Method, apparatus, device and program for determining severity of fault of vehicle

Abstract

The embodiment of the invention relates to a method, device, equipment and program for determining the severity of a fault of a vehicle. The method includes obtaining a set of fault codes related to a fault of a target vehicle. The method further includes determining a severity of the fault of the target vehicle using a data processing model related to the plurality of vehicles based on the set of fault codes, the data processing model determined based on a target knowledge graph, the target knowledge graph including a plurality of fault codes related to the plurality of vehicles and a plurality of severity for the fault. According to the embodiment of the invention, the severity of the fault of the vehicle can be quickly determined according to a group of fault codes, the fault processing efficiency is improved, and the user experience is improved.

Description

TECHNICAL FIELD

[0001] Embodiments of the present disclosure generally relate to the field of fault detection of vehicles, and in particular to a method, apparatus, device and program for determining the severity of a fault of a vehicle. BACKGROUND

[0002] With the rapid development of technology, the functions and performance of vehicles are also rapidly developing. With the continuous progress of new energy vehicle technology, various advanced vehicles, especially electric vehicles, are gradually becoming an important trend in the future development of vehicles. They not only reduce energy consumption, but also provide users with a comfortable experience.

[0003] With the increasing complexity of vehicle technology, vehicle faults have become a problem that cannot be ignored. Common vehicle faults include engine faults, brake system faults, electrical system faults, etc. Usually for a vehicle that has a fault, it needs to be detected by a maintenance personnel, and then the parts are replaced or some components are repaired to remove the fault. SUMMARY

[0004] Embodiments of the present disclosure provide a method, apparatus, device and program for determining the severity of a fault of a vehicle.

[0005] According to a first aspect of the present disclosure, a method for determining the severity of a fault of a vehicle is provided. The method comprises obtaining a set of fault codes related to a fault of a target vehicle. The method further comprises determining the severity of the fault of the target vehicle based on the set of fault codes by using a data processing model related to a plurality of vehicles, the data processing model being determined based on a target knowledge graph, the target knowledge graph comprising a plurality of fault codes related to the plurality of vehicles and a plurality of severities for the fault.

[0006] According to a second aspect of the present disclosure, an apparatus for determining the severity of a fault of a vehicle is provided. The apparatus comprises a fault code obtaining module configured to obtain a set of fault codes related to a fault of a target vehicle; and a target severity determining module configured to determine the severity of the fault of the target vehicle based on the set of fault codes by using a data processing model related to a plurality of vehicles, the data processing model being determined based on a target knowledge graph, the target knowledge graph comprising a plurality of fault codes related to the plurality of vehicles and a plurality of severities for the fault.

[0007] According to a third aspect of the present disclosure, a computing device is provided. The computing device comprises at least one processor; and a memory coupled to the at least one processor and having stored therein instructions that, when executed by the at least one processor, cause the computing device to perform the steps of the method of the first aspect of the present disclosure.

[0008] According to a fourth aspect of the present disclosure, a machine program product is provided. The machine program product comprises machine executable instructions, wherein the machine executable instructions are executed by a processor to implement the steps of the method in the first aspect of the present disclosure.

[0009] According to a fifth aspect of the present disclosure, a machine readable storage medium is provided. The machine readable storage medium has stored thereon machine executable instructions, wherein the machine executable instructions are executed by a processor to implement the steps of the method in the first aspect of the present disclosure. BRIEF DESCRIPTION OF DRAWINGS

[0010] Figure 1 schematic diagram illustrating an example environment in which computing devices and / or methods according to embodiments of the present disclosure can be implemented;

[0011] Figure 2 schematic diagram illustrating an example of determining severity of a vehicle according to embodiments of the present disclosure;

[0012] Figure 3 schematic diagram illustrating an example of workflow of an embedding-based link prediction algorithm according to embodiments of the present disclosure;

[0013] Figure 4 schematic diagram illustrating an example of data for a link prediction algorithm according to embodiments of the present disclosure;

[0014] Figure 5 flowchart of a method of determining severity of a fault of a vehicle according to embodiments of the present disclosure;

[0015] Figure 6 schematic diagram of an apparatus for determining severity of a fault of a vehicle according to embodiments of the present disclosure;

[0016] Figure 7 schematic block diagram of an example computing device suitable for use in implementing embodiments of the present disclosure.

[0017] In the various drawings, like or corresponding elements are denoted by like or corresponding reference numerals. DETAILED DESCRIPTION

[0018] Embodiments of the present disclosure will be described below in greater detail with reference to the accompanying drawings. While certain embodiments of the present disclosure are shown in the drawings, it is understood that the present disclosure can be embodied in various forms and should not be interpreted as being limited to the embodiments set forth herein, but rather these embodiments are provided so as to more thoroughly and completely understand the present disclosure. It is understood that the drawings and embodiments of the present disclosure are for exemplary purposes only and are not intended to limit the scope of protection of the present disclosure.

