostic information.

[0071] Step S2032: Adjust the confidence weight of the abnormal detection data according to the confidence of the third diagnostic information, and combine it with historical correlation data to perform fault diagnosis on the vehicle and generate fault diagnosis results.

[0072] Specifically, for example, when a vehicle is driving at high temperatures and speeds, the driving scenario data includes parameters such as ambient temperature, vehicle speed, and engine load. A sliding window analysis is used to extract its temporal characteristics, such as peak detection, trend changes, and periodic fluctuations. Next, causal correlation analysis is performed between these temporal characteristics and abnormal detection data (such as battery temperature sensor readings). Granger causality tests or Bayesian network models are used to identify whether the control signal failure is a fundamental fault (such as a blocked cooling system) or a derivative fault (such as a false alarm caused by voltage fluctuations). If the confidence level of the third diagnostic information is higher than a preset threshold (such as 0.85), the weight of the abnormal detection data is increased to 1.2 times; otherwise, it is reduced to 0.8 times to optimize diagnostic priority. Finally, by combining similar fault cases from a historical database, a weighted fusion algorithm is used to generate fault diagnosis results, ensuring that the output includes structured information covering faulty components, root cause analysis, and repair recommendations.

[0073] As an example, in low-temperature, low-speed driving scenarios, driving scenario data includes parameters such as ambient temperature, vehicle speed, engine idle speed, and tire pressure monitoring. Temporal features, such as temperature gradient changes, idle speed stability, and pressure fluctuation patterns, are extracted using sliding window analysis. Next, causal correlation analysis is performed between these temporal features and abnormal detection data (such as brake system sensor readings). Granger causality tests or Bayesian network models are used to identify whether control signal failure is a fundamental fault (such as brake fluid leakage) or a derivative fault (such as ABS module erroneous triggering). If the confidence level of the third diagnostic information is higher than a preset threshold, the weight of the abnormal detection data is increased to 1.5 times; otherwise, it is reduced to 0.7 times to optimize diagnostic priority. Finally, by combining similar fault cases from a historical database, a weighted fusion algorithm is used to generate fault diagnosis results, ensuring that the output includes structured information covering faulty components, root cause analysis, and repair recommendations.

[0074] Furthermore, in rainy, snowy, and slippery scenarios, driving scenario data can include parameters such as road surface humidity, vehicle speed, and steering angle. Temporal features, such as slip ratio changes, steering response delays, and traction control fluctuations, can be extracted through sliding window analysis. Causal correlation analysis is then performed between these temporal features and abnormal detection data (such as wheel speed sensor readings). A Bayesian network model is used to identify fundamental faults (such as sensor contamination) or derivative faults (such as ESP system false alarms). Weights are dynamically adjusted based on confidence levels, and the diagnostic results are optimized by combining historical correlation data. A diagnostic report is generated based on the fault diagnosis results, including the time series of fault events, confidence scores, and repair priorities, which facilitates subsequent maintenance decisions.

[0075] The fault diagnosis method provided in this invention effectively distinguishes between abnormal data caused by real component failures and transient pseudo-anomalies caused by specific operating conditions (such as extreme road surfaces and severe weather) through dynamic analysis of driving scenario data. When a strong correlation is detected between abnormal data and the current driving scenario, causal correlation analysis is performed using time-series features to accurately isolate the influence of environmental interference factors on the diagnostic conclusions, significantly reducing the false alarm rate. At the same time, the confidence weight of abnormal detection data is dynamically adjusted based on the scenario correlation analysis results, enabling the diagnostic logic to have environmental adaptability, improving the diagnostic robustness of sensitive devices under complex operating conditions such as autonomous driving systems, and further enhancing the system's ability to identify fault modes in rare or complex driving scenarios, ensuring the stability and reliability of diagnostic results in changing environments.

