Determination method of fault identification model and determination method of vehicle fault

Abstract

The invention discloses a method for determining a fault recognition model and a method for determining a vehicle fault, and belongs to the technical field of vehicle detection. The method for determining the fault recognition model comprises the steps of classifying sample vehicle state data corresponding to historical vehicles according to fault types and fault levels in sequence, and determining sample fault labels corresponding to the sample vehicle state data, the sample fault tag is used for representing a tag including a fault type and a fault level of sample vehicle state data; inputting the state data of each sample vehicle and the corresponding target sample fault tag into the initial fault recognition model for training, and determining a candidate fault recognition model corresponding to the historical vehicle; and based on the candidate fault recognition model, a preset text similarity algorithm and a preset error algorithm, determining a trained standard fault recognition model corresponding to the historical vehicle. According to the invention, more flexible and accurate early warning and identification can be carried out on faults of different types of vehicles based on deep learning.

Description

Technical Field

[0001] The present application relates to the field of vehicle detection technology, and in particular to a method for determining a fault identification model and a method for determining a vehicle fault. Background Art

[0002] At present, in the field of automobile fault warning, it is crucial to accurately identify potential vehicle problems. However, in existing vehicle fault diagnosis systems, fault detection and warning usually rely on fixed thresholds and rules to identify anomalies. These rules and thresholds are based on the standards and experience provided by vehicle manufacturers. Therefore, these rules and thresholds cannot adapt to different types of vehicles, and have poor flexibility and accuracy, which further leads to safety hazards in the corresponding vehicles. Summary of the invention

[0003] An embodiment of the present application provides a method that solves the technical problem in the prior art that the fault identification system cannot adapt to different types of vehicles and has poor flexibility and accuracy. The embodiment of the present application enables more flexible and accurate early warning and identification of faults of different types of vehicles based on deep learning.

[0004] In a first aspect of an embodiment of the present application, an embodiment of the present application provides a method for determining a fault identification model, and the method for determining a fault identification model includes:

[0005] Classify the sample vehicle status data corresponding to the historical vehicles according to the fault type and fault level, and determine the sample fault labels corresponding to each of the historical vehicles, wherein the sample fault labels are used to characterize the labels including the fault type and fault level of the sample vehicle status data;

[0006] Inputting each of the sample vehicle status data and the corresponding target sample fault label into an initial fault recognition model for training, and determining a candidate fault recognition model corresponding to the historical vehicle;

[0007] Based on the candidate fault recognition model, a preset text similarity algorithm and a preset error algorithm, a trained standard fault recognition model corresponding to the historical vehicle is determined.

[0008] In a feasible implementation manner, the sample vehicle status data corresponding to the historical vehicle is determined in the following manner, and the method for determining the fault identification model includes:

[0009] Obtain the initial vehicle status data corresponding to the historical vehicle;

[0010] Determining first candidate vehicle state data corresponding to the historical vehicle based on the initial vehicle state data and a preset local abnormality factor algorithm;

[0011] Determining second candidate vehicle state data corresponding to the historical vehicle based on the first candidate vehicle state data and a preset normalization algorithm;

[0012] Data expansion is performed based on the second candidate vehicle status data to determine sample vehicle status data corresponding to the historical vehicle.

[0013] In a feasible implementation manner, the inputting of each of the sample vehicle status data and the corresponding target sample fault label into the initial fault recognition model for training to determine the candidate fault recognition model corresponding to the historical vehicle includes:

[0014] Each sample vehicle status data and the corresponding target sample fault label are divided according to the preset training task, and the first sub-sample vehicle status data and the corresponding first target sub-sample fault label under different training tasks are determined;

[0015] Inputting each first sub-sample vehicle state data and the corresponding first target sub-sample fault label into the initial fault recognition model for training, and determining the training data under each training task;

[0016] Based on the training data under each of the training tasks, a candidate fault recognition model corresponding to the historical vehicle is determined.

[0017] In a feasible implementation manner, the initial fault identification model includes at least one first initial fault sub-identification model, and the inputting of each of the sample vehicle status data and the corresponding target sample fault label into the initial fault identification model for training to determine the candidate fault identification model corresponding to the historical vehicle also includes:

[0018] Each sample vehicle status data and the corresponding target sample fault label are divided according to the preset vehicle model, and the second sub-sample vehicle status data and the corresponding second target sub-sample fault label under different models are determined;

[0019] Inputting each second sub-sample vehicle status data and the corresponding second target sub-sample fault label into each corresponding first initial fault sub-recognition model for training to determine the prediction type under each of the models;

[0020] Based on each of the prediction types, each of the first initial fault sub-identification models is integrated to determine a candidate fault identification model corresponding to the historical vehicle.

[0021] In a feasible implementation manner, the determining the trained standard fault recognition model corresponding to the historical vehicle based on the candidate fault recognition model, the preset text similarity algorithm and the preset error algorithm includes:

[0022] Based on the candidate fault identification model and a preset error algorithm, determining the error between the candidate prediction label predicted by the candidate fault identification model and the sample fault label;

[0023] Based on the error, the candidate fault recognition model is trained to determine an intermediate fault recognition model corresponding to the historical vehicle;

[0024] Based on the external fault label-historical maintenance database and a preset text similarity algorithm, determining the similarity between the intermediate prediction label predicted by the intermediate fault recognition model and the sample fault label;

[0025] Based on the similarity, the intermediate fault recognition model is trained to determine the trained standard fault recognition model corresponding to the historical vehicle.

[0026] According to a second aspect of the embodiments of the present application, the embodiments of the present application provide a method for determining a vehicle fault, the method for determining a vehicle fault comprising:

[0027] The target vehicle status data of the target vehicles in the batch to be tested are input into the trained standard fault recognition model to determine whether each of the target vehicles has a fault and the target fault type corresponding to the fault.