[0019] In the description of embodiments of the disclosure, the term "includes" and its conjugates are to be construed as open-ended, i.e., "including but not limited to". The term "based on" is to be construed as "based at least in part on". The term "one embodiment" or "an embodiment" is to be construed as "at least one embodiment". The terms "first", "second", etc. can refer to different or same objects. Other explicit or implicit definitions can also be included below.

[0020] As mentioned earlier, there are many problems that need to be solved in the process of handling vehicle failures. For example, during the driving of a vehicle, technical problems can be encountered, which can be solved by following predefined procedures at a repair point. However, in complex vehicle systems, unexpected error messages that have never been encountered before can occur. These unexpected problems can cause serious vehicle problems that need to be addressed immediately. When a combination of error messages occurs, it is usually necessary to first assess the severity of the failure.

[0021] However, in related solutions, the process of assessing the severity of the failure involves a lot of manual work, such as data extraction, conversion and loading. This situation can cause a considerable delay between the field problem and the development of a solution, sometimes up to several months. Therefore, this delay can cause the original equipment manufacturer to face significant loss of time and resources. The main reason for this delay is the challenge of data heterogeneity, as information is stored in isolation in different systems.

[0022] In order to quickly determine the severity of such a failure, one related method is to introduce additional hardware to monitor the running state of the vehicle, internal components and the surrounding environment. This hardware is used to detect alarms and abnormal situations when certain parameters exceed the set critical threshold. However, this method is operationally complex and requires complex hardware, and also requires the deployment of related algorithms on this hardware. Another related technology suggests deploying software in the cloud to continuously monitor the status of the vehicle. Although this eliminates the cost of additional hardware, it causes a delay in response time due to cloud-based data access, and generates operating costs related to cloud storage. Another related technology involves matching vehicle failure information with corresponding entries in a vehicle failure knowledge graph, and analyzing the co-occurrence of various vehicle failures. However, this technology is only used to identify failures that have occurred in multiple vehicles, and cannot assess the severity of the existing failure.

[0023] To this end, embodiments of the present disclosure provide a method for determining the severity of a fault of a vehicle. In the method, a computing device can quickly obtain a set of fault codes related to a fault of a target vehicle. After obtaining the set of fault codes, the computing device can determine the severity of the fault using a target knowledge graph related to a plurality of vehicles, where the target knowledge graph includes a plurality of fault codes related to the plurality of vehicles. In this way, the knowledge graph related to the fault codes is applied to the scenario when predicting the severity of the fault, thereby enabling the quick determination of the severity of the fault corresponding to the set of fault codes, improving the fault handling efficiency, and improving the user experience.

[0024] Embodiments of the present disclosure will be described in detail below with further reference to the drawings, in which Figure 1 An example environment in which computing devices and / or methods of embodiments of the present disclosure can be implemented is shown.

[0025] Figure 1 A schematic diagram of an example environment 100 in which a plurality of embodiments of the present disclosure can be implemented is shown. As Figure 1 shown, a computing device 106 can obtain a set of fault codes 102 of a target vehicle, and then determine the severity 108 of a fault of the target vehicle according to the set of fault codes 102. In some embodiments, the set of fault codes 102 are fault codes displayed on a display screen of the target vehicle when the target vehicle has a fault.

[0026] In some embodiments, the set of fault codes can include at least one type of fault code among a diagnostic trouble code (DTC), a fault byte, and a monitoring ID. In one example, the set of fault codes is composed of a diagnostic trouble code and a fault byte, e.g., information composed of these two codes indicates one or more faults. In another example, the set of fault codes includes a monitoring ID, e.g., one monitoring ID indicates one or more faults. In yet another example, the set of fault codes includes a diagnostic trouble code, a fault byte, and a monitoring ID. This set of fault codes forms an alarm signal for a user to determine the fault.

[0027] The computing device 106 includes, but is not limited to, a personal computer, a server computer, a handheld or laptop device, a mobile device such as a mobile phone, a personal digital assistant (PDA), a media player, etc., a multiprocessor system, a consumer electronic product, a minicomputer, a mainframe computer, a distributed computing environment including any of the above systems or devices, etc.

[0028] In some embodiments, when computing device 106 acquires a set of fault codes 102, it typically also acquires a data processing model, such as a machine learning model, related to a target knowledge graph 104 generated using historical data from multiple vehicles. In one example, the multiple vehicles may include a target vehicle, in which case the target knowledge graph does not include the fault severity of a set of fault codes for the target vehicle. In another example, the multiple vehicles do not include the target vehicle. The data processing model is trained using the target knowledge graph 104. The target knowledge graph 104 includes multiple fault codes related to the multiple vehicles. For example, which vehicle among the multiple vehicles has which type of fault code, and what the value of the fault code is. Additionally, the target knowledge graph 104 also includes the fault severity related to the fault codes in the multiple vehicles.

[0029] The data processing model trained using this target knowledge graph 104 can be used to predict the severity of faults indicated by fault codes on target vehicles. For example, by inputting relevant information about the target vehicle with fault codes into the data processing model, it can be predicted whether the vehicle's fault is serious or minor. If the vehicle's fault is serious, timely intervention is required, improving vehicle safety and enhancing the user experience.

