Vehicle fault diagnosis and maintenance evaluation method based on artificial intelligence

Abstract

The invention discloses a vehicle fault diagnosis and maintenance evaluation method based on artificial intelligence, and the method comprises the steps: deploying a vehicle-mounted sensor network, collecting a plurality of heterogeneous data, and processing the heterogeneous data; based on the processed heterogeneous data, constructing a fault detection model, and performing fault judgment according to a result of the fault detection model; and constructing a vehicle knowledge graph, and generating a maintenance evaluation scheme for a result of the fault detection model. By deploying the advanced vehicle-mounted sensor network and establishing the comprehensive historical maintenance database, comprehensive acquisition of multi-source heterogeneous data is realized, and a rich and reliable data basis is provided for subsequent analysis. And secondly, an edge computing technology and a multi-level adaptive fusion strategy are adopted, and an improved Dempster-Shafer evidence theory and a Kalman filter are combined, so that efficient real-time processing and fusion of heterogeneous data are realized, and the accuracy and the real-time performance of data processing are remarkably improved.

Description

Technical Field

[0001] The present invention relates to the technical field of vehicle intelligent diagnosis and maintenance, and in particular to a vehicle fault diagnosis and maintenance evaluation method based on artificial intelligence. Background Art

[0002] With the rapid development of the automobile industry and the continuous improvement of its intelligence, vehicle fault diagnosis and maintenance assessment technology has also evolved from traditional manual experience judgment to data-driven intelligent diagnosis. Early fault diagnosis mainly relied on the experience of mechanical engineers and simple diagnostic instruments, which was often inefficient and had limited accuracy when dealing with complex faults. Subsequently, fault diagnosis methods based on expert systems have been widely used, which simulate the diagnostic process of human experts by encoding expert knowledge into a rule base. However, the adaptability and scalability of this method are severely limited when faced with new faults or complex systems. In recent years, with the rapid development of artificial intelligence technology, fault diagnosis methods based on machine learning have begun to emerge. These methods can learn fault modes from a large amount of historical data and improve the accuracy and efficiency of diagnosis.

[0003] Despite this, existing vehicle fault diagnosis and maintenance assessment technologies still have some significant shortcomings. First, most systems fail to fully utilize the multi-source heterogeneous data generated during vehicle operation, which affects the comprehensiveness and accuracy of the diagnostic results. Secondly, traditional fault detection models are often static and lack the ability to adapt to the dynamic operating status of the vehicle in real time, making it difficult to cope with complex and changing driving environments and usage conditions. Furthermore, existing maintenance assessment schemes are usually generated based on preset rules or simple statistical models, which fail to fully consider the individual differences and historical maintenance records of vehicles, resulting in insufficient pertinence and effectiveness of maintenance recommendations. In addition, most systems lack structured representation and utilization of knowledge, and cannot effectively integrate and apply the empirical knowledge of domain experts, thereby limiting the reasoning ability and interpretability of the system. Summary of the invention

[0004] The present invention is proposed in view of the problems existing in the above-mentioned vehicle fault diagnosis and maintenance assessment technology.

[0005] Therefore, the problem to be solved by the present invention is how to provide a vehicle fault diagnosis and maintenance assessment technology based on artificial intelligence.

[0006] In order to solve the above technical problems, the present invention provides the following technical solutions: In a first aspect, an embodiment of the present invention provides a vehicle fault diagnosis and maintenance assessment method based on artificial intelligence, which comprises: Deploy a vehicle-mounted sensor network, collect multiple heterogeneous data, and process the heterogeneous data; Building a fault detection model based on the processed heterogeneous data, and performing fault judgment based on the results of the fault detection model; A vehicle knowledge graph is constructed, and a maintenance assessment plan is generated based on the results of the fault detection model.

[0007] As a preferred solution of the vehicle fault diagnosis and maintenance assessment method based on artificial intelligence of the present invention, wherein: the heterogeneous data is subjected to multi-source heterogeneous data processing according to the evidence theory framework and the filter; The theoretical framework of evidence is as follows: ; in, and are the weights of the first and second data sources at time t, respectively. m1 and m2 are the trust functions of the two independent evidence sources on the hypothesis, respectively. A is a set of possible hypotheses. B and C are subsets of the basic probability distribution, representing different hypotheses on the proposition.