[0028] According to a third aspect of the embodiments of the present application, the embodiments of the present application provide a device for determining a fault identification model, wherein the device for determining a fault identification model comprises:

[0029] A first determination module is used to classify the sample vehicle status data corresponding to the historical vehicles according to the fault type and the fault level, and determine the sample fault label corresponding to each of the historical vehicles, wherein the sample fault label is used to characterize the label including the fault type and the fault level of the sample vehicle status data;

[0030] A second determination module is used to input each of the sample vehicle status data and the corresponding target sample fault label into an initial fault recognition model for training, and determine a candidate fault recognition model corresponding to the historical vehicle;

[0031] The third determination module is used to determine the trained standard fault recognition model corresponding to the historical vehicle based on the candidate fault recognition model, the preset text similarity algorithm and the preset error algorithm.

[0032] According to a fourth aspect of the embodiments of the present application, an embodiment of the present application provides a vehicle fault determination device, the vehicle fault determination device comprising:

[0033] The determination module is used to input the target vehicle status data of the target vehicles in the batch to be detected into the trained standard fault recognition model to determine whether each of the target vehicles has a fault and the target fault type corresponding to the fault.

[0034] In the fifth aspect of the embodiments of the present application, the embodiments of the present application provide an electronic device, comprising: a processor, a memory and a bus, the memory storing machine-readable instructions executable by the processor, when the electronic device is running, the processor and the memory communicate through the bus, and the machine-readable instructions are executed by the processor to execute the steps of the method for determining the fault identification model or the steps of the method for determining the vehicle fault as mentioned above.

[0035] In the sixth aspect of the embodiments of the present application, the embodiments of the present application provide a computer-readable storage medium, on which a computer program is stored. When the computer program is executed by a processor, the steps of the method for determining the fault identification model or the steps of the method for determining the vehicle fault as described above are executed.

[0036] The method for determining a fault identification model and a method for determining a vehicle fault provided in the embodiments of the present application, compared with the prior art, the embodiments provided in the present application classify the sample vehicle status data corresponding to the historical vehicles according to the fault type and fault level in turn, determine the sample fault label corresponding to each sample vehicle status data, and then input each sample vehicle status data and the corresponding target sample fault label into the initial fault identification model for training, determine the candidate fault identification model corresponding to the historical vehicle, and determine the trained standard fault identification model corresponding to the historical vehicle based on the candidate fault identification model, the preset text similarity algorithm and the preset error algorithm, so as to provide more flexible and accurate early warning and identification of faults of different types of vehicles based on deep learning. BRIEF DESCRIPTION OF THE DRAWINGS

[0037] Figure 1 A flowchart of a method for determining a fault identification model provided in an embodiment of the present application is shown;

[0038] Figure 2 A screenshot showing a method for determining a fault identification model provided in an embodiment of the present application using a specific protocol to obtain sample vehicle status data;

[0039] Figure 3 A flowchart of a method for determining a vehicle fault provided by an embodiment of the present application is shown;

[0040] Figure 4 A structural block diagram of a device for determining a fault identification model provided in an embodiment of the present application is shown;

[0041] Figure 5 A structural block diagram of a vehicle fault determination device provided by an embodiment of the present application is shown;

[0042] Figure 6 A schematic structural diagram of an electronic device provided in an embodiment of the present application is shown.

[0043] Figure 5 and Figure 6 The corresponding relationship between the reference numerals and the names of the drawings is as follows:

[0044] 400 fault identification model determination device; 410 first determination module; 420 second determination module; 430 third determination module; 500 vehicle fault determination device; 510 determination module; 600 electronic device; 610 processor; 620 memory; 630 bus. DETAILED DESCRIPTION

[0045] In order to better understand the technical solutions provided by the embodiments of this specification, the technical solutions of the embodiments of this specification are described in detail below through the accompanying drawings and specific embodiments. It should be understood that the embodiments of this specification and the specific features in the embodiments are detailed descriptions of the technical solutions of the embodiments of this specification, rather than limitations on the technical solutions of this specification. In the absence of conflict, the embodiments of this specification and the technical features in the embodiments can be combined with each other.

[0046] In this article, relational terms such as first and second, etc. are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply that there is any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that the process, method, article or equipment including a series of elements includes not only those elements, but also includes other elements not explicitly listed, or also includes elements inherent to such process, method, article or equipment. In the absence of more restrictions, the elements limited by the statement "comprise one..." do not exclude the existence of other identical elements in the process, method, article or equipment including the elements. The term "more than two" includes two or more than two situations.

[0047] First, the applicable application scenarios of the present application are introduced. The embodiments provided in the present application are applicable to the field of vehicle detection technology.

[0048] At present, in the field of automobile fault warning, it is crucial to accurately identify potential vehicle problems. In existing vehicle fault diagnosis systems, fault detection and warning usually rely on fixed thresholds and rules to identify anomalies. These rules and thresholds are based on the standards and experience provided by vehicle manufacturers. Therefore, these rules and thresholds cannot adapt to different types of vehicles, and have poor flexibility and accuracy, which further leads to safety hazards in the corresponding vehicles.

[0049] The existing reliance on fixed thresholds and rules cannot adapt to all vehicle models or vehicle behaviors under specific conditions, which may lead to false alarms or missed alarms. Traditionally collected vehicle data comes from different sensors and systems, which are highly diverse and complex. Therefore, it is difficult to effectively integrate and analyze these data using traditional diagnostic methods, which in turn leads to the inability to identify more complex and hidden vehicle failure modes in real time. In addition, due to the different usage and wear conditions of different vehicles, traditional fault detection systems cannot provide more personalized and adaptive warnings for each vehicle's faults.

[0050] Based on this, an embodiment of the present application provides a method for determining a fault identification model and a method for determining a vehicle fault, which can provide more flexible and accurate early warning and identification of faults of different types of vehicles based on deep learning.

[0051] See also Figure 1 , Figure 1 A flowchart of a method for determining a fault identification model provided in an embodiment of the present application is shown in FIG. Figure 1 As shown, the method for determining the fault identification model includes the following steps:

[0052] S101. Classify sample vehicle status data corresponding to historical vehicles according to fault type and fault level, and determine sample fault labels corresponding to each of the historical vehicles, wherein the sample fault labels are used to characterize labels including the fault type and fault level of the sample vehicle status data.