[0076] In one possible implementation, vehicle fault diagnosis can be performed based on correlation analysis results and historical correlation data, i.e., driver operating habits, to obtain fault diagnosis results corresponding to fault events. Specifically, this includes: analyzing the degree of matching between correlation analysis results and driving scenario data to identify whether abnormal data is related to the current operating condition; establishing a personalized behavioral baseline model using driver operating habits (such as acceleration, braking, and steering modes) from historical correlation data, and dynamically calibrating the anomaly detection threshold; performing causal inference by combining temporal features to distinguish between anomalies caused by actual component failures and transient deviations caused by driver operating habits; integrating correlation analysis results with historical habit data through a weighted fusion mechanism to generate environment-adaptive fault diagnosis results; and finally, dynamically adjusting weights based on diagnostic confidence to output the type, location, and repair recommendations of the fault event.

[0077] As an example, when an abnormal acceleration event is detected while the vehicle is driving on the highway, the correlation analysis results are first analyzed to determine the degree of matching with the current driving scenario (such as changes in slope or strong wind conditions) to identify whether the acceleration anomaly originates from interference under real-world conditions. Subsequently, the driver's acceleration habits in historical correlation data (e.g., an average acceleration depth of 30% on flat roads and a response delay of less than 0.5 seconds) are used to construct a personalized behavioral baseline model, dynamically calibrating the detection threshold to a deviation range of ±10%. Causal inference is then performed by combining temporal characteristics (such as the fluctuation trend of acceleration data within a 5-second window) to distinguish whether the persistent abnormality is caused by a throttle sensor malfunction or a transient deviation caused by the driver's habitual rapid acceleration. The correlation analysis results and habitual data are integrated through a weighted fusion mechanism (environmental factors weight 0.6, historical habits weight 0.4) to generate an environment-adaptive fault diagnosis result. Finally, the output weights are dynamically adjusted based on the diagnostic confidence level (e.g., 85%) to confirm the fault event as "throttle actuator sticking," pinpointing the right front wheel drive unit, and recommending priority replacement of the component and calibration testing.

[0078] This embodiment also provides a fault diagnosis device for implementing the above embodiments and preferred embodiments; details already described will not be repeated. As used below, the term "module" can refer to a combination of software and / or hardware that performs a predetermined function. Although the device described in the following embodiments is preferably implemented in software, hardware implementation, or a combination of software and hardware, is also possible and contemplated.

[0079] This embodiment provides a fault diagnosis device, such as Figure 3 As shown, it includes: The acquisition module 301 is used to acquire abnormal detection data and driving scenario data associated with the fault event during the vehicle's operation after the vehicle triggers a fault event; Analysis module 302 is used to analyze the correlation between anomaly detection data and driving scenario data to obtain correlation analysis results; The diagnostic module 303 is used to perform fault diagnosis on the vehicle based on the correlation analysis results and historical correlation data, and generate fault diagnosis results. The historical correlation data is determined based on the vehicle's historical operating data and / or abnormal data of other vehicles when a fault event occurs.

[0080] In one possible implementation, the acquisition module 301 includes: The information acquisition unit is used to acquire the vehicle type and the diagnostic scenario for vehicle diagnosis. The data acquisition unit is used to acquire vehicle driving data and driving scenario data within a target time period based on the diagnostic scenario and vehicle type. The feature extraction unit is used to extract features from driving data based on fault events to obtain abnormal detection data of the vehicle within the target time period.

[0081] In one possible implementation, the analysis module 302 includes: The scene data acquisition unit is used to acquire spatiotemporal data and environmental data of the vehicle's location based on driving scene data. The correlation coefficient calculation unit is used to calculate the first correlation coefficient between anomaly detection data and spatiotemporal data, and the second correlation coefficient between anomaly detection data and environmental data; The correlation analysis unit is used to analyze the correlation between anomaly detection data and driving scenario data based on the first correlation coefficient and / or the second correlation coefficient, and obtain the correlation analysis results.

[0082] In one possible implementation, the diagnostic module 303 includes: The preliminary diagnosis unit is used to perform vehicle fault diagnosis based on the statistical characteristics of the abnormal detection data if the correlation analysis results indicate that the correlation between the abnormal detection data and the driving scenario data is less than a threshold, and to obtain a preliminary diagnosis result. The diagnostic matching unit is used to match the preliminary diagnostic results with the historical diagnostic results in the historical associated data to obtain the diagnostic matching results; The first diagnostic unit is used to update the preliminary diagnostic result based on the maintenance record corresponding to the historical diagnostic result if the diagnostic matching result indicates that the preliminary diagnostic result is the same as the historical diagnostic result, so as to obtain the fault diagnosis result.