[0030] The above example illustrates that severity includes two levels: severe and not severe, and is merely an example and not a specific limitation of this disclosure. In some embodiments, severity may include multiple levels. For example, severity may include three levels: not severe, moderately severe, and very severe.

[0031] By utilizing the knowledge graph related to fault codes, the severity of a fault corresponding to a set of fault codes for a vehicle can be quickly determined, thereby improving fault handling efficiency and enhancing user experience.

[0032] The above combination Figure 1 An example environment in which embodiments of this disclosure can be implemented is described. The following is in conjunction with... Figure 2 A schematic diagram illustrating Example 200 for determining the severity of a vehicle according to embodiments of the present disclosure.

[0033] To analyze the severity of vehicle faults, data from heterogeneous sources is needed, such as vehicle field data, analytical data, production data, and fault description data. Among them, vehicle field data mainly refers to various data generated during the actual vehicle operation, testing, or maintenance process. These data have important reference value for understanding the vehicle state, analyzing the cause of the fault, and formulating maintenance plans. Analytical data refers to some statistical information related to the vehicle, such as the statistics of the failure of a certain module of the vehicle or the statistics of the module usage or sales. Production data refers to various data generated during the automobile manufacturing process, which covers the entire production process from raw material procurement to final product off-line. Fault description data can include user descriptions of vehicle faults and other data. In addition, alarm signals or fault codes of the vehicle that have failed, such as diagnostic fault codes, fault bytes, and monitoring identifiers, are also obtained from one or more databases.

[0034] Then, using these different data sources and fault codes, at block 202, a knowledge graph for a plurality of vehicles is constructed, which includes the operating state of the vehicle. Vehicle field data is also input into the knowledge graph, which is the basis for applying a knowledge graph embedding (KGE) algorithm to learn representations and use a data processing model to predict links pointing to severity.

[0035] For the knowledge graph constructed at block 202, at block 204, the severity of the faults indicated by the plurality of fault codes also needs to be manually determined and input by technicians. Then, at block 206, the severity is combined into the knowledge graph and a triple data set for the combined knowledge graph is constructed. Then, at block 208, a knowledge graph embedding algorithm is used to learn node representations. The learned node representations are then applied to a link prediction algorithm 210 to train the link prediction algorithm. The link prediction algorithm 210 is a machine learning model. In some embodiments, the link prediction algorithm 210 is a neural network model. In some embodiments, the link prediction algorithm 210 is a rule-based model.

[0036] When using a trained link prediction algorithm or machine learning model, upon receiving a new set of fault codes for a target vehicle, the link prediction algorithm 210, combined with other vehicle information, predicts missing or new links in the knowledge graph at box 212, thereby determining the severity of the fault indicated by the new set of fault codes at box 214. Other vehicle information may include vehicle identification information, such as vehicle identification or vehicle type. The above examples are merely illustrative of this disclosure and not intended to limit its specific scope. In some embodiments, an association is established between the vehicle's identification information and each fault code in the set of fault codes. In one example, this association can be described by triples, such as (Vehicle 1 has fault code 1). This association is then transformed or adjusted to obtain the input to the link prediction algorithm.

[0037] This approach is flexible and adaptable. By capturing complex, nonlinear, and high-dimensional relationships within the dataset, KGE algorithms can adapt to the underlying data structure and discover complex patterns that rule-learning algorithms may struggle to find. Furthermore, their ability to generalize and learn continuously makes them well-suited for this use case, especially where new entities frequently emerge.

[0038] The method described above creates a knowledge graph that includes associations between faulty vehicles and multiple fault codes, as well as associations between faulty vehicles and the severity of their faults. By implementing machine learning models or link prediction algorithms on the knowledge graph, its construction and utilization can be greatly enhanced, aiding in the early detection and diagnosis of emerging technical problems. This approach significantly reduces manual intervention and helps accelerate problem-solving.

[0039] The above combination Figure 2 A schematic diagram illustrating an example of determining the severity of a vehicle according to embodiments of the present disclosure is shown below. Figure 3 A schematic diagram illustrating an example workflow of an embedded link prediction algorithm according to an embodiment of the present disclosure.

[0040] Figure 3An example 300 of the workflow of link prediction algorithms is shown. At the start stage 302, data preparation is performed in block 304, including data extraction, pre-processing, and organizing data into a format suitable for training and testing. The final dataset is split into training, testing, and validation sets, each consisting of Resource Description Framework (RDF) triples in the format of head entity, relation, tail entity (tab separated strings). For example, (monitoring identification entity, association relation, severity entity) and (another monitoring identification entity, association relation, fault byte entity). Embedding techniques are then applied to each set to convert entities and their relations into entity and relation embeddings in a vector space 306. These initial embeddings are the starting point for the training stage 308.