[0053] In this step, the embodiment provided by the present application first needs to use a preset acquisition protocol to obtain sample vehicle status data corresponding to the historical vehicle, and then organize the sample vehicle status data (specifically, further data screening and further data preprocessing to ensure the consistency and accuracy of the sample vehicle status data, and remove duplicate, erroneous and invalid sample vehicle status data), and after further organization, the sample vehicle status data is classified according to the type and characteristics of the vehicle fault corresponding to the historical vehicle in turn according to the fault type and fault level. It is assumed that the fault type in the embodiment provided by the present application can be specific but not limited to: engine fault, chassis fault, electronic and electrical equipment fault, etc., and under each fault type, there are further subdivided sub-class faults, such as ignition system fault and fuel supply system fault in engine fault, etc. After the sample vehicle status data is classified according to the fault type, the fault level of the historical vehicle is set and determined according to the degree of damage of the historical vehicle with different types of faults, and the sample fault label is generated / determined according to the fault type and fault level corresponding to each historical vehicle. Here, it is assumed that the sample fault label in the embodiment provided by the present application is set to three levels: high, medium and low.

[0054] In the above, it is assumed that the sample fault labels in the embodiments provided in the present application can be specifically but not limited to being defined as: such as "battery undervoltage-medium level" or "engine abnormality-medium level" and the like.

[0055] It is assumed that the method for determining the fault identification model in the embodiment provided in the present application is mainly used in a fault warning platform.

[0056] Among them, it is assumed that the sample vehicle status data in the embodiments provided in the present application can be specific but not limited to data of various vehicle parameters such as historical vehicle static data, whole vehicle data and power battery pack data. It is assumed that the static data includes: vehicle identification code (VIN number) and vehicle model: whole vehicle data includes: vehicle status, charging status, vehicle speed, battery SOC (state of charge and discharge) and insulation resistance, etc. These sample vehicle status data not only provide detailed usage records for each historical vehicle, but also with the daily operation of historical vehicles, these data are gradually accumulated to form a complete life cycle data record, which is automatically uploaded to the preset external big data platform for backup through the on-board communication equipment, providing important data support for the maintenance and fault warning of the corresponding vehicle.

[0057] Among them, the preset acquisition protocol in the embodiment provided by this application can be customized and used according to different application scenarios and usage scenarios. It is assumed that the preset acquisition protocol in the embodiment provided by this application can be specifically but not limited to using the 32960 protocol to collect sample vehicle status data. For specific vehicle status data types, please refer to Figure 2 , Figure 2 A screenshot of a method for determining a fault identification model provided in an embodiment of the present application using a specific protocol to obtain sample vehicle status data, such as Figure 2 As shown, Figure 2 This is some of the sample vehicle status data collected.

[0058] In the above, it is assumed that the fault type in the embodiment provided in the present application is obtained based on an external fault database. The following describes the contents of the external fault database:

[0059] It is assumed that the external fault database in the embodiment provided in the present application specifically but not limited to includes: a vehicle information table (Vehicles), a fault type table (FaultTypes) and a fault record table (FaultRecords), wherein the vehicle information table (Vehicles) and the fault record table (FaultRecords) are associated through the vin field, indicating the fault record of a historical vehicle; the fault record table (FaultRecords) and the fault type table (FaultTypes) are associated through the fault_id field, indicating the fault type corresponding to each fault record.

[0060] The vehicle information table (Vehicles) is shown in Table 1:

[0061] Field Name Field Type Field length describe vin VARCHAR(17) 17 Vehicle Identification Number model VARCHAR(50) 50 Vehicle Model manufacturer VARCHAR(50) 50 Manufacturer

[0062] Table 1

[0063] The fault type table (FaultTypes) is shown in Table 2:

[0064]

[0065] Table 2

[0066] The fault record table (FaultRecords) is shown in Table 3:

[0067] Field Name Field Type Field length describe recorg_id INT 11 Fault record ID (primary key) vin VARCHAR(17) 17 Vehicle Identification Number (Primary Key) fault_id INT 11 Fault type ID (foreign key) fault_time DATETIME Failure time details TEXT Detailed description of the fault solve TEXT Troubleshooting Description

[0068] Table 3

[0069] Here, the embodiment provided by the present application adds indexes to fields that are frequently used for query (such as vin, fault_id, etc.) to improve data retrieval efficiency. Analyze and optimize database query statements to avoid inefficient operations such as full table scans to ensure query accuracy and efficiency.

[0070] In the above, the embodiments provided in the present application design a RESTful API interface or a SOAP interface, which is called by a fault identification model or other analysis and control device to obtain sample vehicle status data in an external fault database corresponding to historical vehicles, and the embodiments provided in the present application can specifically but are not limited to using JSON as the data transmission format to ensure the correct transmission and parsing of data.

[0071] S102: Input each of the sample vehicle status data and the corresponding target sample fault label into an initial fault recognition model for training, and determine a candidate fault recognition model corresponding to the historical vehicle.

[0072] In this step, after the embodiment provided in the present application determines the target sample fault labels of each sample vehicle status data, each target sample fault label of the sample vehicle status data is input into the initial fault recognition model of deep learning for training, and after the training is completed, the candidate fault recognition model corresponding to the historical vehicle is determined.

[0073] It can be understood that the training of the initial fault identification model in the embodiment provided in the present application requires a large amount of computing resources, and distributed training has become a key technology to achieve this goal. The distributed training of the present application can be carried out in two ways: data parallelism (DataParallelism) and model parallelism (ModelParallelism).

[0074] Among them, data parallelism is used to divide the input data into several parts and distribute them to different computing nodes for parallel processing, and complete a part of the training task on each node, update the model parameters together, and finally merge the results; while model parallelism is used to divide the model parameters and calculations into several sub-models, distribute them to different computing nodes for execution, and the sub-models exchange intermediate results through communication to jointly complete the training task.

[0075] In the above, the initial fault recognition model of the degree learning in the embodiment provided in the present application can be but is not limited to adopting the Transformer architecture. Transformer is a network structure based on the attention mechanism, which has powerful parallelization capabilities and can achieve fast training on large-scale data sets.