[0083] In one possible implementation, the preliminary diagnostic unit includes: Faulty component candidate sub-unit, used to identify candidate faulty components of the vehicle based on fault events; The component information acquisition subunit is used to acquire the dependencies and fault priorities among candidate faulty components. The first anomaly analysis subunit is used to analyze the anomaly data corresponding to the candidate faulty components according to the fault priority if there is no dependency between the candidate faulty components, so as to obtain the first diagnostic information of the candidate faulty components and the confidence level corresponding to the first diagnostic information. The first preliminary diagnosis subunit is used to identify candidate faulty components whose confidence level is higher than the first preset threshold corresponding to the first diagnosis information as faulty components, and to generate preliminary diagnosis results based on the faulty components and the corresponding first diagnosis information.

[0084] In one possible implementation, the preliminary diagnostic unit further includes: The associated feature extraction subunit is used to extract the temporal association features between the anomaly detection data of the first candidate faulty component and the anomaly detection data of the second candidate faulty component if there is a dependency between the candidate faulty components. The second anomaly analysis subunit is used to perform causal correlation analysis on the abnormal data of the first candidate faulty component and the second candidate faulty component based on the time-series correlation characteristics, so as to obtain the second diagnostic information and the confidence level corresponding to the second diagnostic information. The second preliminary diagnosis subunit is used to identify candidate faulty components whose confidence level is higher than the second preset threshold corresponding to the second diagnostic information as faulty components, and to generate preliminary diagnosis results based on the faulty components and the corresponding second diagnostic information.

[0085] In one possible implementation, the diagnostic module 303 further includes: The correlation analysis unit is used to extract the temporal features of the driving scenario data and perform causal correlation analysis with the anomaly detection data if the correlation analysis results indicate that the correlation between the anomaly detection data and the driving scenario data is greater than or equal to a threshold, and to obtain the third diagnostic information. The second diagnostic unit is used to adjust the confidence weight of the abnormal detection data based on the confidence of the third diagnostic information, and to perform fault diagnosis on the vehicle by combining historical correlation data, and generate fault diagnosis results.

[0086] The fault diagnosis device in this embodiment is presented in the form of a functional unit. Here, a unit refers to an ASIC circuit, a processor and memory that execute one or more software or fixed programs, and / or other devices that can provide the above functions.

[0087] Further functional descriptions of the above modules and units are the same as those in the corresponding embodiments described above, and will not be repeated here.

[0088] Please see Figure 4 , Figure 4 This is a schematic diagram of the structure of an electronic device provided in an optional embodiment of the present invention, such as... Figure 4 As shown, the electronic device includes one or more processors 10, memory 20, and interfaces for connecting the components, including high-speed interfaces and low-speed interfaces. The components communicate with each other via different buses and can be mounted on a common motherboard or otherwise as required. The processors can process instructions executed within the electronic device, including instructions stored in or on memory to display graphical information of a GUI on external input / output devices (such as display devices coupled to the interfaces). In some alternative implementations, multiple processors and / or multiple buses can be used with multiple memories and multiple memory modules, if desired. Similarly, multiple electronic devices can be connected, each providing some of the necessary operations (e.g., as a server array, a group of blade servers, or a multiprocessor system). Figure 4 Take a processor 10 as an example.

[0089] Processor 10 may be a central processing unit, a network processor, or a combination thereof. Processor 10 may further include a hardware chip. The hardware chip may be an application-specific integrated circuit (ASIC), a programmable logic device (PLD), or a combination thereof. The programmable logic device may be a complex programmable logic device (CAMP), a field-programmable gate array (FPGA), a general-purpose array logic (GPA), or any combination thereof.

[0090] The memory 20 stores instructions executable by at least one processor 10 to cause at least one processor 10 to perform the method shown in the above embodiments.

