Vehicle fault prediction method, device and computer equipment based on V2X

Abstract

The present invention discloses a vehicle fault prediction method, device and computer equipment based on V2X. The method comprises: obtaining data and environmental information from key parts of the vehicle collected by a vehicle communication bus based on V2X to obtain initial data; preprocessing the initial data to obtain a preprocessing result; inputting the preprocessing result into a vehicle fault prediction model to perform real-time fault warning, fault prediction and automatic driving fault prevention prediction to obtain a prediction result; and outputting the prediction result. By implementing the method of the present invention, the problems of insufficient real-time performance and accuracy in traditional fault prediction methods can be solved, and more intelligent and accurate fault prediction can be achieved.

Description

Technical Field

[0001] The present invention relates to the field of vehicle communication technology, and more specifically to a vehicle fault prediction method, device and computer equipment based on V2X. Background Art

[0003] At present, vehicle fault prediction mainly relies on static data such as basic parameters of the vehicle when it leaves the factory, regular inspections, and analysis based on historical records. Although this method can prevent the occurrence of faults to a certain extent, it has many limitations. First, traditional detection methods are difficult to provide instant data feedback, resulting in the inability to monitor and warn the vehicle status in real time. Due to the lack of effective real-time data support, traditional methods often cannot respond quickly when encountering emergencies or complex traffic environments. Secondly, relying on limited historical data and simple statistical analysis makes the prediction results inaccurate and easy to miss potential problem points. Moreover, the current passive maintenance mode means that action is only taken after a fault occurs, which not only increases maintenance costs, but may also affect driving safety.

[0004] With the development of intelligent transportation systems, vehicle communication technologies such as V2X have begun to be introduced into the field of fault prediction. Although V2X technology enhances traffic management and safety through information exchange between vehicles and between vehicles and infrastructure, it mainly focuses on improving traffic flow and driving experience, and its support for vehicle fault prediction is still insufficient. In addition, although the application of artificial intelligence and big data technologies has brought new perspectives to fault prediction, due to the limitations of the computing power and computing speed of on-board equipment, complex AI algorithms are difficult to deploy efficiently, limiting the actual application effect of these technologies.

[0005] Therefore, it is necessary to design a new method to solve the problems of insufficient real-time and accuracy in traditional fault prediction methods and to achieve more intelligent and accurate fault prediction. Summary of the invention

[0007] The purpose of the present invention is to overcome the defects of the prior art and provide a vehicle fault prediction method, device and computer equipment based on V2X.

[0008] To achieve the above object, the present invention adopts the following technical solution: a vehicle fault prediction method based on V2X, comprising: Acquire the V2X-based vehicle communication bus to collect data from key parts of the vehicle and environmental information to obtain initial data; Preprocessing the initial data to obtain a preprocessing result; Inputting the preprocessing results into a vehicle fault prediction model to perform real-time fault warning, fault prediction, and automatic driving fault prevention prediction to obtain a prediction result; The prediction result is output.

[0009] Its further technical solution is: the vehicle fault prediction model is obtained by preprocessing the key parts data of the vehicle and the environmental information obtained by the V2X-based vehicle communication bus and combining the real results as a sample set to train the basic model, wherein the basic model includes the LightGBM model.

[0010] A further technical solution is that the data of the key parts of the vehicle include speed, acceleration and engine status data of key components of the engine management system, gearbox and safety system.

[0011] A further technical solution is: the preprocessing of the initial data to obtain a preprocessing result includes: Performing data cleaning on the initial data to obtain a cleaning result; Performing feature engineering processing on the cleaning result to obtain a feature extraction result; Performing data fusion on the feature extraction results to obtain a fusion result; The fusion result is standardized to obtain a preprocessing result.

[0012] The further technical solution is: the vehicle fault prediction model is obtained by preprocessing the key parts data of the vehicle and the environmental information obtained by the vehicle communication bus based on V2X and combining the real results as the sample set to train the basic model, wherein the basic model includes the LightGBM model, including: The key vehicle parts data and environmental information obtained through the V2X-based vehicle communication bus are pre-processed and combined with the real results as a sample set; Build the basic model; Define the loss function and choose the optimization algorithm; Using the sample set, combined with the loss function and the optimization algorithm, the basic model is forward propagated, loss calculated, back propagated and weight updated, and the model parameters are continuously optimized until the best performance is achieved, so as to obtain a trained model; The K-fold cross-validation method is used to evaluate the model stability and generalization ability of the trained model to determine the vehicle fault prediction model.

[0013] Its further technical solution is: after outputting the prediction result, it also includes: Collect user feedback and actual fault conditions, and retrain the vehicle fault prediction model based on the collected data to update the vehicle fault prediction model.

[0014] A further technical solution is: the preprocessing result is input into a vehicle fault prediction model to perform real-time fault warning, fault prediction and automatic driving fault prevention prediction to obtain a prediction result, including: When real-time fault warning is performed, the preprocessing results are input into the vehicle fault prediction model, the preprocessing data is comprehensively analyzed, the potential fault risk of the vehicle is quickly identified, and an alarm is issued in time; When fault prediction is performed, the preprocessing results are input into the vehicle fault prediction model, and possible problems are foreseen through in-depth analysis of vehicle usage records, driving habits and environmental factors; When performing automatic driving fault prevention prediction, the preprocessing result is input into the vehicle fault prediction model to automatically adjust the vehicle speed, implement emergency braking or perform obstacle avoidance steering operations.

[0015] A further technical solution is: when performing automatic driving fault prevention prediction, the preprocessing result is input into the vehicle fault prediction model to automatically adjust the vehicle speed, implement emergency braking or perform obstacle avoidance steering operations, and also includes: The V2X network continuously monitors the vehicle's operating status and changes in the surrounding environment, and identifies factors that may threaten driving safety in advance.

[0016] The present invention also provides a vehicle fault prediction device based on V2X, comprising: A data acquisition unit, used to acquire data from key parts of the vehicle and environmental information collected by the V2X-based vehicle communication bus to obtain initial data; A preprocessing unit, used for preprocessing the initial data to obtain a preprocessing result; A prediction unit, used for inputting the preprocessing result into a vehicle fault prediction model to perform real-time fault warning, fault prediction and automatic driving fault prevention prediction to obtain a prediction result; An output unit is used to output the prediction result.

[0017] The present invention further provides a computer device, which includes a memory and a processor. The memory stores a computer program, and the processor implements the above method when executing the computer program.

[0018] The beneficial effects of the present invention compared with the prior art are as follows: the present invention, by adopting the vehicle communication bus technology based on V2X, can collect data and environmental information from key components of the vehicle in real time, ensure the timeliness and comprehensiveness of the data, and solve the problem of data collection delay in the traditional method. Then, these initial data are preprocessed, including steps such as cleaning, standardization and feature extraction to improve the data quality, thereby providing a reliable basis for subsequent analysis. Then, the preprocessed data is input into a specially designed vehicle fault prediction model, which uses advanced algorithms for real-time fault warning, fault prediction and fault prevention prediction for autonomous driving vehicles, greatly improving the accuracy and timeliness of the prediction. Since the model continuously learns and adapts to new data patterns, it can make more intelligent and accurate fault predictions, effectively making up for the shortcomings of traditional methods in dealing with complex situations. Finally, the prediction results are output in time to help the driver or the automatic system take action in advance to avoid safety hazards and economic losses caused by potential faults.