[0041] The main goal of link prediction is to evaluate how well the embeddings of the algorithm capture existing and potential links between entities in the knowledge graph. In the training stage 308 of the data processing model or link prediction algorithm, a target function 310 is utilized for the training operation 312. Typically, this target function can be represented as a likelihood function or a loss function, depending on the specific data processing model or link prediction algorithm. The target function is a key component of the training stage, which provides explicit guidance for optimizing the data processing model or link prediction algorithm to make accurate link predictions. In this process, specific links pointing to severity are focused on so that the algorithm can better diagnose vehicle technical issues. For example, the missing link to be predicted is: Appear_FaultByte(Car_03, 42) Appear_Dtc(Car_03, C4042)→?Diagnostic_criticality(Severity, Car_03), where Appear_FaultByte(Car_03, 42) indicates that Car_03 appears fault byte 42, and Appear_Dtc(Car_03, C4042) indicates that Car_03 appears diagnostic fault code C4042, to predict whether the fault of Car_03 is severe “?Diagnostic_criticality(Severity, Car_03)”.

[0042] At this stage, some techniques (including negative sampling 314, optimization 316, and regularization 318) are used to improve model performance. Specifically, negative sampling 314 creates contrastive negative examples to help the model distinguish between positive and negative relations. For example, by modifying positive samples to generate negative samples, the number of negative samples is increased to make training more reasonable. Optimization algorithms such as stochastic gradient descent are used in the optimization operation 316 to adjust the parameters of the embeddings; through the regularization operation 318, to prevent overfitting and maintain the generality of the learned predictions.

[0043] As mentioned above, a portion of the data is set aside for validation purposes. In the validation phase 320, the validation data 322 is used to evaluate the performance of the model during training and make necessary adjustments to avoid overfitting. Selected evaluation metrics 324 include Hits@k and Mean Reciprocal Rank (MRR) to evaluate different aspects of the link prediction algorithm or data processing model.

[0044] For the evaluation metric Hits@k, which represents the proportion of test triples that fall within the top k triples, for example, as represented by equation (1) below:

[0045]

[0046] where K test represents the set of test triples, |K test | represents the number of test sets, and rank(t) represents the link prediction rank of the tth triple.

[0047] In addition, for the evaluation metric MRR, which evaluates the quality of the predicted ranks by considering the reciprocal rank of the first correct prediction for each entity, it can be represented by equation (2) below:

[0048]

[0049] where Q corresponds to the predicted samples, rank i refers to the rank position of the first correct prediction for entity i.

[0050] Further, the model is also tested on a set of unseen examples to evaluate its generalization ability beyond the training data. For example, in the testing phase 326, the test data 328 is input to the data processing model or link prediction algorithm to predict the missing links at block 330. If the data processing model is a rule-based model, post-processing can be performed at operation 332 to make adjustments. Then at block 334, the data processing model or link prediction algorithm is evaluated again. After the data processing model or link prediction algorithm is trained through iterative fine-tuning and post-processing, the model can be deployed to practical applications in industrial scenarios. For example, in the application phase 336, the model or algorithm is deployed. The model can also be updated over time using new data, for example, at block 338, the model can also be continuously learned using newly received data to adapt to the changing graph dynamics, making it more robust in real-world scenarios. Then the model deployment is completed at block 340.

[0051] The above describes an example of the workflow of the embedding-based link prediction algorithm according to an embodiment of the present disclosure, and the following describes an example of the workflow of the data processing model according to an embodiment of the present disclosure. Figure 3 Figure 4 ​A schematic diagram illustrating an example of data for a link prediction algorithm according to an embodiment of the present disclosure.

[0052] Figure 4 Example 400 shown is a partial knowledge graph related to vehicle fault codes. For example... Figure 4 As shown, this part of the knowledge graph includes vehicles 414 and 418. Vehicle 414 has a diagnostic fault code 412, for example, a value of D1042, and a fault byte 420, for example, a value of 42, and the fault is diagnosed as severe (422). Vehicle 418 has a diagnostic fault code 416, for example, a value of C4042, and a fault byte 410, for example, a value of 10, and the fault is diagnosed as non-severe (426). Therefore, there are corresponding associations between the above nodes. In addition, there are also vehicle systems 402 and 404 in this subgraph, which are related to each other; this subgraph further includes vehicle parts 408 and 406. Vehicle system 402 has two subsystems, vehicle parts 408 and 406, and vehicle system 404 is related to vehicle parts 406. In addition, vehicle parts 406 have fault bytes 410 and 420. Vehicle parts 408 have diagnostic fault codes 412 and 416.

[0053] During operation, vehicle 424 issued an alarm message containing two types of fault codes: diagnostic fault code 416 and fault byte 420. A link can be established in the knowledge graph between vehicle 424 and diagnostic fault code 416 and fault byte 420. Next, the aforementioned data processing model can be used, at least leveraging the link between vehicle 424 and diagnostic fault code 416 and fault byte 420, to predict the severity of the fault in vehicle 424. Figure 4 The severity level shown has two levels: one corresponding to severe and the other to non-severe. Therefore, the data processing model can determine whether car 424 has a link with severe nodes or a key link with non-severe nodes (e.g., ...). Figure 4 (As shown by the dashed line). At this point, the prediction of the severity of the vehicle malfunction is transformed into the prediction of the links between nodes in the knowledge graph.