[0091] The memory 20 may include a program storage area and a data storage area. The program storage area may store the operating system and applications required for at least one function; the data storage area may store data created based on the use of the electronic device as displayed on a mini-program landing page. Furthermore, the memory 20 may include high-speed random access memory and may also include non-transitory memory, such as at least one disk storage device, flash memory device, or other non-transitory solid-state storage device. In some alternative embodiments, the memory 20 may optionally include memory remotely located relative to the processor 10, and these remote memories can be connected to the electronic device via a network. Examples of such networks include, but are not limited to, the Internet, intranets, local area networks, mobile communication networks, and combinations thereof.

[0092] The memory 20 may include volatile memory, such as random access memory; the memory may also include non-volatile memory, such as flash memory, hard disk or solid-state drive; the memory 20 may also include a combination of the above types of memory.

[0093] The electronic device also includes an input device 30 and an output device 40. The processor 10, memory 20, input device 30, and output device 40 can be connected via a bus or other means.

[0094] Input device 30 can receive input numerical or character information, and generate key signal inputs related to user settings and function control of the electronic device, such as a touch screen, keypad, mouse, trackpad, touchpad, joystick, one or more mouse buttons, trackball, joystick, etc. Output device 40 may include display devices, auxiliary lighting devices (e.g., LEDs), and haptic feedback devices (e.g., vibration motors). The aforementioned display devices include, but are not limited to, liquid crystal displays, light-emitting diodes, displays, and plasma displays. In some alternative embodiments, the display device may be a touch screen.

[0095] This invention also provides a computer-readable storage medium. The methods described above according to embodiments of the invention can be implemented in hardware or firmware, or implemented as computer code that can be recorded on a storage medium, or implemented as computer code downloaded via a network and originally stored on a remote storage medium or a non-transitory machine-readable storage medium and then stored on a local storage medium. Thus, the methods described herein can be processed by software stored on a storage medium using a general-purpose computer, a dedicated processor, or programmable or dedicated hardware. The storage medium can be a magnetic disk, optical disk, read-only memory, random access memory, flash memory, hard disk, or solid-state drive, etc.; further, the storage medium can also include combinations of the above types of memory. It is understood that computers, processors, microprocessor controllers, or programmable hardware include storage components capable of storing or receiving software or computer code, which, when accessed and executed by the computer, processor, or hardware, implements the methods shown in the above embodiments.

[0096] A portion of the embodiments of this application can be applied as a computer program product, such as computer program instructions. When executed by a computer, these instructions, through the operation of the computer, can invoke or provide the methods and / or technical solutions according to the present invention. Those skilled in the art will understand that the forms in which computer program instructions exist in a computer-readable medium include, but are not limited to, source files, executable files, installation package files, etc. Accordingly, the ways in which computer program instructions are executed by a computer include, but are not limited to: the computer directly executing the instructions, or the computer compiling the instructions and then executing the corresponding compiled program, or the computer reading and executing the instructions, or the computer reading and installing the instructions and then executing the corresponding installed program. Here, the computer-readable medium can be any available computer-readable storage medium or communication medium accessible to a computer.

[0097] Although embodiments of the invention have been described in conjunction with the accompanying drawings, those skilled in the art can make various modifications and variations without departing from the spirit and scope of the invention, and such modifications and variations all fall within the scope defined by the appended claims.

Claims

1. A failure diagnosis method characterized by comprising: The method comprises: After a vehicle triggers a fault event, abnormal detection data associated with the fault event and driving scene data of the vehicle during operation are acquired; Correlation between the abnormal detection data and the driving scene data is analyzed to obtain a correlation analysis result; Based on the correlation analysis result and historical correlation data, fault diagnosis is performed on the vehicle to obtain a fault diagnosis result corresponding to the fault event, wherein the historical correlation data is determined based on historical operation data of the vehicle and / or abnormal data of other vehicles when the fault event occurs.

2. The method of claim 1, wherein, The acquisition of the abnormal detection data associated with the fault event and the driving scene data of the vehicle during operation comprises: The vehicle type of the vehicle and the diagnosis scene for diagnosing the vehicle are acquired; Based on the diagnosis scene and the vehicle type, driving data and driving scene data of the vehicle within a target period are acquired; Based on the fault event, feature extraction is performed on the driving data to obtain abnormal detection data of the vehicle within the target period.