[0008] As a preferred solution of the vehicle fault diagnosis and maintenance assessment method based on artificial intelligence of the present invention, the filter fuses the rapidly changing sensor data as follows: ; in, is the predicted value of the system state at time k, The estimated value of the system state at time k-1, is the dynamic weight, is the estimated value of the system state at time k-1, is the error covariance matrix, is the estimated value of the covariance matrix of the state estimation error at time k−1, is the Kalman gain, is the observed value at time k, The mapping of the predicted value at time k in the measurement space, A, B, H, Q, R, I are the state transfer matrix, control input matrix, observation matrix, process noise covariance matrix, observation noise covariance matrix and identity matrix respectively. is the state update result of the improved Kalman filter, is the updated error covariance matrix.

[0009] As a preferred solution of the vehicle fault diagnosis and maintenance assessment method based on artificial intelligence of the present invention, the processing of the heterogeneous data further includes: According to historical data and current environmental conditions, the preliminary weight of each data source is calculated according to the dynamic adjustment weight formula, and filters are applied to each data source separately to obtain preliminary state predictions and update results. According to the preliminary results of the filter, the weight of each data source is adjusted to obtain dynamic weights, and the dynamic weights are used to combine evidence. The improved evidence combination results are used to further optimize the data source state transfer matrix, control input matrix, observation matrix, process noise covariance matrix and observation noise covariance matrix. The final fusion result is obtained by combining the data source states and weights after multiple optimizations.

[0010] As a preferred solution of the vehicle fault diagnosis and maintenance assessment method based on artificial intelligence of the present invention, wherein: the fault detection model is a dual-stream neural network architecture; The first-stream neural network processes temporally and spatially correlated sensor data, while the second-stream neural network processes data with strong sequentiality; The outputs of the first-stream neural network and the second-stream neural network are dynamically fused through an attention mechanism.

[0011] As a preferred solution of the vehicle fault diagnosis and maintenance assessment method based on artificial intelligence of the present invention, wherein: the fault judgment includes: If the predicted probability of the first fault is greater than the high probability threshold and the uncertainty is less than the low uncertainty threshold LU, the fault is directly reported, and it is recommended to carry out the corresponding maintenance immediately, and a detailed description of the fault and the possible impact are provided; if the uncertainty is greater than or equal to the low uncertainty threshold but less than the medium uncertainty threshold, the fault is reported and it is recommended to carry out maintenance as soon as possible; if there is a second fault and the probability is greater than the high probability threshold, and the probability of the first fault minus the probability of the second fault is less than the probability difference threshold, the second fault is detected, and if the probability of the first fault minus the probability of the second fault is greater than the probability difference threshold, the second fault cannot be determined; If there is no fault with a probability greater than the high probability threshold, but there are multiple faults greater than the medium probability threshold, if multiple faults greater than the medium probability threshold are less than the probability difference threshold or the uncertainty is greater than the medium uncertainty threshold, multiple fault types are reported. If the uncertainty is greater than the high uncertainty threshold, a warning is given that multiple faults exist. Otherwise, a detailed description of the fault type is provided and each fault is checked one by one according to the probability size.

[0012] As a preferred solution of the vehicle fault diagnosis and maintenance assessment method based on artificial intelligence of the present invention, wherein: the vehicle knowledge graph includes: When a specific fault symptom is input, the message is propagated in the vehicle knowledge graph, the importance of different nodes and edges is dynamically adjusted using the attention mechanism, and the most likely fault cause is located through multi-hop reasoning; After the cause of the fault is located, similar situations are retrieved based on the case-based reasoning engine. If similar situations are found, similar cases are used to provide reference for current maintenance decisions. If no similar situations are found, a maintenance plan is generated based on various factors of the vehicle.

[0013] In the second aspect, the present invention is to further solve the safety problems existing in the vehicle fault diagnosis and maintenance assessment method based on artificial intelligence. The embodiment provides a vehicle fault diagnosis and maintenance assessment system based on artificial intelligence, which includes: A data processing module, used to deploy a vehicle-mounted sensor network, collect multiple heterogeneous data, and process the heterogeneous data; A fault judgment module is used to build a fault detection model based on the processed heterogeneous data and perform fault judgment based on the results of the fault detection model; The maintenance assessment module is used to construct a vehicle knowledge graph and generate a maintenance assessment plan based on the results of the fault detection model.

[0014] In a third aspect, an embodiment of the present invention provides a computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: when the computer program is executed by the processor, any step of the artificial intelligence-based vehicle fault diagnosis and maintenance assessment method as described in the first aspect of the present invention is implemented.