[0054] The above combination Figure 4 A schematic diagram illustrating an example of data for a link prediction algorithm according to embodiments of the present disclosure is described below, in conjunction with... Figure 5 A flowchart describing a method for determining the severity of a vehicle malfunction according to embodiments of the present disclosure is provided. This method can be used... Figure 1 Executed on the computing device 106 shown or any suitable computing device.

[0055] In the example 500, at block 502, the computing device obtains a set of fault codes related to a fault of a target vehicle. After the vehicle has a fault, a set of fault codes is usually provided to a display device of the vehicle.

[0056] In some embodiments, the knowledge graph related to the vehicle is also utilized when determining the fault. Thus, a knowledge graph for the vehicle can be generated at the computing device. When generating the knowledge graph, the computing device can obtain a plurality of fault codes related to a plurality of vehicles, e.g., from different sources of fault codes data. In addition, the computing device can also obtain associated data for the plurality of vehicles from a plurality of data sources. For example, at least one of vehicle field data, analysis data, production data, fault description data can be obtained from different data sources as the associated data. Then, the computing device utilizes the plurality of fault codes and the associated data to generate a target knowledge graph related to the plurality of vehicles. The above example is only used to describe the present disclosure, but not to limit the present disclosure. The knowledge graph can also be pre-generated at the computing device or other computing devices.

[0057] In some embodiments, when generating the target knowledge graph by utilizing the plurality of fault codes and the associated data, the computing device can first generate an initial knowledge graph related to the plurality of vehicles by utilizing the plurality of fault codes and the associated data. Since the severity of the fault is determined by the technician, the information obtained from the above data sources cannot determine the severity of the fault. Then, the technician can determine a plurality of severities of a plurality of faults in the plurality of vehicles according to the fault conditions. Thus, a corresponding severity can be determined for each fault or vehicle. Then, the computing device adds the plurality of severities to the initial knowledge graph and identifies the association between the severities and the corresponding vehicles or faults. Thus, the target knowledge graph is generated.

[0058] At block 504, the computing device determines the severity of the fault by utilizing a data processing model related to a plurality of vehicles based on the set of fault codes. The data processing model is determined based on a target knowledge graph, and the target knowledge graph includes a plurality of fault codes related to the plurality of vehicles. After obtaining the set of fault codes, the computing device determines the severity of the fault by utilizing the data processing model obtained from the target knowledge graph described above.

[0059] In some embodiments, in the process of determining the severity of the fault by utilizing the data processing model, the computing device obtains the data processing model trained by the target knowledge graph. Then, the computing device utilizes the data processing model to process the set of fault codes to determine the severity of the fault.

[0060] In some embodiments, in determining the severity of the fault, the computing device determines an association of the target vehicle with the fault codes in the set of fault codes. The computing device then applies the association to a data processing model to determine the severity of the fault. For example, a triple consisting of the vehicle and the fault codes is applied to the data processing model to determine the severity of the fault. The target knowledge graph includes a plurality of severities. Thus, by applying the association to the data processing model, the severity corresponding to the target vehicle can be determined from the plurality of severities.

[0061] In some embodiments, the computing device can pre-determine a mapping relationship between various information related to vehicles in the target knowledge graph and the severity of the fault, the various information including the fault codes. The computing device then determines the severity of the fault of the target vehicle by computing an association of the information of the target vehicle including the set of fault codes with the vehicles in the target knowledge graph. For example, the similarity of two vehicles is computed by the various information, and the severity of the vehicle in the target knowledge graph with high similarity is determined as the severity of the target vehicle or the severity of the fault of the target vehicle. The above example is only used to describe the present disclosure, and is not a specific limitation of the present disclosure.

[0062] In some embodiments, the computing device also trains the data processing model using the target knowledge graph, where the data processing model is a machine learning model. In this process, the computing device generates an embedding representation corresponding to the target knowledge graph. For example, the computing device first converts the target knowledge graph into triples, and then uses a knowledge graph embedding algorithm to generate an embedding representation corresponding to the target knowledge graph. Next, the computing device trains the machine learning model using the embedding representation.

[0063] In some embodiments, in the training process of the machine learning model, the computing device applies the embedding representation to the machine learning model to predict the severity for a sample vehicle in the plurality of vehicles. Then, the computing device computes an objective function according to the predicted severity of the sample vehicle and a sample severity of the sample vehicle. The machine learning model is then optimized using the objective function. In one example, the objective function can be a loss function. In another example, the objective function is a likelihood function.

[0064] In some embodiments, after training the machine learning model using the training data, the machine learning model can also be further verified using validation data. The validation data can also come from the target knowledge graph. Then, the computing device applies the validation data to the machine learning model to compute an evaluation metric for the trained machine learning model. For example, the Hits@k and the mean reciprocal rank of the machine learning model. Then, the computing device evaluates the trained machine learning model according to the evaluation metric.