[0019] The present invention is further described below in conjunction with the accompanying drawings and specific embodiments. BRIEF DESCRIPTION OF THE DRAWINGS

[0021] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings required for use in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other accompanying drawings can be obtained based on these accompanying drawings without paying any creative work.

[0022] Figure 1 A schematic diagram of an application scenario of a V2X-based vehicle fault prediction method provided in an embodiment of the present invention; Figure 2 A schematic diagram of a flow chart of a vehicle fault prediction method based on V2X provided in an embodiment of the present invention; Figure 3 A schematic diagram of a sub-process of a vehicle fault prediction method based on V2X provided in an embodiment of the present invention; Figure 4 A schematic diagram of a process flow of a vehicle fault prediction model training process provided by an embodiment of the present invention; Figure 5 A schematic block diagram of a vehicle fault prediction device based on V2X provided in an embodiment of the present invention; Figure 6 A schematic block diagram of a preprocessing unit of a V2X-based vehicle fault prediction device provided in an embodiment of the present invention; Figure 7 A schematic block diagram of a computer device provided in an embodiment of the present invention. DETAILED DESCRIPTION

[0024] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0025] It should be understood that when used in this specification and the appended claims, the terms "include" and "comprises" indicate the presence of described features, integers, steps, operations, elements and / or components, but do not exclude the presence or addition of one or more other features, integers, steps, operations, elements, components and / or combinations thereof.

[0026] It should also be understood that the terms used in this specification of the present invention are only for the purpose of describing specific embodiments and are not intended to limit the present invention. As used in the specification of the present invention and the appended claims, unless the context clearly indicates otherwise, the singular forms "a", "an" and "the" are intended to include plural forms.

[0027] It should be further understood that the term "and / or" used in the present description and the appended claims refers to and includes any and all possible combinations of one or more of the associated listed items.

[0028] See also Figure 1 and Figure 2 , Figure 1 A schematic diagram of an application scenario of a V2X-based vehicle fault prediction method provided in an embodiment of the present invention. Figure 2A schematic flow chart of a vehicle fault prediction method based on V2X provided in an embodiment of the present invention. The vehicle fault prediction method based on V2X is applied to a system with integrated vehicle-mounted equipment, roadside equipment and intelligent transportation system, and uses V2X technology to realize seamless information exchange between vehicles and vehicles (V2V), vehicles and infrastructure (V2I), vehicles and pedestrians (V2P) and vehicles and networks (V2N), ensuring real-time update and transmission of vehicle status data. This not only helps to accurately grasp the current road conditions, but also can identify and warn potential fault risks in advance. With the help of powerful cloud computing resources and advanced machine learning algorithms such as support vector machines SVM, convolutional neural networks CNN, long short-term memory networks LSTM, etc., the collected large-scale multi-source heterogeneous data including but not limited to sensor data, driving behavior data, environmental factors, etc. are deeply analyzed, so as to more accurately predict the possibility of fault occurrence and its development trend. Taking into account the differences between different models and technical levels, the system is flexible in design and can adapt to various types of vehicles, and the prediction model can be continuously optimized and upgraded through OTA (Over-The-Air) technology to maintain its advancement and adaptability. Through centralized cloud management, the system supports automated model training and update mechanisms, ensuring that as new data accumulates and technology advances, the prediction model is always kept up to date, thereby improving prediction performance. Through continuous monitoring and analysis of vehicle operation data, car owners and managers can obtain more detailed vehicle health reports, plan maintenance work in advance, reduce the incidence of unexpected failures, reduce maintenance costs, and extend the service life of vehicles.

[0029] In summary, the method of this embodiment not only solves the problems of insufficient real-time and accuracy in traditional fault prediction methods, but also realizes more intelligent and accurate fault prediction by introducing advanced AI technology and cloud computing platform, providing a solid guarantee for the safe driving of vehicles. In addition, this innovative solution will also help promote the further development of intelligent transportation systems and lay the foundation for future smart travel.

[0030] In this embodiment, the vehicle-mounted device is the basis of the entire fault prediction system. It can monitor and obtain key operating data from various sensors in real time by connecting to the vehicle's communication bus. These data include but are not limited to engine temperature, oil pressure, vehicle speed, and tire status, providing a detailed information basis for subsequent analysis. The vehicle data detection and optimization module is responsible for optimizing and screening these real-time collected data to ensure that the data fed into the fault prediction model has high precision, thereby improving the accuracy of the analysis.

[0031] The driving decision module uses optimized data and information fed back by the fault prediction model to provide safe driving suggestions to the driver. These suggestions may involve reminding the driver of potential safety risks, or executing more accurate driving decisions through the system in autonomous driving mode to reduce the risks caused by human error. In addition, V2X on-board equipment also has the ability to exchange information with other vehicles, road infrastructure and pedestrians, further enriching the environmental data required for fault prediction and improving the comprehensiveness and accuracy of the prediction.

[0032] As part of the ground support, roadside equipment works closely with vehicle-mounted equipment to focus on collecting information about road conditions and traffic dynamics. Through smart sensors and cameras deployed on the road, roadside equipment can transmit key information such as road conditions, traffic density, and traffic accidents to the fault prediction model in real time. For example, when encountering traffic jams or traffic accidents, the system can issue early warnings to vehicles to reduce the possibility of failures, while also providing an important basis for traffic management and emergency response.

[0033] The fault prediction model is deployed in the network platform, whether it is located on the roadside equipment or the server of the intelligent transportation system. The model combines the data collected by the on-board equipment and the roadside equipment, and uses advanced algorithms for real-time analysis to predict the possible failure of the vehicle and notify the driver or the management center in time. This process not only helps to reduce the incidence of sudden failures, but also optimizes the vehicle maintenance plan and extends its service life through the analysis of failure data.

[0034] The intelligent transportation system plays a core role in coordinating resources from all parties, and mainly undertakes three key functions: first, traffic flow monitoring and management, that is, real-time monitoring of traffic conditions and reasonable scheduling based on prediction results; second, the coordination work of the management center to ensure the efficient operation of the entire system; third, real-time monitoring and response to ensure that the system can quickly adapt to changes and make timely adjustments. By integrating on-board equipment, roadside equipment and intelligent transportation systems, real-time monitoring and prediction of vehicle failures are achieved, which greatly improves driving safety and optimizes traffic management efficiency.

[0035] In summary, the method of this embodiment not only enhances the safety of the vehicle itself, but also contributes to the improvement and development of the intelligent transportation system through the fusion of multi-source data and the support of advanced algorithms.

[0036] The method of this embodiment gives full play to the advantages of the collaborative work of the end (vehicle equipment) and the cloud (cloud computing platform), which can not only predict vehicle faults in real time, but also dynamically optimize and upgrade the prediction model. The system has significant advantages such as strong real-time performance, high prediction accuracy, wide applicability, and continuous iterative upgrades, providing strong technical support for vehicle management and traffic safety.