[0076] Transformer includes: Multi-head Self-Attention, Positional Encoding, and Feed-forward Neural Network.

[0077] Among them, Multi-head Self-Attention is specifically used to enable the model to pay attention to information at all positions when processing sequence data and capture global contextual relationships.

[0078] Positional Encoding is specifically used to introduce position information to ensure that the model can understand the order relationship in the sequence.

[0079] The feed-forward neural network is specifically used to further process and nonlinearly transform the information at each position.

[0080] S103 , determining a trained standard fault recognition model corresponding to the historical vehicle based on the candidate fault recognition model, a preset text similarity algorithm, and a preset error algorithm.

[0081] In this step, after the candidate fault identification model is determined, the embodiment provided in the present application determines the consistency between the sample fault label and the predicted fault label according to the preset error algorithm and the candidate fault identification model, and calculates indicators such as precision, recall and precision score according to the consistency rate between the two to quantitatively analyze the performance of the candidate fault identification model. After the candidate model is trained according to the preset error algorithm, the candidate model trained by the preset error algorithm is trained for similarity according to the preset text similarity algorithm to determine the trained standard fault identification model corresponding to the historical vehicle.

[0082] It can be understood that the preset error algorithm in the embodiments provided in the present application can specifically but not limited to use a confusion matrix algorithm; the preset text similarity algorithm can specifically but not limited to use a BERTScore algorithm.

[0083] The method for determining a fault identification model provided in an embodiment of the present application, compared with the prior art, classifies the sample vehicle status data corresponding to historical vehicles according to fault type and fault level in turn, determines the sample fault label corresponding to each sample vehicle status data, and then inputs each sample vehicle status data and the corresponding target sample fault label into the initial fault identification model for training, determines the candidate fault identification model corresponding to the historical vehicle, and determines the trained standard fault identification model corresponding to the historical vehicle based on the candidate fault identification model, a preset text similarity algorithm, and a preset error algorithm, which can provide more flexible and accurate early warning and identification of faults of different types of vehicles based on deep learning.

[0084] In a feasible embodiment, the sample vehicle status data corresponding to the historical vehicle is determined through the following sub-steps:

[0085] Sub-step 1: Obtain the initial vehicle status data corresponding to the historical vehicle.

[0086] In this step, the embodiment provided in the present application first obtains the initial vehicle status data corresponding to the historical vehicle. Here, the initial vehicle status data is the initial vehicle status data obtained without any processing.

[0087] Sub-step 2: determining first candidate vehicle state data corresponding to the historical vehicle based on the initial vehicle state data and a preset local abnormality factor algorithm.

[0088] In this step, since the initial vehicle state data of historical vehicle characteristics such as power batteries may be mixed with error interference data, obvious outliers may appear in the initial vehicle state data. This may be caused by data acquisition errors and vehicle degradation characteristics. In order to solve this problem, the present invention uses a preset local anomaly factor algorithm to process the outlier data to generate the first candidate vehicle state data. The use of the initial vehicle state data can eliminate the influence of the test error on the regularity of the feature quantity, thereby effectively improving the information amount of the feature quantity and the learning efficiency of the model.

[0089] Assume that the local outlier factor algorithm preset in the embodiment provided by the present application can be but is not limited to a local outlier factor (LOF) algorithm. The local outlier factor (LOF) algorithm is a density-based outlier detection method that aims to quantify the degree of anomaly in data points. It determines whether the data point is an outlier and its degree of outlier by calculating the local outlier factor (LOF) of each data point. According to the set threshold, invalid outliers can be accurately located, avoiding the problem of incorrect cleaning of valid data by the outlier discrimination method based on data distribution. In the feature quantity, the average ratio of the local reachable density of a point to the local reachable density of its neighboring points is expressed as the LOF of point n. The specific calculation formula is as follows:

[0090]

[0091] Among them, N K (n) represents the k distance neighborhoods of point n, and point m is a point in the n neighborhood. k (n) and lrd k (m) represents the local reachable density of point n and point m respectively. When the LOF value is close to 1, it means that the surrounding density of point n and its neighboring points is basically the same. The greater the probability that n and its neighboring points are in the same cluster, on the contrary, n may be an outlier.

[0092] Sub-step 3: determining the second candidate vehicle state data corresponding to the historical vehicle based on the first candidate vehicle state data and a preset normalization algorithm.

[0093] In this step, after generating the first candidate vehicle status data, the embodiment provided by the present application performs corresponding normalization processing on the corresponding first candidate vehicle status data according to a preset normalization algorithm to determine and generate the second candidate vehicle status data.

[0094] It is assumed that the preset normalization algorithm in the embodiment provided in the present application includes but is not limited to the Z-score method, which is implemented by converting the data into a standard normal distribution (mean is 0, standard deviation is 1), and the principle of normalization using the Z-score method is to subtract the mean of each feature column from the data of the column, and then divide it by the standard deviation of the column. The specific calculation formula is as follows:

[0095]

[0096] Here, the Z-score processing is used to make the data of each feature column conform to the standard normal distribution, that is, the mean is 0 and the standard deviation is 1. The advantage of this is that it can eliminate the dimensional influence between different features, make different features have the same scale, and can better meet the requirements of some deep learning algorithms.

[0097] Sub-step 4: performing data expansion based on the second candidate vehicle status data to determine sample vehicle status data corresponding to the historical vehicle.

[0098] In this step, since the second candidate vehicle status data may have a class imbalance problem in actual operation, the embodiment provided by the present application needs to perform data expansion on the second candidate vehicle status data. The specific method may be: using the Safe-LevelSMOTE algorithm to balance the data set by synthesizing new minority class samples, while trying to maintain the distribution characteristics of the data and avoid the risk of overfitting. The following are the calculation steps of the Safe-LevelSMOTE algorithm:

[0099] a. Calculate the distance matrix: For each minority class sample x i , calculate its difference with all other minority class samples x j The Euclidean distance d(x i ,x j ), the formula is as follows:

[0100]

[0101] Here, n is the number of features.