[0065] In some embodiments, the computing device also uses test data to test the machine learning model. During testing, the machine learning model can be post-processed as needed, and further evaluation can be performed. Thus, the trained machine learning model can be fine-tuned as required.

[0066] In some embodiments, after deploying the machine learning model, data from the application process can be added to a knowledge graph to further train the machine learning model. This allows the machine learning model to be continuously adjusted and improved, enabling it to better predict the severity of faults. Additionally, this method can predict the severity of fault codes corresponding to problems in a vehicle steering system.

[0067] By utilizing the knowledge graph related to fault codes, the severity of a vehicle's fault can be quickly determined using a set of fault codes, thereby improving fault handling efficiency and enhancing the user experience.

[0068] Figure 6 A schematic diagram of an apparatus for determining the severity of a vehicle malfunction according to an embodiment of the present disclosure is further shown. The apparatus 600 can be applied to a computing device 106, which may include multiple modules for performing tasks such as... Figure 5 The corresponding steps in method 500 discussed herein. For example... Figure 6 As shown, the device 600 includes a fault code acquisition module 602, configured to acquire a set of fault codes related to a fault of a target vehicle; and a severity determination module 604, configured to determine the severity of a fault of the target vehicle based on a set of fault codes using a data processing model related to multiple vehicles. The data processing model is determined based on a target knowledge graph, which includes multiple fault codes related to multiple vehicles and multiple severity levels for the fault.

[0069] In some embodiments, the apparatus 600 further includes: a fault code and associated data acquisition module, configured to acquire multiple fault codes related to multiple vehicles and associated data for multiple vehicles from multiple data sources; and a target knowledge graph generation module, configured to generate a target knowledge graph related to multiple vehicles based on the multiple fault codes and associated data.

[0070] In some embodiments, the associated data includes at least one of the following: vehicle field data, analysis data, production data, and fault description data.

[0071] In some embodiments, the target knowledge graph generation module includes: an initial knowledge graph generation module configured to generate an initial knowledge graph related to the plurality of vehicles based on the plurality of fault codes and the association data; a plurality of severity determination modules configured to determine a plurality of severities of a plurality of faults in the plurality of vehicles; and a target knowledge graph obtaining module configured to obtain the target knowledge graph by adding the plurality of severities to the initial knowledge graph.

[0072] In some embodiments, the severity determination module 604 includes: a data processing model obtaining module configured to obtain a data processing model for the target knowledge graph; and a severity determination module configured to determine the severity of the fault of the target vehicle by applying the set of fault codes to the data processing model.

[0073] In some embodiments, the severity determination module includes: an association relationship determination module configured to determine an association relationship between the target vehicle and the fault codes in the set of fault codes; and a relationship application module configured to determine the severity of the fault of the target vehicle by applying the association relationship to the data processing model.

[0074] In some embodiments, the association relationship determination module includes: an identification information determination module configured to determine identification information of the target vehicle; and an association relationship generation module configured to generate the association relationship between the identification information and the fault codes.

[0075] In some embodiments, the target knowledge graph includes a plurality of severities, and the relationship application module includes: a data processing model application module configured to determine the severity corresponding to the target vehicle from the plurality of severities by applying the association relationship to the data processing model.

[0076] In some embodiments, the data processing model is a machine learning model, and the apparatus 600 further includes: an embedding representation generation module configured to generate an embedding representation corresponding to the target knowledge graph; and a training module configured to train the machine learning model using the embedding representation.

[0077] In some embodiments, the embedding representation generation module includes: an algorithm application module configured to generate the embedding representation corresponding to the target knowledge graph using a knowledge graph embedding algorithm.

[0078] In some embodiments, the training module includes: a representation application module configured to apply the embedding representation to the machine learning model to predict the severity for a sample vehicle in the plurality of vehicles; a target function determination module configured to determine a target function based on the predicted severity of the sample vehicle and a sample severity of the sample vehicle; and an optimization module configured to optimize the machine learning model based on the target function.

[0079] In some embodiments, the apparatus 600 further comprises a validation data obtaining module configured to obtain validation data for the trained machine learning model; an evaluation metric determining module configured to determine an evaluation metric for the trained machine learning model by applying the validation data to the machine learning model; and an evaluation module configured to evaluate the trained machine learning model based on the evaluation metric.

[0080] In some embodiments, the set of fault codes comprises at least one of: a diagnostic fault code, a fault byte, and a watchdog identification.

[0081] Figure 7 A schematic block diagram of an example computing device 700 that can be used to implement embodiments of the present disclosure is shown. Figure 1 The computing device 106 in FIG. 1 can be implemented with the computing device 700. As shown, the computing device 700 includes a central processing unit (CPU) 701, which can perform various suitable actions and processes according to computer program instructions stored in a read-only memory (ROM) 702 or loaded into a random access memory (RAM) 703 from a storage unit 708. Various programs and data required by the computing device 700 for operation can also be stored in the RAM 703. The CPU 701, the ROM 702, and the RAM 703 are connected to each other via a bus 704. An input / output (I / O) interface 705 is also connected to the bus 704.