3. The method of claim 1, wherein, The analysis of the correlation between the abnormal detection data and the driving scene data to obtain a correlation analysis result comprises: Based on the driving scene data, spatio-temporal data and environmental data of the space-time where the vehicle is located are acquired; A first correlation coefficient between the abnormal detection data and the spatio-temporal data, and a second correlation coefficient between the abnormal detection data and the environmental data are calculated; Based on the first correlation coefficient and / or the second correlation coefficient, the correlation between the abnormal detection data and the driving scene data is analyzed to obtain the correlation analysis result.

4. The method of claim 1, wherein, The fault diagnosis based on the correlation analysis result and the historical correlation data to obtain the fault diagnosis result corresponding to the fault event comprises: If the correlation analysis result represents that the correlation between the abnormal detection data and the driving scene data is less than a threshold value, fault diagnosis is performed on the vehicle based on statistical features of the abnormal detection data to obtain a preliminary diagnosis result; The preliminary diagnosis result is matched with historical diagnosis results in the historical correlation data to obtain a diagnosis matching result; If the diagnosis matching result represents that the preliminary diagnosis result is the same as the historical diagnosis result, the preliminary diagnosis result is updated based on a maintenance record corresponding to the historical diagnosis result to obtain the fault diagnosis result.

5. The method of claim 4, wherein, The fault diagnosis based on the statistical features of the abnormal detection data to obtain a preliminary diagnosis result comprises: Based on the fault event, candidate fault components of the vehicle are determined; Dependency relationships and fault priorities between the candidate fault components are acquired; If there is no dependency relationship between the candidate fault components, the abnormal data corresponding to the candidate fault components is analyzed according to the fault priorities to obtain first diagnosis information of the candidate fault components and a confidence degree corresponding to the first diagnosis information; Candidate fault components with a confidence degree corresponding to the first diagnosis information higher than a first preset threshold value are determined as fault components, and the preliminary diagnosis result is generated based on the fault components and the corresponding first diagnosis information.

6. The method of claim 5, wherein, The preliminary diagnosis result is obtained by performing fault diagnosis on the vehicle based on the statistical features of the abnormal detection data. If there is a dependency relationship between the candidate fault components, a time sequence correlation feature of the abnormal detection data of the first candidate fault component and the abnormal detection data of the second candidate fault component is extracted. Causal correlation analysis is performed on the abnormal data of the first candidate fault component and the second candidate fault component based on the time sequence correlation feature, and second diagnosis information and a confidence degree corresponding to the second diagnosis information are obtained. If the confidence degree corresponding to the second diagnosis information is higher than a second preset threshold, the candidate fault component is determined as a fault component, and the preliminary diagnosis result is generated based on the fault component and the corresponding second diagnosis information.

7. The method of claim 1, wherein, The fault diagnosis result corresponding to the fault event is obtained by performing fault diagnosis on the vehicle based on the correlation analysis result and historical correlation data, and the fault diagnosis result further includes: If the correlation analysis result indicates that the correlation between the abnormal detection data and the driving scene data is greater than or equal to a threshold, a time sequence feature of the driving scene data is extracted and causal correlation analysis is performed on the abnormal detection data, and third diagnosis information is obtained. The confidence degree of the abnormal detection data is adjusted according to the confidence degree of the third diagnosis information, and fault diagnosis is performed on the vehicle in combination with the historical correlation data to generate the fault diagnosis result.

8. A failure diagnosing apparatus characterized by comprising: The device includes: An acquisition module is configured to acquire abnormal detection data and driving scene data associated with a fault event of a vehicle during operation of the vehicle after the vehicle triggers the fault event; An analysis module is configured to analyze the correlation between the abnormal detection data and the driving scene data to obtain a correlation analysis result; A diagnosis module is configured to perform fault diagnosis on the vehicle based on the correlation analysis result and historical correlation data, and generate a fault diagnosis result, wherein the historical correlation data is determined based on historical operation data of the vehicle and / or abnormal data of other vehicles when the fault event occurs.

9. An electronic device, comprising: The device includes: A memory and a processor are communicatively connected, and the memory stores computer instructions, and the processor executes the computer instructions to perform the method of any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that, The computer readable storage medium stores computer instructions for causing a computer to execute the method of any one of claims 1 to 7.

Images