[0037] Figure 2 FIG. 1 is a flow chart of a vehicle fault prediction method based on V2X provided in an embodiment of the present invention. Figure 2 As shown, the method includes the following steps S110 to S140.

[0038] S110: Acquire the V2X-based vehicle communication bus to collect data from key parts of the vehicle and environmental information to obtain initial data.

[0039] In this embodiment, the initial data includes data of key parts of the vehicle and environmental information. The data of key parts of the vehicle include speed, acceleration, and engine status data of key parts of the engine management system, gearbox, and safety system. The collection of initial data is completed through the V2X (Vehicle-to-Everything) on-board equipment integrated in the vehicle. This process aims to obtain comprehensive and real-time data streams from key parts of the vehicle and the surrounding environment, providing a solid foundation for subsequent analysis and processing.

[0040] V2X vehicle-mounted devices access and collect operating data from different key systems by connecting to various communication bus interfaces inside the vehicle, such as CAN (Controller Area Network) bus, LIN (Local Interconnect Network) bus, FlexRay bus, MOST (Media Oriented Systems Transport) bus, and vehicle Ethernet. These systems cover: Engine management system: responsible for controlling the engine's operating parameters, such as speed, temperature, throttle position, etc.

[0041] Transmission: monitors shift timing, gear ratio and other information.

[0042] Airbag control system: records the occurrence and response of collision events.

[0043] ABS (Anti-lock Braking System): monitors brake status and pressure.

[0044] EPS (Electric Power Steering) electronic power steering system: tracks steering wheel angle and steering torque.

[0045] ESC (Electronic Stability Control) Electronic Stability Control System: manages vehicle stability related parameters.

[0046] In addition, it also involves important information such as vehicle speed, acceleration, steering wheel angle, engine speed, throttle position, brake status and brake pressure. These data are crucial for understanding the vehicle's current operating mode and driving behavior.

[0047] In addition to in-vehicle data, V2X onboard equipment also needs to interact with other vehicles (V2V), road infrastructure (V2I) and pedestrians (V2P) to obtain detailed information about the external environment. This includes but is not limited to the location, speed, and driving direction of other vehicles, as well as the status of traffic lights, road conditions, weather conditions, etc. This cross-entity information exchange helps to build a complete perception picture of the vehicle and its surrounding environment, thereby supporting more accurate fault prediction and auxiliary decision-making.

[0048] In order to ensure the quality of the collected data, a series of preprocessing steps are performed before the data is transmitted to the central processing unit or cloud platform. For example, missing values ​​and outliers are cleaned and data normalization or standardization is performed to ensure the consistency and accuracy of the data. These measures not only improve the reliability of the data, but also lay a good foundation for subsequent data fusion and advanced analysis.

[0049] In summary, in the S110 stage, through the collaborative work of V2X on-board equipment, the vehicle's internal communication bus and other external entities, extensive and in-depth collection of vehicle key parts data and environmental information is achieved, forming an initial data set for further analysis and application. This data set will serve as an important input source for data fusion, feature extraction, model training and other operations in subsequent steps.

[0050] S120: preprocess the initial data to obtain a preprocessing result.

[0051] In this embodiment, the preprocessing result refers to the result obtained after comprehensive data cleaning, feature engineering, data fusion and standardization of the initial data collected from the vehicle communication bus and environmental information source. This process ensures the quality and applicability of the data and provides a solid foundation for subsequent analysis and model training.

[0052] In one embodiment, see Figure 3 , the above-mentioned step S120 may include steps S121~S124.

[0053] S121, performing data cleaning on the initial data to obtain a cleaning result.

[0054] In this embodiment, data cleaning is the first step to ensure data quality, which involves removing or correcting noise, errors and inconsistencies in the original data. Specifically: For missing data points, they can be completed by deleting related records, filling the mean / median / mode, using interpolation or other prediction methods.

[0055] Identify and process data points that are outside the normal range through statistical methods (such as the 3σ principle), box plot analysis, machine learning algorithms, etc.

[0056] Make sure all data follows consistent timestamp format, units, etc. to facilitate subsequent processing.

[0057] S122: Perform feature engineering processing on the cleaning result to obtain a feature extraction result.

[0058] In this embodiment, feature engineering aims to extract key features from the cleaned data that help improve model performance. This step includes but is not limited to: Based on domain knowledge and technical means, we select the parameters that best represent the vehicle status and driving behavior, such as vehicle speed, engine speed, temperature, current, battery level, tire pressure, vibration data, etc.

[0059] Create new features based on existing features, such as calculating acceleration, deceleration rate, fuel efficiency and other indicators; you can also consider time series features, such as historical averages, trend changes, etc.

[0060] Methods such as principal component analysis (PCA) and factor analysis are applied to reduce redundant features and simplify the model input while retaining as much information as possible.

[0061] S123: performing data fusion on the feature extraction results to obtain a fusion result.

[0062] In this embodiment, data fusion combines data from different sources to form a more complete and higher-dimensional data set. This process can: Combine the external environment information collected by V2X devices (such as the location, speed, driving direction of other vehicles, the status of traffic lights, road conditions, weather conditions, etc.) with the operating data provided by the vehicle's internal system. If the vehicle is equipped with multiple types of sensors (cameras, radars, lidars, etc.), these different types of data sources need to be synchronized and fused to provide richer perception capabilities.

[0063] S124: Standardize the fusion result to obtain a preprocessing result.

[0064] In this embodiment, the purpose of standardization is to eliminate the dimensional differences between different features so that they can participate in modeling on a relatively fair basis. Commonly used methods include: Normalization: Scaling the values ​​to the range [0, 1] is applicable to situations where certain algorithms require the data to be non-negative.

[0065] Standardization: usually refers to Z-score standardization, which converts the data into a standard normal distribution with a mean of 0 and a standard deviation of 1, which is suitable for most machine learning algorithms.

[0066] Encoding categorical variables: For non-numeric data (such as color, brand, etc.), use one-hot encoding or label encoding to convert it into a form that the machine learning model can understand.

[0067] In summary, in step S120, by performing data cleaning, feature engineering, data fusion and standardization processing on the initial data in sequence, a high-quality, structured preprocessing result is finally obtained, which can be directly used for subsequent fault prediction, real-time decision support and other advanced analysis tasks.

[0068] In this embodiment, in order to improve the accuracy of vehicle fault prediction and real-time decision-making, all collected data must be processed by data fusion technology. Commonly used data fusion algorithms include Kalman filters, etc. These methods can effectively improve the accuracy and integrity of data. In addition, it is also important to flexibly set the data collection trigger mechanism according to different driving conditions. For example, when abnormal speed or emergency braking is detected, the system should be able to automatically speed up the frequency of data collection in order to capture key information more timely.

[0069] After data collection is completed, a series of preprocessing steps must be performed on the data to ensure that it is suitable for subsequent analysis and model training. These steps include: Data cleaning: used to process missing values, outliers and noise data reported by on-board equipment to improve data quality.