[0102] b. Standardized distance matrix: Standardize the distance matrix to ensure that all distance indicators have the same weight. The standardized distance calculation formula is:

[0103]

[0104] Among them, min(d(x i ,·)) and max(d(x i ,·)) are X i The minimum and maximum distances to all other samples.

[0105] c. Calculate Safe-Level: For each minority class sample, calculate its x i Safe-Level is the ratio of the closest distance of the sample to the minority class sample to the closest distance to the majority class sample. The calculation formula of Safe-Level is:

[0106]

[0107] in, Yes X i The minimum value of the standardized distance from Xi to other minority class samples, and the other is the minimum value of the standardized distance from Xi to other majority class samples.

[0108] d. Select reference sample: Based on the calculated Safe-Level, select a neighbor with the smallest Safe-Level for the current sample as a reference sample.

[0109] e. Number of generated samples: According to the value of Safe-Level, determine the number of synthetic samples (N) that need to be generated. The calculation formula is:

[0110] N=(Safe-Level(x i )-1)×α;

[0111] Among them, α is a parameter that controls the number of generated samples.

[0112] f. Generate samples by linear interpolation: Generate synthetic samples by linear interpolation between the current sample Xi and the selected reference sample Xr. The formula for linear interpolation is:

[0113] x new =x i +β×(x r -x i ),β∈[0,1];

[0114] Where β is the interpolation ratio determined according to N and α.

[0115] g. Add synthetic samples: Add the generated synthetic samples to the minority class sample set to increase its number.

[0116] Here, the Safe-Level SMOTE algorithm in the embodiment provided in the present application adds consideration of the local structure of the data on the basis of the traditional SMOTE algorithm. By introducing the concept of Safe-Level, the oversampling process is more refined and safe. This algorithm is particularly suitable for situations where minority class samples are surrounded by majority class samples. It can effectively increase the number of minority class samples without destroying the original structure of the data, while maintaining the distribution characteristics of the data and avoiding the risk of overfitting, thereby improving the generalization ability of the model.

[0117] In a feasible embodiment, it is assumed that the candidate fault identification model in the embodiment provided in the present application is trained using data parallelism, specifically: first, each sample vehicle status data and the corresponding target sample fault label are divided according to a preset training task, and the first sub-sample vehicle status data and the corresponding first target sub-sample fault label under different training tasks are determined. Then, each first sub-sample vehicle status data and the corresponding first target sub-sample fault label are input into the initial fault identification model for training, and the training data under each training task is determined. Finally, based on the training data under each training task, the candidate fault identification model corresponding to the historical vehicle is determined.

[0118] In a feasible embodiment, it is assumed that the candidate fault identification model in the embodiment provided in the present application is trained using model parallelism, specifically: first, each sample vehicle status data and the corresponding target sample fault label are divided according to the preset vehicle model, and the second sub-sample vehicle status data and the corresponding second target sub-sample fault label under different models are determined; then, each second sub-sample vehicle status data and the corresponding second target sub-sample fault label are input into the corresponding first initial fault sub-identification model for training to determine the prediction type under each model; finally, based on each prediction type, each first initial fault sub-identification model is integrated to determine the candidate fault identification model corresponding to the historical vehicle.

[0119] In a feasible embodiment, step S103 includes the following sub-steps:

[0120] Sub-step 1031 : determining the error between the candidate prediction label predicted by the candidate fault identification model and the sample fault label based on the candidate fault identification model and a preset error algorithm.

[0121] In this step, after the candidate fault identification model is determined, the embodiment provided in the present application evaluates the error between the candidate prediction label and the sample fault label output by the candidate fault identification model according to the confusion matrix in the preset error algorithm, that is, the error between the sample fault label and the predicted fault label, and based on the error, trains the candidate fault identification model to determine the intermediate fault identification model corresponding to the historical vehicle.

[0122] Sub-step 1032: determining the similarity between the intermediate prediction label predicted by the intermediate fault recognition model and the sample fault label based on the external fault label-historical maintenance database and a preset text similarity algorithm.

[0123] In this step, after determining the intermediate fault identification model, the historical maintenance data corresponding to the predicted label output by the intermediate fault identification model is determined from the external fault label-historical maintenance database, and then the preset text similarity algorithm is used to determine the similarity between the historical maintenance data and the sample fault label.

[0124] Among them, assuming that the sample fault label in the embodiment provided in this application is: engine failure, and the corresponding predictive maintenance suggestion is: check the engine oil and coolant. At this time, a preset text similarity algorithm (such as the BERTScore algorithm) is used to calculate the similarity of the two texts.

[0125] Sub-step 1032: Based on the similarity, the intermediate fault recognition model is trained to determine the trained standard fault recognition model corresponding to the historical vehicle.

[0126] In this step, the intermediate fault recognition model is trained according to the similarity between the two texts obtained in the previous step, and the trained standard fault recognition model corresponding to the historical vehicle is determined.

[0127] It can be understood that if the similarity score of the two texts is high, it means that the preset maintenance suggestion has a high similarity with the actual sample fault label, which means that the generated preset maintenance suggestion is more accurate. Otherwise, the suggestion deviation is large. After generating the preset maintenance suggestion, collect user feedback on the fault diagnosis results and predictive maintenance suggestions, evaluate user satisfaction and model practicality, and monitor the performance of the trained standard fault recognition model in actual applications in real time, graph response time, fault detection accuracy and other key indicators.

[0128] Here, the embodiments provided in this application may use different methods to train and optimize the standard fault identification model until a trained standard fault identification model is determined. The specific model training and optimization methods may be as follows:

[0129] a. Model structure adjustment: increase the number of multi-head self-attention heads:

[0130] According to the predicted label results, the model performed poorly in capturing global contextual relationships, so we tried to increase the number of multi-head self-attention heads in the Transformer architecture. By increasing the number of heads, the standard fault recognition model can pay attention to information at more locations in parallel, thereby improving its ability to capture global contextual relationships.