[0082] A plurality of components in the computing device 700 are connected to the I / O interface 705, which can include, for example: an input unit 706, such as a keyboard, a mouse, etc.; an output unit 707, such as various types of displays, a speaker, etc.; a storage unit 708, such as a magnetic disk, an optical disk, etc.; and a communication unit 709, such as a network card, a modem, a wireless communication transceiver, etc. The communication unit 709 allows the computing device 700 to exchange information / data with other devices via a computer network, such as the Internet, and / or various telecommunication networks.

[0083] The various processes and processes described above, such as the method 500, can be performed by the processing unit 701. For example, in some embodiments, the method 500 can be implemented as a computer software program tangibly embodied in a machine-readable medium, such as the storage unit 708. In some embodiments, part or all of the computer program can be loaded and / or installed on the computing device 700 via the ROM 702 and / or the communication unit 709. When the computer program is loaded into the RAM 703 and executed by the CPU 701, one or more actions of the method 500 described above can be performed.

[0084] The present disclosure can be a method, apparatus, system, and / or computer program product. The computer program product can include a computer-readable storage medium (or media) having computer readable program instructions thereon for performing various aspects of the present disclosure.

[0085] The computer-readable storage medium can be a tangible device that can retain and store instructions for use by an instruction execution device. The computer-readable storage medium can be, for example, but is not limited to, an electronic storage device, a magnetic storage device, an optical storage device, an electromagnetic storage device, a semiconductor storage device, or any suitable combination of the foregoing. More specific examples (a non-exhaustive list) of the computer-readable storage medium include the following: a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or Flash memory), a static random access memory (SRAM), and any suitable combination of the foregoing. Computer-readable storage media is not to be construed as being transitory signals per se, such as radio waves or other freely propagating electromagnetic waves, electromagnetic waves propagating through a waveguide or other transmission media (e.g., light pulses passing through a fiber-optic cable), or electrical signals transmitted through a wire.

[0086] Computer readable program instructions described herein can be downloaded to respective computing / processing devices from a computer readable storage medium or to an external computer or external storage device via a network, for example, the Internet, a local area network, a wide area network and / or a wireless network. The network can comprise copper transmission cables, optical transmission fibers, wireless transmission, routers, firewalls, switches, gateway computers and / or edge servers. A network adapter card or network interface in each computing / processing device receives computer readable program instructions from the network and forwards the computer readable program instructions for storage in a computer readable storage medium within the respective computing / processing device.

[0087] Computer readable program instructions for carrying out operations of the present disclosure can be assembler instructions, instruction-set-architecture (ISA) instructions, machine instructions, machine dependent instructions, microcode, firmware instructions, state-setting data, or either source code or object code written in any combination of one or more programming languages, including an object oriented programming language such as Smalltalk, C++ or the like and conventional procedural programming languages, such as the "C" programming language or similar programming languages. The computer readable program instructions can execute entirely on the user's computer, partly on the user's computer, as a stand-alone software package, partly on the user's computer and partly on a remote computer or entirely on the remote computer or server. In the latter scenario, the remote computer can be connected to the user's computer through any type of network, including a local area network (LAN) or a wide area network (WAN), or the connection can be made to an external computer (for example, through the Internet using an Internet Service Provider). In some embodiments, electronic circuitry including, for example, programmable logic circuitry, field-programmable gate array (FPGA), or programmable logic array (PLA) can execute the computer readable program instructions by utilizing state information of the computer readable program instructions to personalize the electronic circuitry, in order to perform aspects of the present disclosure.

[0088] The computer readable program instructions can also be loaded onto a computer, other programmable data processing apparatus, or other device to cause a series of operational steps to be performed on the computer, other programmable apparatus or other device to produce a computer implemented process such that the instructions which execute on the computer or other programmable apparatus provide processes for implementing the functions / acts specified in the flowchart and / or block diagram block or blocks.

[0089] These computer readable program instructions can be provided to a processor of a general purpose computer, special purpose computer, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, create means for implementing the functions / acts specified in the flowchart and / or block diagram block or blocks. These computer readable program instructions can also be stored in a computer readable storage medium that can include random access memory (RAM), read only memory (ROM), electrically erasable programmable read only memory (EEPROM), flash memory or other data storage device. When the computer readable program instructions are loaded into the computer and other programmable data processing apparatus, a series of operational steps are implemented that provide processes such that the computer or other programmable apparatus provide processes for implementing the functions / acts specified in the flowchart and / or block diagram block or blocks.

[0090] The computer readable program instructions can also be loaded onto a computer, other programmable data processing apparatus, or other device to cause a series of operational steps to be performed on the computer, other programmable data processing apparatus or other device to produce a computer implemented process such that the instructions which execute on the computer, other programmable data processing apparatus, or other device implement the functions / acts specified in the flowchart and / or block diagram block or blocks.