[0070] Feature selection: Identify the most valuable parameters for fault prediction, such as vehicle operating status (vehicle speed, engine speed, temperature, current, power, tire pressure, vibration data, lateral acceleration, roll angle, etc.), environmental factors (temperature, humidity, air pressure, road conditions, traffic flow, etc.) and other relevant features (historical maintenance records, driving behavior, vehicle age, etc.). These features provide important information for predicting potential faults.

[0071] Data encoding: For categorical variables, one-hot encoding or label encoding is usually used for conversion.

[0072] Data normalization / standardization: adjust the scale of different features, eliminate dimensional differences, and ensure that each feature has a balanced impact in model training.

[0073] Finally, the processed and integrated data will be generated into messages in the V2X communication protocol format and reported by the V2X on-board equipment to the V2X roadside equipment or the remote network platform of the intelligent transportation system for tasks such as real-time monitoring, fault prediction and decision support.

[0074] S130: Input the preprocessing result into a vehicle fault prediction model to perform real-time fault warning, fault prediction, and automatic driving fault prevention prediction to obtain a prediction result.

[0075] In this embodiment, the vehicle fault prediction model is obtained by preprocessing the key parts data of the vehicle and the environmental information obtained by the V2X-based vehicle communication bus and combining the real results as the sample set to train the basic model, wherein the basic model includes the LightGBM model.

[0076] The vehicle fault prediction model obtains data from key parts of the vehicle and environmental information, combines the real results of these data as a sample set for preprocessing, and then uses them to train the basic model. Through a series of optimization and evaluation steps, an efficient, accurate and highly generalized fault prediction model is finally obtained.

[0077] Specifically, see Figure 4 The vehicle fault prediction model is obtained by preprocessing the key parts data of the vehicle and the environmental information obtained by the vehicle communication bus based on V2X and combining the real results as the sample set to train the basic model, and may include steps S131~S135.

[0078] S131. Preprocess the key vehicle part data and environmental information obtained through the V2X-based vehicle communication bus and combine them with the actual results as a sample set.

[0079] In this embodiment, this step includes collecting key sensor data from the vehicle communication bus (such as the CAN bus), such as engine status, vehicle speed, acceleration, etc., while integrating external environmental information such as weather conditions, road conditions, and the location and speed of other vehicles. These raw data are processed through the aforementioned preprocessing steps (such as data cleaning, feature engineering, data fusion, and standardization) and combined with the actual fault records or maintenance reports at the corresponding timestamps to form a labeled training sample set. This step ensures that the basic model used for training can learn the mode transition from normal operation to the occurrence of a fault.

[0080] S132. Construct a basic model.

[0081] In this embodiment, a machine learning or deep learning algorithm suitable for vehicle fault prediction tasks is selected to build a basic model. Considering the time series characteristics and real-time requirements, gradient boosting decision tree algorithms such as LightGBM may be one of the preferred ones. In addition, according to specific needs, other types of models such as recurrent neural networks (RNNs), long short-term memory networks (LSTMs), or convolutional neural networks (CNNs) can also be considered to capture feature associations in different dimensions.

[0082] S133. Define the loss function and select an optimization algorithm.

[0083] In this embodiment, in order to adapt to the mixed task characteristics of classification and regression, a composite loss function can be used, that is, the classification task uses the cross entropy loss function, and the mean square error (MSE) is used for the continuous value output (such as the probability of failure). This is done to more accurately reflect the difference between the model prediction and the actual situation.

[0084] Choose an appropriate optimizer, such as the Adam optimizer, which is able to adaptively adjust the learning rate in different parameter spaces, accelerate the convergence process and improve the final performance.

[0085] S134, using the sample set, combining the loss function and the optimization algorithm to perform forward propagation, loss calculation, back propagation and weight update on the basic model, and continuously optimizing the model parameters until the best performance is achieved to obtain a trained model.

[0086] In this example, at this stage, the model begins to learn how to predict possible failures based on the input data. Specifically: Forward propagation: The preprocessed data is input into the model and calculated layer by layer until the predicted output is generated.

[0087] Loss calculation: Compare the predicted output with the true label and use the loss function to quantify the difference between the two.

[0088] Backpropagation: Based on the loss value, apply the chain rule to calculate the gradient of each layer's parameters to determine the impact of each weight on the total loss.

[0089] Weight update: According to the selected optimization algorithm, the weights in the model are adjusted based on the calculated gradient to minimize the loss. This process is repeated until the preset stopping condition is met (such as the maximum number of iterations, loss threshold, or performance on the validation set is no longer improved).

[0090] S135. Use a K-fold cross-validation method to evaluate the model stability and generalization ability of the trained model to determine a vehicle fault prediction model.

[0091] In this embodiment, in order to ensure that the model not only performs well on the training data, but also has good generalization ability on unseen data, a K-fold cross-validation method is used. This means that the original sample set is divided into K subsets, which are used as validation sets in turn, and the rest are used as training sets. By averaging the results of all K runs, more reliable and stable evaluation indicators such as accuracy, recall, F1 score, and ROC-AUC curve can be obtained, thereby confirming the effectiveness and stability of the selected model.

[0092] Specifically, user feedback and actual fault conditions are collected, and the vehicle fault prediction model is retrained based on the collected data to update the vehicle fault prediction model.

[0093] In this embodiment, once the model is deployed in the actual environment, the system will continue to collect new data points, including but not limited to user feedback information and actual failure cases. These new data are critical to keeping the model up to date, as they can help the model adapt to changing operating conditions and technological advances. Therefore, it is necessary to regularly retrain the existing model with the latest data, so as to continuously improve the model's prediction accuracy and service quality and ensure its long-term effectiveness.

[0094] In summary, the entire training process is a closed-loop system. From data collection, preprocessing, model building, training optimization to evaluation deployment, and then to subsequent continuous learning, each link is closely linked and works together to improve the overall performance and reliability of the vehicle fault prediction system.

[0095] For the vehicle fault prediction model based on V2X technology, the algorithm accuracy, real-time performance, scalability, and resource limitations (memory and computing power) of the vehicle and roadside equipment were comprehensively considered, and the ensemble learning algorithm LightGBM was finally selected. This algorithm improves the prediction performance by combining multiple models, mainly by building a series of decision trees to achieve the prediction goal. Incremental training is used in the training process. In each round of iteration, the tree structure is gradually optimized by calculating the residual (that is, the difference between the predicted value and the true value), and a deeper tree is built according to the leaf growth strategy to enhance the prediction accuracy, and the loss function is minimized for optimization.

[0096] LightGBM performs well in data processing. It uses histogram algorithms and data bucketing to reduce memory consumption, and is particularly suitable for large-scale real-time data processing. By sorting the mean of each category, the continuous features are discretized into multiple discrete features (for example, k discrete features), and a histogram with a width of k is constructed for statistics. This allows the algorithm to quickly locate the best split point by scanning k bins without traversing the entire data set. This process not only improves the processing speed, but also improves the cache hit rate through cache optimization.