[0131] b. Loss function optimization: weighted cross entropy loss function:

[0132] Due to the problem of class imbalance in the data set (i.e., the number of samples in some sample fault categories is much larger than that in other categories), the present invention uses a weighted cross entropy loss function to optimize the standard fault recognition model. By assigning different weights to samples of different categories, the model pays more attention to categories with fewer samples, thereby improving the performance of the model on these categories.

[0133] c. Regularization and hyperparameter adjustment: L2 regularization and grid search:

[0134] In order to prevent the standard fault recognition model from overfitting, regularization is used to constrain the parameter size of the standard fault recognition model. At the same time, the grid search method is used to adjust the hyperparameters of the standard fault recognition model, such as learning rate and batch size. By traversing different hyperparameter combinations, the hyperparameter settings that make the performance of the standard fault recognition model optimal are found.

[0135] d. Continuous learning and feedback: Online learning and user feedback system:

[0136] The standard fault identification model is deployed online, and online learning technology is used to continuously update and optimize the standard fault identification model. By collecting real-time data and fine-tuning the standard fault identification model, the standard fault identification model can adapt to new predicted fault types and changing environmental conditions. At the same time, a user feedback system is established to collect user feedback on the maintenance suggestions generated by the standard fault identification model. These feedbacks can be used to guide further optimization and improvement of the model.

[0137] The method for determining a fault identification model provided in an embodiment of the present application, compared with the prior art, classifies the sample vehicle status data corresponding to the historical vehicles according to the fault type and fault level in turn, determines the sample fault label corresponding to each sample vehicle status data, and then inputs each sample vehicle status data and the corresponding target sample fault label into the initial fault identification model for training, determines the candidate fault identification model corresponding to the historical vehicle, and determines the trained standard fault identification model corresponding to the historical vehicle based on the candidate fault identification model, the preset text similarity algorithm and the preset error algorithm, which can provide more flexible and accurate early warning and identification of faults of different types of vehicles based on deep learning, and the embodiment provided by the present application can provide personalized fault analysis and maintenance suggestions for users and after-sales, and can dynamically adjust the fault detection and analysis strategy according to the actual usage and data characteristics of the vehicle, thereby optimizing the vehicle maintenance process and improving the safety and reliability of the vehicle.

[0138] See also Figure 3 , Figure 3 A flowchart of a method for determining a vehicle fault provided in an embodiment of the present application is shown in FIG. Figure 3 As shown, the method for determining a vehicle fault includes the following steps:

[0139] S301, inputting the target vehicle status data of the target vehicles in the batch to be detected into the trained standard fault recognition model to determine whether each of the target vehicles has a fault and the target fault type corresponding to the fault.

[0140] In this step, after the trained standard fault recognition model is determined, the target vehicle status data of the target vehicle in the batch to be tested is directly input into the standard fault recognition model to realize the recognition and prediction of the target vehicle fault and the target fault type.

[0141] Compared with the prior art, the method for determining vehicle faults provided in the embodiments of the present application more accurately and automatically predicts vehicle faults and can dynamically adjust the fault detection and analysis strategy according to the actual usage and data characteristics of the vehicle, thereby optimizing the vehicle maintenance process and improving the safety and reliability of the vehicle.

[0142] See also Figure 4 , Figure 4 A structural block diagram of a device for determining a fault identification model provided in an embodiment of the present application, such as Figure 4 As shown, the fault identification model determination device 400 includes:

[0143] The first determination module 410 is used to classify the sample vehicle status data corresponding to the historical vehicles according to the fault type and fault level, and determine the sample fault label corresponding to each of the historical vehicles, wherein the sample fault label is used to characterize the label including the fault type and fault level of the sample vehicle status data.

[0144] The second determination module 420 is used to input each of the sample vehicle status data and the corresponding target sample fault label into the initial fault recognition model for training, and determine the candidate fault recognition model corresponding to the historical vehicle.

[0145] The third determination module 430 is used to determine the trained standard fault recognition model corresponding to the historical vehicle based on the candidate fault recognition model, the preset text similarity algorithm and the preset error algorithm.

[0146] Optionally, the first determination module 410 determines the sample vehicle status data corresponding to the historical vehicle in the following manner, wherein the method for determining the fault identification model includes:

[0147] Get the initial vehicle status data corresponding to the historical vehicle.

[0148] Based on the initial vehicle state data and a preset local abnormality factor algorithm, first candidate vehicle state data corresponding to the historical vehicle is determined.

[0149] Based on the first candidate vehicle state data and a preset normalization algorithm, second candidate vehicle state data corresponding to the historical vehicle is determined.

[0150] Data expansion is performed based on the second candidate vehicle status data to determine sample vehicle status data corresponding to the historical vehicle.

[0151] Optionally, the second determining module 420 is specifically configured to:

[0152] Each sample vehicle status data and the corresponding target sample fault label are divided according to the preset training task, and the first sub-sample vehicle status data and the corresponding first target sub-sample fault label under different training tasks are determined.

[0153] Each first sub-sample vehicle status data and the corresponding first target sub-sample fault label are input into the initial fault recognition model for training to determine the training data under each of the training tasks.

[0154] Based on the training data under each of the training tasks, a candidate fault recognition model corresponding to the historical vehicle is determined.

[0155] Optionally, the second determining module 420 is further specifically configured to:

[0156] Each sample vehicle status data and the corresponding target sample fault label are segmented according to the preset vehicle model, and the second sub-sample vehicle status data and the corresponding second target sub-sample fault label under different models are determined.

[0157] The vehicle status data of each second sub-sample and the corresponding fault label of the second target sub-sample are input into the corresponding first initial fault sub-recognition model for training to determine the prediction type under each of the models.

[0158] Based on each of the prediction types, each of the first initial fault sub-identification models is integrated to determine a candidate fault identification model corresponding to the historical vehicle.

[0159] Optionally, the determining, based on the candidate fault recognition model, a preset text similarity algorithm, and a preset error algorithm, a trained standard fault recognition model corresponding to the historical vehicle includes:

[0160] Based on the candidate fault identification model and a preset error algorithm, an error between a candidate prediction label predicted by the candidate fault identification model and a sample fault label is determined.