[0015] In a fourth aspect, an embodiment of the present invention provides a computer-readable storage medium having a computer program stored thereon, wherein: when the computer program is executed by a processor, it implements any step of the artificial intelligence-based vehicle fault diagnosis and maintenance assessment method as described in the first aspect of the present invention.

[0016] The beneficial effect of the present invention is that by deploying advanced vehicle-mounted sensor networks and establishing a comprehensive historical maintenance database, the comprehensive collection of multi-source heterogeneous data is realized, providing a rich and reliable data foundation for subsequent analysis. Secondly, by adopting edge computing technology and multi-level adaptive fusion strategy, combined with the improved Dempster-Shafer evidence theory and Kalman filter, efficient real-time processing and fusion of heterogeneous data are realized, and the accuracy and real-time performance of data processing are significantly improved. Furthermore, by designing an innovative dual-stream neural network architecture, combined with 3D-CNN and LSTM, effective processing of spatiotemporal related data and sequential data is achieved, greatly improving the accuracy of fault diagnosis and prediction. In addition, a complex vehicle maintenance knowledge graph is constructed, and combined with graph neural network technology, accurate fault cause location and personalized maintenance plan generation are achieved. Finally, by introducing technologies such as meta-learning, reinforcement learning and case reasoning, the adaptability, decision-making ability and interpretability of the system are further enhanced. This comprehensive method not only significantly improves the accuracy, efficiency and personalization of vehicle fault diagnosis and maintenance evaluation, but also provides strong support for preventive maintenance and full life cycle management of vehicles, and has important practical application value. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following briefly introduces the drawings required for describing the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work. Among them: Figure 1 Flowchart of the artificial intelligence-based vehicle fault diagnosis and maintenance assessment method. DETAILED DESCRIPTION

[0018] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, the specific implementation methods of the present invention are described in detail below in conjunction with the accompanying drawings.

[0019] In the following description, many specific details are set forth to facilitate a full understanding of the present invention, but the present invention may also be implemented in other ways different from those described herein, and those skilled in the art may make similar generalizations without violating the connotation of the present invention. Therefore, the present invention is not limited to the specific embodiments disclosed below.

[0020] Secondly, the term "one embodiment" or "embodiment" as used herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The term "in one embodiment" that appears in different places in this specification does not necessarily refer to the same embodiment, nor does it refer to a separate or selective embodiment that is mutually exclusive with other embodiments.

[0021] Embodiment 1: Reference Figure 1 , which is the first embodiment of the present invention, provides a vehicle fault diagnosis and maintenance assessment method based on artificial intelligence, comprising the following steps: First, design and deploy an advanced vehicle sensor network, including various types of sensors, such as engine temperature sensors, vibration sensors and tire pressure monitors, which realize real-time data collection through a carefully designed low-power, high-reliability network architecture. At the same time, establish a comprehensive vehicle maintenance history database, integrate maintenance records from vehicle manufacturers, repair shops and users, and design a standardized data format to ensure the consistency and availability of information.

[0022] In order to capture the details of driving behavior, a dedicated data collection system is developed, including in-vehicle behavior monitoring devices and mobile applications, which record operations such as acceleration, braking, steering, and driving mileage, frequency, and road conditions. In order to achieve real-time processing and preliminary analysis of data, an edge computing platform is deployed in the in-vehicle system to run data preprocessing algorithms for noise filtering and outlier detection.

[0023] Furthermore, in the process of real-time processing and fusion of heterogeneous data using edge computing technology, a multi-level, adaptive fusion strategy is adopted. First, the system pre-processes the input from multiple heterogeneous data sources such as the vehicle sensor network, vehicle historical maintenance records, and driving behavior data, including data cleaning, normalization, and time synchronization.

[0024] Subsequently, an improved Dempster-Shafer evidence theory framework is applied to deal with data uncertainty and conflict, in which each data source is assigned a dynamic weight that is adjusted in real time based on the historical reliability of the data and current environmental conditions. On this basis, a Kalman filter is introduced to optimize the prediction and update of time series data, especially for rapidly changing sensor data.