[0097] In feature processing, LightGBM calculates gradients in the same way for all features, achieving continuous access and further improving cache hit rate. Since there is no need to store the index of features and samples, storage consumption is greatly reduced, while avoiding the problem of Cache Miss. This design enables LightGBM to make full use of limited hardware resources while maintaining high efficiency, providing a powerful solution for vehicle fault prediction based on V2X technology.

[0098] In summary, through sophisticated data processing procedures and efficient model selection, the vehicle fault prediction system can achieve efficient and accurate fault prediction in practical applications, fully meeting business needs.

[0099] The purpose of the above training process is to continuously optimize the model parameters by analyzing historical data to ensure that the model can make accurate predictions when facing unknown data. In vehicle fault prediction based on V2X communication technology, this training process is particularly critical because it is directly related to the accuracy of fault prediction and the real-time response capability of the system. The entire training process can be divided into two main parts: training configuration and model training.

[0100] The training configuration phase is the starting point of the training process. The training configuration involves selecting an appropriate loss function to guide the learning direction of the model. The loss function is used to quantify the gap between the model's predicted results and the actual results, and is the basis for optimizing model performance. For vehicle fault prediction tasks, classification loss functions (such as cross entropy) are usually used in combination to predict the type of fault, and regression loss functions (such as mean square error MSE) are used to estimate continuous variables, such as the probability of a fault. In addition, an optimization algorithm must be selected to minimize the loss function value; common choices include gradient descent and its improved version Adam, which has shown higher efficiency and better convergence in practice due to its adaptive learning rate adjustment mechanism.

[0101] Learning rate: determines the pace of each weight update. A too high learning rate may cause the model to be unable to converge stably or even diverge, while a too low learning rate will make the convergence speed very slow. Therefore, it is usually necessary to find a suitable learning rate through experiments or learning rate scheduling strategies.

[0102] Batch size: refers to the amount of data passed to the model at a time. Smaller batches help update weights more frequently but may cause training instability; larger batches provide more stable gradient estimates but also increase training time. Therefore, in the vehicle fault prediction scenario, a compromise batch size is generally selected to balance computational efficiency and model performance.

[0103] The training phase includes the following core steps: Forward propagation: In this process, the input data is passed layer by layer to finally generate the predicted output. The weights and activation functions between the layers together determine the learning ability of the model.

[0104] Loss calculation: Compare the prediction results obtained by forward propagation with the true labels, and calculate the prediction error through the loss function to evaluate the current performance of the model.

[0105] Back propagation: Use the chain rule to efficiently calculate the contribution of each parameter to the total loss, and adjust the weights accordingly to push the model in a better direction.

[0106] Weight update: Update the weights in the model based on the calculated gradient information and the selected optimization algorithm. This process is repeated until the preset conditions are met, such as the loss reaches a minimum point or the performance on the validation set no longer improves.

[0107] Hyperparameter tuning: In order to further improve the performance of the model, some key hyperparameters (such as learning rate, batch size, number of network layers, etc.) are fine-tuned. Common methods include grid search, random search, and Bayesian optimization. Bayesian optimization significantly improves search efficiency by building a probabilistic model to predict the optimal combination.

[0108] In order to ensure efficient training and optimization, an early stopping mechanism is used to prevent overfitting, and regularization and data enhancement techniques are used to improve the generalization ability of the model. Through these steps, the model gradually approaches the optimal solution.

[0109] After training, a comprehensive validation is required to ensure the effectiveness and reliability of the model in practical applications. First, the K-fold cross-validation method is used to evaluate the stability and generalization ability of the model. This method obtains more reliable and stable evaluation indicators by dividing the dataset into several subsets, which are used as validation sets in turn, and the rest as training sets. Subsequently, the cross-validated model is deployed on an independent test set to further verify its prediction accuracy. In this process, performance indicators such as accuracy, recall, and F1 score are calculated to fully understand the overall performance of the model. High recall and high accuracy indicate that the model is both sensitive and reliable in fault prediction.

[0110] Based on the results of verification and testing, the best model can be selected for practical application, and detailed adjustments and optimizations can be made to the problems found, such as modifying hyperparameters, improving data processing procedures, or introducing new algorithms to deal with specific patterns. This not only improves the prediction accuracy and robustness of the model, but also ensures that it can maintain a high level of performance in various complex and dynamic environments.

[0111] Considering the need for real-time monitoring and prediction, the choice of hardware platform is crucial. Edge computing devices are very suitable for data processing and prediction tasks within the vehicle due to their low latency and real-time processing advantages; cloud servers are suitable for large-scale data processing and remote model updates with their powerful computing power. After deployment, the system can receive and process vehicle status data in real time, quickly identify potential faults and generate warnings, notify drivers or management systems to take corresponding measures, ensure driving safety, reduce maintenance costs, and improve operating efficiency.

[0112] As time goes by, new data is constantly generated, and the model's predictive ability and adaptability may change, so it is necessary to retrain the model regularly with the latest data. To this end, a complete monitoring system must be established to track the model's prediction effect and vehicle status in real time, and promptly detect and correct any deviations or abnormalities. At the same time, user feedback and actual fault cases are collected as an important basis for subsequent optimization. By analyzing this information, the deficiencies of the model can be identified and the direction of future improvements can be guided. Regularly updating and retraining the model can not only improve its ability to identify new patterns, but also ensure that it always adapts to the changing vehicle status and environmental conditions. Through such a continuous learning mechanism, the vehicle fault prediction system can be continuously optimized to provide users with more accurate and timely fault warning services, ultimately improving overall safety and operating efficiency.

[0113] The trained model is applied in actual scenarios. When real-time fault warning is performed, the preprocessing results are input into the vehicle fault prediction model, the preprocessing data is comprehensively analyzed, the potential fault risk of the vehicle is quickly identified, and an alarm is issued in time; When fault prediction is performed, the preprocessing results are input into the vehicle fault prediction model, and possible problems are foreseen through in-depth analysis of vehicle usage records, driving habits and environmental factors. The vehicle operating status and surrounding environmental changes are continuously monitored through the V2X network to identify factors that may threaten driving safety in advance.

[0114] When performing automatic driving fault prevention prediction, the preprocessing result is input into the vehicle fault prediction model to automatically adjust the vehicle speed, implement emergency braking or perform obstacle avoidance steering operations.

[0115] Specifically, the vehicle fault prediction model can provide real-time fault warnings in actual operation. The system can identify potential fault risks and issue warnings to the driver in advance by continuously collecting and analyzing the vehicle's operating status data, including but not limited to engine temperature, tire pressure, brake system performance, throttle response, steering wheel operation, and environmental information such as traffic flow, weather conditions, and road conditions. For example, when abnormal brake system or engine temperature is detected, the model can quickly identify abnormal patterns, evaluate possible fault types and their severity, and promptly notify the driver to take maintenance or repair measures, thereby preventing traffic accidents.

[0116] In addition to real-time warnings, the model also has the ability to predict faults and make maintenance recommendations. By learning from historical data, it can accurately predict possible problems with the vehicle and provide recommendations for optimizing vehicle performance, such as adjusting engine parameters to improve fuel efficiency. By analyzing vehicle usage history, driving behavior characteristics, and environmental conditions, the model can predict possible problems with the engine or battery and give specific maintenance recommendations. This allows vehicle managers to develop reasonable maintenance plans based on the probability and impact of failures, reducing the high maintenance costs caused by sudden failures.