[0161] Based on the error, the candidate fault recognition model is trained to determine an intermediate fault recognition model corresponding to the historical vehicle.

[0162] Based on the external fault label-historical maintenance database and a preset text similarity algorithm, determining the similarity between the intermediate prediction label predicted by the intermediate fault recognition model and the sample fault label;

[0163] Based on the similarity, the intermediate fault recognition model is trained to determine the trained standard fault recognition model corresponding to the historical vehicle.

[0164] The fault identification model determination device 400 provided in the embodiment of the present application, compared with the prior art, the embodiment provided by the present application classifies the sample vehicle status data corresponding to the historical vehicle according to the fault type and fault level in turn, determines the sample fault label corresponding to each sample vehicle status data, and then inputs each sample vehicle status data and the corresponding target sample fault label into the initial fault identification model for training, determines the candidate fault identification model corresponding to the historical vehicle, and determines the trained standard fault identification model corresponding to the historical vehicle based on the candidate fault identification model, the preset text similarity algorithm and the preset error algorithm, which can provide more flexible and accurate early warning and identification of faults of different types of vehicles based on deep learning, and the embodiment provided by the present application can provide personalized fault analysis and maintenance suggestions for users and after-sales, and can dynamically adjust the fault detection and analysis strategy according to the actual usage and data characteristics of the vehicle, thereby optimizing the vehicle maintenance process and improving the safety and reliability of the vehicle.

[0165] See also Figure 5 , Figure 5 A structural block diagram of a vehicle fault determination device 500 provided in an embodiment of the present application, such as Figure 5 As shown, the vehicle fault determination device 500 includes:

[0166] The determination module 510 is used to input the target vehicle status data of the target vehicles in the batch to be detected into the trained standard fault recognition model to determine whether each of the target vehicles has a fault and the target fault type corresponding to the fault.

[0167] Compared with the prior art, the vehicle fault determination device 500 provided in the embodiment of the present application realizes the prediction of vehicle faults in a more accurate and automated manner, and can dynamically adjust the fault detection and analysis strategy according to the actual usage and data characteristics of the vehicle, thereby optimizing the vehicle maintenance process and improving the safety and reliability of the vehicle.

[0168] See also Figure 6 , Figure 6 A schematic diagram of the structure of an electronic device provided in an embodiment of the present application is shown in FIG. Figure 6 As shown, the electronic device 600 includes a processor 610 , a memory 620 , and a bus 630 .

[0169] The memory 620 stores machine-readable instructions executable by the processor 610. When the electronic device 600 is running, the processor 610 communicates with the memory 620 via the bus 630. When the machine-readable instructions are executed by the processor 610, the above-mentioned Figure 1 to Figure 2 The method for determining the fault identification model in the method embodiment shown in the figure or Figure 3The specific implementation of the steps of the method for determining a vehicle fault in the method embodiment shown can be found in the method embodiment and will not be described in detail here.

[0170] The present application also provides a computer-readable storage medium on which a computer program is stored. When the computer program is executed by a processor, the computer program can execute the above-mentioned Figure 1 to Figure 2 The method for determining the fault identification model in the method embodiment shown in the figure or Figure 3 The specific implementation of the steps of the method for determining a vehicle fault in the method embodiment shown can be found in the method embodiment and will not be described in detail here.

[0171] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working processes of the systems, devices and units described above can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0172] It should be noted that in the above embodiments, the description of each embodiment has its own emphasis, and for parts that are not described in detail in a certain embodiment, reference can be made to the relevant descriptions of other embodiments.

[0173] Those skilled in the art will appreciate that the embodiments of the present application may be provided as methods, systems, or computer program products. Therefore, the present application may adopt the form of a complete hardware embodiment, a complete software embodiment, or an embodiment in combination with software and hardware. Moreover, the present application may adopt the form of a computer program product implemented in one or more computer-readable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) that contain computer-readable program code.

[0174] The present application is described with reference to the flowcharts and / or block diagrams of the methods, devices (systems), and computer program products according to the embodiments of the present application. It should be understood that each process and / or box in the flowchart and / or block diagram, as well as the combination of the processes and / or boxes in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded computer, or other programmable data processing device to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0175] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0176] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0177] An embodiment of the present application further provides a computer program product, which includes computer software instructions. When the computer software instructions are executed on a processing device, the processing device executes the process of a method for determining a fault identification model.

[0178] The computer program product includes one or more computer instructions. When the computer program instructions are loaded and executed on the computer, the process or function according to the embodiment of the present application is generated in whole or in part. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable devices. The computer instructions can be stored in a computer-readable storage medium, or transmitted from one computer-readable storage medium to another computer-readable storage medium. For example, the computer instructions can be transmitted from a website site, a computer, a server or a data center by wired (e.g., coaxial cable, optical fiber, digital subscriber line (digital subscriber line, DSL)) or wireless (e.g., infrared, wireless, microwave, etc.) mode to another website site, computer, server or data center. The computer-readable storage medium can be any available medium that a computer can store or a data storage device such as a server or a data center that includes one or more available media integration. The available medium can be a magnetic medium, (e.g., a floppy disk, a hard disk, a tape), an optical medium (e.g., a DVD), or a semiconductor medium (e.g., a solid state drive (SSD)), etc.

[0179] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working processes of the systems, devices and units described above can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0180] In the several embodiments provided in the present application, it should be understood that the disclosed devices, apparatuses and methods can be implemented in other ways. For example, the device embodiments described above are only schematic, for example, the division of units is only a logical function division, and there may be other division methods in actual implementation, such as multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be an indirect coupling or communication connection through some interfaces, devices or units, which can be electrical, mechanical or other forms.

[0181] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0182] In addition, each functional unit in each embodiment of the present application may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit. The above-mentioned integrated unit may be implemented in the form of hardware or in the form of software functional units.

[0183] 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 computer-readable storage medium. Based on this understanding, the technical solution of the present application 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, and the computer software product is stored in a storage medium, including a number of instructions to enable a computer device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the various embodiments of the present application. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (Read-Only Memory, ROM), random access memory (Random Access Memory, RAM), disk or optical disk and other media that can store program codes.