[0025] The performance of edge computing nodes is taken into account, and the kernel function K(x, y) is used to evaluate and optimize the allocation of computing resources. In addition, information entropy H(X) is introduced to quantify the uncertainty of the data, and the hyperbolic tangent function is used to smooth the impact of data processing speed. This comprehensive approach can not only effectively fuse heterogeneous data from different sources, but also dynamically adjust the fusion strategy according to real-time conditions, ultimately forming a highly accurate and real-time updated vehicle "digital twin" model. The whole process is iterative, and the system will continue to learn and optimize the fusion algorithm to adapt to different vehicle types, usage environments, and data characteristics, thereby providing the most accurate vehicle status representation.

[0026] For example, in a feasible embodiment, the improved Dempster-Shafer evidence theory framework is as follows: Assuming that the weight of each data source is dynamically adjusted based on historical reliability and current environmental conditions, the weight function wi(t) is introduced here to represent the weight of the i-th data source at time t. The Kalman filter is used to optimize time series data. Specifically, the combination rules are as follows: ; in, and are the weights of the first and second data sources at time t, respectively. m1 and m2 are the trust functions of the two independent evidence sources on the hypothesis, respectively. A is a set of possible hypotheses. B and C are subsets of the basic probability distribution, representing different hypotheses on the proposition.

[0027] Specifically, the dynamic adjustment weight is shown in the following formula: ; in, represents the historical reliability of the i-th data source, represents the impact of the current environmental conditions on the i data sources, is the adjustment parameter.

[0028] Furthermore, combined with the prediction and update steps of the Kalman filter, the improved fusion strategy for rapidly changing sensor data is as follows: ; in, is the predicted value of the system state at time k, The estimated value of the system state at time k-1, is the dynamic weight, is the estimated value of the system state at time k-1, is the error covariance matrix, is the estimated value of the covariance matrix of the state estimation error at time k−1, is the Kalman gain, is the observed value at time k, The mapping of the predicted value at time k in the measurement space, A, B, H, Q, R, I are the state transfer matrix, control input matrix, observation matrix, process noise covariance matrix, observation noise covariance matrix and identity matrix respectively. is the state update result of the improved Kalman filter, is the updated error covariance matrix.

[0029] Specifically, according to historical data and current environmental conditions, the initial weight of each data source is calculated according to the dynamic adjustment weight formula, and the Kalman filter is applied to each data source separately to obtain the initial state prediction and update results. According to the preliminary results of the Kalman filter, the weight of each data source is adjusted to obtain the dynamic weight, and the dynamic weight is used to combine evidence. The improved evidence combination results are used to further optimize the state transfer matrix, control input matrix, observation matrix, process noise covariance matrix and observation noise covariance matrix of the data source, and the final fusion result is obtained by combining the data source states and weights after multiple optimizations. Through the above multi-level nested interactive logic, complex fusion and optimization processing of heterogeneous data sources can be achieved to ensure the accuracy and real-time performance of data processing.

[0030] Furthermore, the system makes full use of the heterogeneous data sources that have been fused and optimized in the previous stage to build a highly accurate and adaptive deep learning model. First, the system receives heterogeneous data that has been processed and fused in real time by edge computing technology. These data include real-time data from the vehicle sensor network, long-term trend information from the vehicle's historical maintenance records, and driving behavior data. In the data preprocessing stage, these fused data will be further standardized and time-aligned to ensure that data from different sources can be seamlessly input into the deep learning model.

[0031] Subsequently, an innovative two-stream neural network architecture was designed, in which one stream uses a 3D convolutional neural network (3D-CNN) to process spatiotemporally related sensor data, such as engine vibration spectrum, temperature distribution, etc., to extract high-level spatial-temporal features; the other stream uses a long short-term memory network (LSTM) to process highly sequential data, such as historical maintenance records and driving behavior data, to capture long-term dependencies.

[0032] The outputs of the two streams are dynamically fused through an attention mechanism, which enables the model to adaptively focus on key information in different data sources. In order to deal with the possible uncertainty in the fused data, Bayesian deep learning technology is introduced to estimate the posterior distribution of model parameters using variational inference, thereby providing reliable uncertainty estimates.

[0033] The model also integrates a multi-task learning framework to simultaneously perform current fault diagnosis and future fault risk prediction. Through residual connections and gating mechanisms, the model can effectively utilize information at different time scales to improve the accuracy and robustness of predictions. Finally, the output of the model will pass through a post-processing module based on a graph neural network (GNN), which uses the relationship graph between the various system components of the vehicle to further optimize the fault diagnosis results and provide more interpretable fault location and propagation analysis. This deep learning method that fully integrates heterogeneous data sources not only greatly improves the accuracy of fault identification and prediction, but also significantly enhances the system's adaptability to complex and dynamic vehicle environments, laying a solid foundation for realizing intelligent and personalized preventive maintenance.