[0117] For autonomous vehicles, the role of fault prediction models is particularly critical. With the help of V2X communication, the autonomous driving system can obtain data on the vehicle status and surrounding environment in real time, conduct comprehensive analysis, and detect any faults that may affect driving safety in advance. After detecting potential problems, the system can automatically take measures such as slowing down, stopping, or changing the driving route to avoid accidents. This continuous fault detection and prediction mechanism ensures that autonomous vehicles maintain efficient and safe operation in complex and changing traffic environments.

[0118] In summary, the vehicle fault prediction method based on V2X not only ensures the real-time and rapid response capabilities of the data, but also improves the accuracy of fault prediction by integrating multiple sensor data. At the same time, the model has good scalability and can be easily integrated into the existing intelligent transportation system to provide fault prediction services for a wider range of vehicles. This technical solution can use a variety of advanced technologies such as industrial Internet, Internet of Things, edge computing, artificial intelligence, big data and cloud computing to perform intelligent health management of vehicles and realize the closed loop of equipment fault prediction and full life cycle management. This not only reduces the high maintenance costs caused by sudden failures, but also avoids the inconvenience caused to users by vehicle breakdowns, significantly improves the operational efficiency and economic benefits of manufacturing companies, and helps to create a safer, more efficient and environmentally friendly intelligent transportation system.

[0119] S140: Output the prediction result.

[0120] In this embodiment, the prediction result is output to the terminal for display.

[0121] The above-mentioned V2X-based vehicle fault prediction method, by adopting V2X-based vehicle communication bus technology, can collect data and environmental information from key components of the vehicle in real time, ensure the timeliness and comprehensiveness of the data, and solve the problem of data collection delay in traditional methods. Then, these initial data are preprocessed, including cleaning, standardization and feature extraction steps to improve data quality, thereby providing a reliable basis for subsequent analysis. Then, the preprocessed data is input into a specially designed vehicle fault prediction model. The model uses advanced algorithms for real-time fault warning, fault prediction and fault prevention prediction for autonomous driving vehicles, which greatly improves the accuracy and timeliness of the prediction. Since the model continuously learns and adapts to new data patterns, it can make more intelligent and accurate fault predictions, effectively making up for the shortcomings of traditional methods in dealing with complex situations. Finally, the prediction results are output in a timely manner to help drivers or automatic systems take action in advance to avoid safety hazards and economic losses caused by potential faults.

[0122] Figure 5 is a schematic block diagram of a vehicle fault prediction device 300 based on V2X provided in an embodiment of the present invention. Figure 5 As shown, corresponding to the above V2X-based vehicle fault prediction method, the present invention also provides a V2X-based vehicle fault prediction device 300. The V2X-based vehicle fault prediction device includes a unit for executing the above V2X-based vehicle fault prediction method, and the device can be configured in a server. Figure 5 The V2X-based vehicle fault prediction device includes a data acquisition unit 301, a preprocessing unit 302, a prediction unit 303 and an output unit 304.

[0123] The data acquisition unit 301 is used to acquire data and environmental information from key parts of the vehicle collected by the V2X-based vehicle communication bus to obtain initial data; the preprocessing unit 302 is used to preprocess the initial data to obtain a preprocessing result; the prediction unit 303 is used to input the preprocessing result into a vehicle fault prediction model for real-time fault warning, fault prediction and automatic driving fault prevention prediction to obtain a prediction result; the output unit 304 is used to output the prediction result.

[0124] In one embodiment, if Figure 6 As shown, the preprocessing unit 302 includes a cleaning subunit 3021 , a feature engineering subunit 3022 , a data fusion subunit 3023 and a standardization subunit 3024 .

[0125] The cleaning subunit 3021 is used to perform data cleaning on the initial data to obtain a cleaning result; the feature engineering subunit 3022 is used to perform feature engineering processing on the cleaning result to obtain a feature extraction result; the data fusion subunit 3023 is used to perform data fusion on the feature extraction result to obtain a fusion result; the standardization subunit 3024 is used to perform standardization processing on the fusion result to obtain a preprocessing result.

[0126] In one embodiment, the device also includes: a training unit, which is used to pre-process the vehicle key part data and environmental information obtained through the V2X-based vehicle communication bus and combine them with real results as a sample set to train a basic model to obtain a vehicle fault prediction model, wherein the basic model includes a LightGBM model.

[0127] In one embodiment, the training unit comprises: A sample set determination subunit is used to pre-process the vehicle key part data and environmental information obtained through the V2X-based vehicle communication bus and combine the real results as a sample set; A model building subunit, used to build a basic model; Define subunits to define loss functions and select optimization algorithms; A model training subunit, for performing forward propagation, loss calculation, back propagation and weight update on the basic model by using the sample set, combining the loss function and the optimization algorithm, and continuously optimizing the model parameters until the best performance is achieved, so as to obtain a trained model; The evaluation subunit is used to evaluate the model stability and generalization ability of the trained model using a K-fold cross-validation method to determine a vehicle fault prediction model.

[0128] The updating subunit is used to collect user feedback and actual fault conditions, and retrain the vehicle fault prediction model based on the collected data to update the vehicle fault prediction model.

[0129] In one embodiment, the prediction unit 303 is used to: When real-time fault warning is performed, the preprocessing results are input into the vehicle fault prediction model, the preprocessing data is comprehensively analyzed, the potential fault risk of the vehicle is quickly identified, and an alarm is issued in time; When predicting a fault, the preprocessing result is input into the vehicle fault prediction model, and possible problems are foreseen through in-depth analysis of vehicle usage records, driving habits and environmental factors; when predicting automatic driving fault prevention, the preprocessing result is input into the vehicle fault prediction model to automatically adjust the vehicle speed, implement emergency braking or perform obstacle avoidance steering operations. The vehicle operating status and surrounding environmental changes are continuously monitored through the V2X network to identify factors that may threaten driving safety in advance.

[0130] It should be noted that those skilled in the art can clearly understand that the specific implementation process of the above-mentioned V2X-based vehicle fault prediction device and each unit can refer to the corresponding description in the aforementioned method embodiment, and for the convenience and conciseness of the description, it will not be repeated here.

[0131] The above-mentioned vehicle fault prediction device based on V2X can be implemented in the form of a computer program. The computer program can be used in Figure 7 Runs on the computer device shown.

[0132] See also Figure 7 , Figure 7 5 is a schematic block diagram of a computer device provided in an embodiment of the present application. The computer device 500 may be a server, wherein the server may be an independent server or a server cluster composed of multiple servers.

[0133] See also Figure 7 The computer device 500 includes a processor 502 , a memory and a network interface 505 connected via a system bus 501 , wherein the memory may include a non-volatile storage medium 503 and an internal memory 504 .

[0134] The non-volatile storage medium 503 may store an operating system 5031 and a computer program 5032. The computer program 5032 includes program instructions, which, when executed, may enable the processor 502 to execute a vehicle fault prediction method based on V2X.