[0184] The above embodiments are only used to illustrate the technical solutions of the present application, rather than to limit them. Although the present application has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present application.

[0185] Although the preferred embodiments of this specification have been described, those skilled in the art may make additional changes and modifications to these embodiments once they have learned the basic creative concept. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments and all changes and modifications that fall within the scope of this specification.

[0186] Obviously, those skilled in the art can make various changes and modifications to this specification without departing from the spirit and scope of this specification. Thus, if these modifications and variations of this specification fall within the scope of the claims of this specification and their equivalents, this specification is also intended to include these modifications and variations.

Claims

1. A method for determining a fault identification model, characterized in that: The method for determining the fault identification model includes: Classify the sample vehicle status data corresponding to the historical vehicles according to the fault type and fault level, and determine the sample fault labels corresponding to each of the historical vehicles, wherein the sample fault labels are used to characterize the labels including the fault type and fault level of the sample vehicle status data; Inputting each of the sample vehicle status data and the corresponding target sample fault label into 42 for training to determine a candidate fault recognition model corresponding to the historical vehicle; Based on the candidate fault recognition model, a preset text similarity algorithm and a preset error algorithm, a trained standard fault recognition model corresponding to the historical vehicle is determined.

2. The method for determining a fault identification model according to claim 1, characterized in that: The sample vehicle status data corresponding to the historical vehicle is determined in the following manner, and the method for determining the fault identification model includes: Obtain the initial vehicle status data corresponding to the historical vehicle; Determining first candidate vehicle state data corresponding to the historical vehicle based on the initial vehicle state data and a preset local abnormality factor algorithm; Determining second candidate vehicle state data corresponding to the historical vehicle based on the first candidate vehicle state data and a preset normalization algorithm; Data expansion is performed based on the second candidate vehicle status data to determine sample vehicle status data corresponding to the historical vehicle.

3. The method for determining a fault identification model according to claim 1, characterized in that: The inputting each of the sample vehicle status data and the corresponding target sample fault label into the initial fault recognition model for training to determine the candidate fault recognition model corresponding to the historical vehicle comprises: Each sample vehicle status data and the corresponding target sample fault label are divided according to the preset training task, and the first sub-sample vehicle status data and the corresponding first target sub-sample fault label under different training tasks are determined; Inputting each first sub-sample vehicle state data and the corresponding first target sub-sample fault label into the initial fault recognition model for training, and determining the training data under each training task; Based on the training data under each of the training tasks, a candidate fault recognition model corresponding to the historical vehicle is determined.

4. The method for determining a fault identification model according to claim 1, characterized in that: The initial fault recognition model includes at least one first initial fault sub-recognition model, and the inputting of each of the sample vehicle status data and the corresponding target sample fault label into the initial fault recognition model for training to determine the candidate fault recognition model corresponding to the historical vehicle also includes: Each sample vehicle status data and the corresponding target sample fault label are divided according to the preset vehicle model, and the second sub-sample vehicle status data and the corresponding second target sub-sample fault label under different models are determined; Inputting each second sub-sample vehicle status data and the corresponding second target sub-sample fault label into each corresponding first initial fault sub-recognition model for training to determine the prediction type under each of the models; Based on each of the prediction types, each of the first initial fault sub-identification models is integrated to determine a candidate fault identification model corresponding to the historical vehicle.

5. The method for determining a fault identification model according to claim 1, characterized in that: The determining, based on the candidate fault recognition model, a preset text similarity algorithm, and a preset error algorithm, a trained standard fault recognition model corresponding to the historical vehicle includes: Based on the candidate fault identification model and a preset error algorithm, determining the error between the candidate prediction label predicted by the candidate fault identification model and the sample fault label; Based on the error, the candidate fault recognition model is trained to determine an intermediate fault recognition model corresponding to the historical vehicle; Based on the external fault label-historical maintenance database and a preset text similarity algorithm, determining the similarity between the intermediate prediction label predicted by the intermediate fault recognition model and the sample fault label; Based on the similarity, the intermediate fault recognition model is trained to determine the trained standard fault recognition model corresponding to the historical vehicle.

6. A method for determining vehicle faults, using the method for determining the fault identification model according to any one of claims 1 to 5, characterized in that: The method for determining a vehicle fault comprises: The target vehicle status data of the target vehicles in the batch to be tested are input into the trained standard fault recognition model to determine whether each of the target vehicles has a fault and the target fault type corresponding to the fault.

7. A device for determining a fault identification model, characterized in that: The device for determining the fault identification model comprises: A first determination module is used to classify the sample vehicle status data corresponding to the historical vehicles according to the fault type and the fault level, and determine the sample fault label corresponding to each of the historical vehicles, wherein the sample fault label is used to characterize the label including the fault type and the fault level of the sample vehicle status data; A second determination module is used to input each of the sample vehicle status data and the corresponding target sample fault label into an initial fault recognition model for training, and determine a candidate fault recognition model corresponding to the historical vehicle; The third determination module is used to determine the trained standard fault recognition model corresponding to the historical vehicle based on the candidate fault recognition model, the preset text similarity algorithm and the preset error algorithm.

8. A vehicle fault determination device, characterized in that: Using the trained standard fault recognition model as claimed in claim 6, the vehicle fault determination device comprises: The determination module is used to input the target vehicle status data of the target vehicles in the batch to be detected into the trained standard fault recognition model to determine whether each of the target vehicles has a fault and the target fault type corresponding to the fault.

9. An electronic device, characterized in that: include: A processor, a memory and a bus, wherein the memory stores machine-readable instructions executable by the processor, and when the electronic device is running, the processor and the memory communicate via the bus, and when the processor is running, the machine-readable instructions are executed by the steps of the method for determining the fault identification model as described in any one of claims 1 to 5 or the steps of the method for determining the vehicle fault as described in claim 6.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the computer program executes the steps of the method for determining a fault identification model as described in any one of claims 1 to 5 or the steps of the method for determining a vehicle fault as described in claim 6.

Images