[0034] For example, in one feasible embodiment, the two-stream neural network architecture is as follows: 3D-CNN stream: This part processes the sensor data that is related to time and space. 3D-CNN can capture spatial and temporal patterns in the data. For example, it may identify a specific vibration pattern that often occurs before a vehicle is about to break down. LSTM stream: This part processes data with strong sequentiality. LSTM can capture long-term dependencies. For example, it may find a trend of gradually shortening maintenance intervals in the last three times, which may indicate that the vehicle performance is declining. Specifically, the formula is expressed as follows: ; The specific expression of the two-stream neural network architecture is as follows: ; in, F 3D-CNN is the output feature of the 3D convolutional neural network, F LSTM is the output feature of the long short-term memory network, Attention is the attention mechanism function, MLP is the multi-layer perceptron, α is the attention weight coefficient, G is the nonlinear activation function, H is the Kullback-Leibler divergence, q φ(w) is the variational posterior distribution of the model parameter w, p(w) is the prior distribution of the model parameter w, λ is the entropy regularization coefficient, S is the entropy function, ReLU is the rectified linear unit activation function, σ is the sigmoid activation function, W 3D is the 3D convolution kernel, b 3D is the 3D convolution bias, W f is the weight matrix of LSTM, b f is the bias vector of LSTM, x spatial-temporalis the temporal and spatial related sensor data, x sequential The data has strong sequentiality.

[0035] Specifically, the attention mechanism function is shown as follows: ; in, W Q , W K , W V is the weight matrix in the attention mechanism, d k is the scaling factor of the attention mechanism.

[0036] ; ; in, F 3D-CNN is the output feature of the 3D convolutional neural network, F LSTM is the output feature of the long short-term memory network.

[0037] Further, define the key parameters: High probability threshold (HP), medium probability threshold (MP), low probability threshold (LP), high uncertainty threshold (HU), medium uncertainty threshold (MU), low uncertainty threshold (LU), probability difference threshold (PD).

[0038] If the predicted probability of the first fault is greater than the high probability threshold and the uncertainty is less than the low uncertainty threshold LU, the fault is directly reported, and it is recommended to carry out the corresponding maintenance immediately, and a detailed description of the fault and possible impact is provided. If the uncertainty is greater than or equal to the low uncertainty threshold but less than the medium uncertainty threshold, the fault is reported and it is recommended to carry out maintenance as soon as possible; if there is a second fault and the probability is greater than the high probability threshold, and the probability of the first fault minus the probability of the second fault is less than the probability difference threshold, the second fault is detected. If the probability of the first fault minus the probability of the second fault is greater than the probability difference threshold, the second fault cannot be determined.

[0039] If there is no fault with a probability greater than the high probability threshold, but there are multiple faults greater than the medium probability threshold, if multiple faults greater than the medium probability threshold are less than the probability difference threshold or the uncertainty is greater than the medium uncertainty threshold, multiple fault types are reported. If the uncertainty is greater than the high uncertainty threshold, a warning is given that multiple faults exist. Otherwise, a detailed description of the fault type is provided and each fault is checked one by one according to the probability size.

[0040] If the probability of failure is less than the medium probability threshold, and the probability of existence of failure is greater than the low probability threshold, the vehicle has a minor failure. If the failure is safety-related (brakes, etc.), the problem is reported.

[0041] Furthermore, a large and complex vehicle maintenance knowledge graph was constructed, which not only includes the hierarchical structure of the vehicle system, the relationship between the components, common fault symptoms and their corresponding possible causes, but also integrates a large number of historical maintenance records and expert experience. This knowledge graph transforms unstructured maintenance manuals, technical documents and expert opinions into structured knowledge representation through ontology modeling and natural language processing technology. When a specific fault symptom is input, the message is propagated on the graph, and the attention mechanism is used to dynamically adjust the importance of different nodes and edges, and then the possible cause of the fault is located through multi-hop reasoning.

[0042] In order to improve the accuracy of reasoning, a meta-learning module is also introduced to enable the model to quickly adapt to different vehicle models and emerging fault modes. After locating the cause of the fault, the system activates a decision-making module based on reinforcement learning, which takes into account multiple factors such as the vehicle's usage history, current status, maintenance costs, and future usage plans to generate a personalized maintenance plan. This plan not only includes specific maintenance steps, but also provides priority recommendations and expected repair effects.