[0135] The processor 502 is used to provide computing and control capabilities to support the operation of the entire computer device 500 .

[0136] The internal memory 504 provides an environment for the operation of the computer program 5032 in the non-volatile storage medium 503. When the computer program 5032 is executed by the processor 502, the processor 502 can execute a vehicle fault prediction method based on V2X.

[0137] The network interface 505 is used to communicate with other devices over the network. Figure 7The structure shown in the figure is merely a block diagram of a portion of the structure related to the solution of the present application, and does not constitute a limitation on the computer device 500 to which the solution of the present application is applied. The specific computer device 500 may include more or fewer components than those shown in the figure, or combine certain components, or have a different arrangement of components.

[0138] The processor 502 is used to run the computer program 5032 stored in the memory to implement the following steps: Acquire a V2X-based vehicle communication bus to collect data from key parts of the vehicle and environmental information to obtain initial data; preprocess the initial data to obtain a preprocessing result; input the preprocessing result into a vehicle fault prediction model to perform real-time fault warning, fault prediction, and automatic driving fault prevention prediction to obtain a prediction result; and output the prediction result.

[0139] The vehicle fault prediction model is obtained by preprocessing the key vehicle parts data and environmental information obtained by the V2X-based vehicle communication bus and combining the real results as a sample set to train the basic model, wherein the basic model includes the LightGBM model.

[0140] The data of the key parts of the vehicle include speed, acceleration and engine status data of the key parts of the engine management system, gearbox and safety system.

[0141] In one embodiment, when the processor 502 implements the step of preprocessing the initial data to obtain a preprocessing result, the processor 502 specifically implements the following steps: The initial data is cleaned to obtain a cleaning result; the cleaning result is subjected to feature engineering processing to obtain a feature extraction result; the feature extraction result is subjected to data fusion to obtain a fusion result; the fusion result is subjected to standardization processing to obtain a preprocessing result.

[0142] In one embodiment, the processor 502 implements the vehicle fault prediction model by preprocessing the key vehicle part data and environmental information obtained by the vehicle communication bus based on V2X and combining the real results as the sample set to train the basic model, wherein when the basic model includes the LightGBM model step, the following steps are specifically implemented: The vehicle key parts data and environmental information obtained through the V2X-based vehicle communication bus are preprocessed and combined with real results as a sample set; a basic model is constructed; a loss function is defined, and an optimization algorithm is selected; the basic model is forward propagated, loss calculated, back propagated, and weighted updated using the sample set, combined with the loss function and the optimization algorithm, and the model parameters are continuously optimized until the best performance is achieved to obtain a trained model; the trained model is evaluated for model stability and generalization ability using the K-fold cross-validation method to determine a vehicle fault prediction model.

[0143] In one embodiment, after implementing the step of outputting the prediction result, the processor 502 further implements the following steps: Collect user feedback and actual fault conditions, and retrain the vehicle fault prediction model based on the collected data to update the vehicle fault prediction model.

[0144] In one embodiment, when the processor 502 implements the step of inputting the preprocessing result into the vehicle fault prediction model for real-time fault warning, fault prediction, and automatic driving fault prevention prediction to obtain the prediction result, the processor 502 specifically implements the following steps: When performing real-time fault warning, the preprocessing results are input into the vehicle fault prediction model, and the preprocessing data is comprehensively analyzed to quickly identify potential fault risks of the vehicle and issue an alarm in a timely manner; when performing fault prediction, the preprocessing results are input into the vehicle fault prediction model, and through in-depth analysis of vehicle usage records, driving habits and environmental factors, possible problems are foreseen; when performing automatic driving fault prevention prediction, the preprocessing results are input into the vehicle fault prediction model to automatically adjust the vehicle speed, implement emergency braking or perform obstacle avoidance steering operations.

[0145] In one embodiment, when the processor 502 implements the automatic driving fault prevention prediction, inputs the preprocessing result into the vehicle fault prediction model, and automatically adjusts the vehicle speed, implements emergency braking, or performs obstacle avoidance steering operation steps, the following steps are specifically implemented: The V2X network continuously monitors the vehicle's operating status and changes in the surrounding environment, and identifies factors that may threaten driving safety in advance.

[0146] It should be understood that in the embodiment of the present application, the processor 502 may be a central processing unit (CPU), and the processor 502 may also be other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASIC), field-programmable gate arrays (FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. Among them, the general-purpose processor may be a microprocessor or the processor may also be any conventional processor, etc.

[0147] It can be understood by those skilled in the art that all or part of the processes in the method for implementing the above embodiment can be completed by instructing the relevant hardware through a computer program. The computer program includes program instructions, and the computer program can be stored in a storage medium, which is a computer-readable storage medium. The program instructions are executed by at least one processor in the computer system to implement the process steps of the embodiment of the above method.

[0148] Therefore, the present invention also provides a storage medium. The storage medium may be a computer-readable storage medium. The storage medium stores a computer program, wherein when the computer program is executed by a processor, the processor executes the following steps: Acquire a V2X-based vehicle communication bus to collect data from key parts of the vehicle and environmental information to obtain initial data; preprocess the initial data to obtain a preprocessing result; input the preprocessing result into a vehicle fault prediction model to perform real-time fault warning, fault prediction, and automatic driving fault prevention prediction to obtain a prediction result; and output the prediction result.

[0149] The vehicle fault prediction model is obtained by preprocessing the key vehicle parts data and environmental information obtained by the V2X-based vehicle communication bus and combining the real results as a sample set to train the basic model, wherein the basic model includes the LightGBM model.

[0150] The data of the key parts of the vehicle include speed, acceleration and engine status data of the key parts of the engine management system, gearbox and safety system.

[0151] In one embodiment, when the processor executes the computer program to implement the step of preprocessing the initial data to obtain a preprocessing result, the processor specifically implements the following steps: The initial data is cleaned to obtain a cleaning result; the cleaning result is subjected to feature engineering processing to obtain a feature extraction result; the feature extraction result is subjected to data fusion to obtain a fusion result; the fusion result is subjected to standardization processing to obtain a preprocessing result.

[0152] In one embodiment, the processor implements the vehicle fault prediction model by executing the computer program, which is obtained by preprocessing the key vehicle part data and environmental information obtained by the vehicle communication bus based on V2X and combining the real results as a sample set to train the basic model, wherein when the basic model includes the LightGBM model step, the following steps are specifically implemented: The vehicle key parts data and environmental information obtained through the V2X-based vehicle communication bus are preprocessed and combined with real results as a sample set; a basic model is constructed; a loss function is defined, and an optimization algorithm is selected; the basic model is forward propagated, loss calculated, back propagated, and weighted updated using the sample set, combined with the loss function and the optimization algorithm, and the model parameters are continuously optimized until the best performance is achieved to obtain a trained model; the trained model is evaluated for model stability and generalization ability using the K-fold cross-validation method to determine a vehicle fault prediction model.

[0153] In one embodiment, after executing the computer program to implement the step of outputting the prediction result, the processor further implements the following steps: Collect user feedback and actual fault conditions, and retrain the vehicle fault prediction model based on the collected data to update the vehicle fault prediction model.