[0091] The flow diagrams and the block diagrams in the drawings are presented to illustrate the architecture, functionality, and operation of possible implementations of systems, methods and computer program products according to various embodiments of the present disclosure. In this regard, each block in the flow diagrams and the block diagrams can represent a module, segment, or portion of instructions, which comprises one or more executable instructions for implementing the specified logic functions. In some alternative implementations, the functions noted in the blocks can occur out of the order noted in the figures. For example, two blocks shown in succession may, in fact, be executed substantially concurrently, or the blocks can sometimes be executed in the reverse order, depending upon the functionality involved. It will also be noted that each block of the block diagrams and / or flow diagrams, and combinations thereof, can be implemented by special purpose hardware-based systems that perform the specified functions or acts, or combinations of special purpose hardware and

[0092] Embodiments of the present disclosure have been described above, and the description is intended to be illustrative of the embodiments and not restrictive. Many modifications and variations of the described embodiments are possible and are within the scope of the disclosure. The selection of terms is intended to best describe the principles of the embodiments, practical application, or technical improvements over the technology found in the market, or to enable others skilled in the art to understand the embodiments disclosed herein.

Claims

1. A method (500) for determining a severity of a fault of a vehicle, comprising: obtaining (502) a set of fault codes related to a fault of a target vehicle; and determining (504) a severity of the fault of the target vehicle based on the set of fault codes using a data processing model related to a plurality of vehicles, the data processing model being determined based on a target knowledge graph, the target knowledge graph comprising a plurality of fault codes related to the plurality of vehicles and a plurality of severities for faults.

2. The method (500) of claim 1, further comprising: obtaining the plurality of fault codes related to the plurality of vehicles and associated data for the plurality of vehicles from a plurality of data sources; and generating the target knowledge graph related to the plurality of vehicles based on the plurality of fault codes and the associated data.

3. The method (500) of claim 2, wherein the associated data comprises at least one of: vehicle field data, analytics data, production data, fault description data.

4. The method (500) of claim 2, wherein generating the target knowledge graph related to the plurality of vehicles based on the plurality of fault codes and the associated data comprises: generating an initial knowledge graph related to the plurality of vehicles based on the plurality of fault codes and the associated data; determining the plurality of severities for a plurality of faults in the plurality of vehicles; and obtaining the target knowledge graph by adding the plurality of severities to the initial knowledge graph.

5. The method (500) of claim 1, wherein determining the severity of the fault of the target vehicle based on the set of fault codes using a data processing model related to a plurality of vehicles comprises: obtaining a data processing model for the target knowledge graph; and determining the severity of the fault of the target vehicle by applying the set of fault codes to the data processing model.

6. The method (500) of claim 5, wherein determining the severity of the fault of the target vehicle by applying the set of fault codes to the data processing model comprises: determining an association of the target vehicle with a fault code in the set of fault codes; and determining the severity of the fault of the target vehicle by applying the association to the data processing model.

7. The method (500) of claim 6, wherein determining the association of the target vehicle with a fault code in the set of fault codes comprises: determining identification information of the target vehicle; and generating the association of the identification information with the fault code.

8. The method (500) of claim 6, wherein determining the severity of the fault of the target vehicle by applying the association to the data processing model comprises: determining a severity corresponding to the target vehicle from the plurality of severities by applying the association to the data processing model. ​ ​ ​ ​ ​ ​ 9. The method (500) of claim 5, wherein the data processing model is a machine learning model, and the method further comprises: generating an embedding representation corresponding to the target knowledge graph; and training the machine learning model with the embedding representation.

10. The method (500) of claim 9, wherein generating an embedding representation corresponding to the target knowledge graph comprises: generating an embedding representation corresponding to the target knowledge graph with a knowledge graph embedding algorithm.

11. The method (500) of claim 9, wherein training the machine learning model with the embedding representation comprises: applying the embedding representation to the machine learning model to predict a severity for a sample vehicle of the plurality of vehicles; determining an objective function based on the predicted severity of the sample vehicle and a sample severity of the sample vehicle; and optimizing the machine learning model based on the objective function.

12. The method (500) of claim 11, further comprising: obtaining validation data for the trained machine learning model; determining an evaluation metric for the trained machine learning model by applying the validation data to the machine learning model; and evaluating the trained machine learning model based on the evaluation metric.

13. The method (500) of claim 1, wherein the set of fault codes comprises at least one of: a diagnostic fault code, a fault byte, and a monitor identification.

14. An apparatus (600) for determining a severity of a fault of a vehicle, comprising: a fault code obtaining module (602) configured to obtain a set of fault codes related to a fault of a target vehicle; and a severity determining module (604) configured to determine a severity of the fault of the target vehicle based on the set of fault codes with a data processing model related to a plurality of vehicles, the data processing model being determined based on a target knowledge graph, the target knowledge graph comprising a plurality of fault codes and a plurality of severities for faults related to the plurality of vehicles.

15. A computing device (700), comprising: at least one processor; and a memory coupled to the at least one processor and having stored thereon instructions that, when executed by the at least one processor, cause the computing device to perform the method of any one of claims 1-13.

16. A computer program product comprising computer executable instructions executed by a processor to implement the method of any one of claims 1 to 13. ​ ​ ​ ​ ​

Images