[0154] In one embodiment, when the processor executes the computer program to implement the step of inputting the preprocessing result into a vehicle fault prediction model for real-time fault warning, fault prediction, and automatic driving fault prevention prediction to obtain a prediction result, the processor specifically implements the following steps: When performing real-time fault warning, the preprocessing results are input into the vehicle fault prediction model, and the preprocessing data is comprehensively analyzed to quickly identify potential fault risks of the vehicle and issue an alarm in a timely manner; when performing fault prediction, the preprocessing results are input into the vehicle fault prediction model, and through in-depth analysis of vehicle usage records, driving habits and environmental factors, possible problems are foreseen; when performing automatic driving fault prevention prediction, the preprocessing results are input into the vehicle fault prediction model to automatically adjust the vehicle speed, implement emergency braking or perform obstacle avoidance steering operations.

[0155] In one embodiment, the processor, when executing the computer program to implement the automatic driving fault prevention prediction, inputting the preprocessing result into the vehicle fault prediction model, automatically adjusting the vehicle speed, implementing emergency braking or performing obstacle avoidance steering operation steps, further implements the following steps: The V2X network continuously monitors the vehicle's operating status and changes in the surrounding environment, and identifies factors that may threaten driving safety in advance.

[0156] The storage medium may be a USB flash drive, a mobile hard disk, a read-only memory (ROM), a magnetic disk, or an optical disk, etc., which are computer-readable storage media that can store program codes.

[0157] Those of ordinary skill in the art will appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, computer software, or a combination of the two. In order to clearly illustrate the interchangeability of hardware and software, the composition and steps of each example have been generally described in terms of function in the above description. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of the present invention.

[0158] In the several embodiments provided by the present invention, it should be understood that the disclosed devices and methods can be implemented in other ways. For example, the device embodiments described above are only schematic. For example, the division of each unit is only a logical function division, and there may be other division methods in actual implementation. For example, multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed.

[0159] The steps in the method of the embodiment of the present invention can be adjusted in order, combined and deleted according to actual needs. The units in the device of the embodiment of the present invention can be combined, divided and deleted according to actual needs. In addition, the functional units in the various embodiments of the present invention can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit.

[0160] If the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a storage medium. Based on this understanding, the technical solution of the present invention is essentially 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. The computer software product is stored in a storage medium and includes several instructions for a computer device (which can be a personal computer, terminal, or network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present invention.

[0161] The above is only a specific embodiment of the present invention, but the protection scope of the present invention is not limited thereto. Any technician familiar with the technical field can easily think of various equivalent modifications or replacements within the technical scope disclosed by the present invention, and these modifications or replacements should be included in the protection scope of the present invention. Therefore, the protection scope of the present invention shall be based on the protection scope of the claims.

Claims

1. A vehicle fault prediction method based on V2X, characterized in that: include: Acquire the V2X-based vehicle communication bus to collect data from key parts of the vehicle and environmental information to obtain initial data; Preprocessing the initial data to obtain a preprocessing result; Inputting the preprocessing results into a vehicle fault prediction model to perform real-time fault warning, fault prediction, and automatic driving fault prevention prediction to obtain a prediction result; The prediction result is outputted.

2. The V2X-based vehicle fault prediction method according to claim 1, characterized in that: The vehicle fault prediction model is obtained by preprocessing the key parts data of the vehicle and the environmental information obtained by the V2X-based vehicle communication bus and combining the real results as the sample set to train the basic model, wherein the basic model includes the LightGBM model.

3. The V2X-based vehicle fault prediction method according to claim 1, characterized in that: The data of the key parts of the vehicle include speed, acceleration and engine status data of the key parts of the engine management system, gearbox and safety system.

4. The V2X-based vehicle fault prediction method according to claim 1, characterized in that: The preprocessing of the initial data to obtain a preprocessing result includes: Performing data cleaning on the initial data to obtain a cleaning result; Performing feature engineering processing on the cleaning result to obtain a feature extraction result; Performing data fusion on the feature extraction results to obtain a fusion result; The fusion result is standardized to obtain a preprocessing result.

5. The V2X-based vehicle fault prediction method according to claim 2, characterized in that: The vehicle fault prediction model is obtained by preprocessing the key vehicle parts data and environmental information obtained by the V2X vehicle communication bus and combining the real results as the sample set to train the basic model, wherein the basic model includes the LightGBM model, including: The key vehicle parts data and environmental information obtained through the V2X-based vehicle communication bus are pre-processed and combined with the real results as a sample set; Build the basic model; Define the loss function and choose the optimization algorithm; Using the sample set, combined with the loss function and the optimization algorithm, the basic model is forward propagated, loss calculated, back propagated and weight updated, and the model parameters are continuously optimized until the best performance is achieved, so as to obtain a trained model; The K-fold cross-validation method is used to evaluate the model stability and generalization ability of the trained model to determine the vehicle fault prediction model.

6. The V2X-based vehicle fault prediction method according to claim 5, characterized in that: After outputting the prediction result, the method further includes: Collect user feedback and actual fault conditions, and retrain the vehicle fault prediction model based on the collected data to update the vehicle fault prediction model.

7. The V2X-based vehicle fault prediction method according to claim 1, characterized in that: The preprocessing result is input into the vehicle fault prediction model to perform real-time fault warning, fault prediction and automatic driving fault prevention prediction to obtain a prediction result, including: When real-time fault warning is performed, the preprocessing results are input into the vehicle fault prediction model, the preprocessing data is comprehensively analyzed, the potential fault risk of the vehicle is quickly identified, and an alarm is issued in time; When fault prediction is performed, the preprocessing results are input into the vehicle fault prediction model, and possible problems are foreseen through in-depth analysis of vehicle usage records, driving habits and environmental factors; When performing automatic driving fault prevention prediction, the preprocessing result is input into the vehicle fault prediction model to automatically adjust the vehicle speed, implement emergency braking or perform obstacle avoidance steering operations.

8. The V2X-based vehicle fault prediction method according to claim 7, characterized in that: When performing automatic driving fault prevention prediction, the preprocessing result is input into the vehicle fault prediction model to automatically adjust the vehicle speed, implement emergency braking or perform obstacle avoidance steering operations, and further includes: The V2X network continuously monitors the vehicle's operating status and changes in the surrounding environment, and identifies factors that may threaten driving safety in advance.

9. A vehicle fault prediction device based on V2X, characterized in that: include: A data acquisition unit, used to acquire data from key parts of the vehicle and environmental information collected by the V2X-based vehicle communication bus to obtain initial data; A preprocessing unit, used for preprocessing the initial data to obtain a preprocessing result; A prediction unit, used for inputting the preprocessing result into a vehicle fault prediction model to perform real-time fault warning, fault prediction and automatic driving fault prevention prediction to obtain a prediction result; An output unit is used to output the prediction result.

10. A computer device, characterized in that: The computer device comprises a memory and a processor, wherein a computer program is stored in the memory, and the processor implements the method according to any one of claims 1 to 8 when executing the computer program.

Images