[0043] In order to further improve the interpretability and credibility of the system, a case-based reasoning engine is integrated, which can retrieve similar situations from historical maintenance cases to provide reference and evidence for current maintenance decisions. Finally, the system also designs an interactive interface that allows human experts to review and adjust the maintenance plan generated by AI to achieve intelligent decision-making in human-machine collaboration. This method of combining knowledge graphs and deep learning not only greatly improves the accuracy of fault diagnosis and the personalization of maintenance plans, but also significantly improves the efficiency and reliability of the maintenance process, providing strong support for preventive maintenance and full life cycle management of vehicles.

[0044] This embodiment also provides a vehicle fault diagnosis and maintenance assessment system based on artificial intelligence, including: A data processing module, used to deploy a vehicle-mounted sensor network, collect multiple heterogeneous data, and process the heterogeneous data; A fault judgment module is used to build a fault detection model based on the processed heterogeneous data and perform fault judgment based on the results of the fault detection model; The maintenance assessment module is used to construct a vehicle knowledge graph and generate a maintenance assessment plan based on the results of the fault detection model.

[0045] This embodiment also provides a computer device, which is suitable for the case of an artificial intelligence-based vehicle fault diagnosis and maintenance assessment method, including: a memory and a processor; the memory is used to store computer-executable instructions, and the processor is used to execute computer-executable instructions to implement the artificial intelligence-based vehicle fault diagnosis and maintenance assessment method proposed in the above embodiment.

[0046] The computer device may be a terminal, and the computer device includes a processor, a memory, a communication interface, a display screen and an input device connected through a system bus. Among them, the processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and a computer program. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The communication interface of the computer device is used to communicate with an external terminal in a wired or wireless manner, and the wireless manner can be achieved through WIFI, an operator network, NFC (near field communication) or other technologies. The display screen of the computer device may be a liquid crystal display screen or an electronic ink display screen, and the input device of the computer device may be a touch layer covered on the display screen, or a key, trackball or touchpad provided on the housing of the computer device, or an external keyboard, touchpad or mouse, etc.

[0047] This embodiment also provides a storage medium on which a computer program is stored. When the program is executed by a processor, the vehicle fault diagnosis and maintenance assessment method based on artificial intelligence as proposed in the above embodiment is implemented; the storage medium can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as static random access memory (Static Random Access Memory, referred to as SRAM), electrically erasable programmable read-only memory (Electrically Erasable Programmable Read-Only Memory, referred to as EEPROM), erasable programmable read-only memory (Erasable Programmable Read Only Memory, referred to as EPROM), programmable read-only memory (Programmable Red-Only Memory, referred to as PROM), read-only memory (Read-Only Memory, referred to as ROM), magnetic storage, flash memory, magnetic disk or optical disk.

[0048] Embodiment 2: This is the second embodiment of the present invention. In order to further verify the beneficial effects of the present invention, experimental data of a vehicle fault diagnosis and maintenance assessment method based on artificial intelligence is provided.

[0049] To verify the effectiveness of the present invention, 100 cars of different models, years of use and mileage were selected as test objects. First, an advanced on-board sensor network containing 20 sensors was installed on each car, including engine temperature sensor, vibration sensor, tire pressure monitor, oil pressure sensor, etc. These sensors are connected to the on-board edge computing unit via low-power Bluetooth 5.0 technology to achieve real-time data collection and preliminary processing. At the same time, a database containing the maintenance records of these 100 vehicles in the past five years was established, including more than 4,500 standardized maintenance records.

[0050] To capture driving behavior, a dedicated mobile application was developed and installed on the driver's smartphone to record parameters such as acceleration, braking force, steering angle, etc.

[0051] In addition, the driving route and road condition information are recorded through the on-board GPS module. All data are sampled at a frequency of 5Hz and transmitted to the cloud server in real time through the 4G network.

[0052] In terms of data processing, the original data is first preprocessed, including removing outliers (using the 3σ rule), filling missing values ​​(using the forward filling method), and standardization (using Z-score standardization). Then, the improved Dempster-Shafer evidence theory framework is applied to fuse multi-source data. Specifically, the historical reliability of each data source is calculated based on its accuracy over the past 30 days, and the weight is dynamically adjusted based on the current environmental conditions (such as weather, road conditions, etc.). For example, on rainy days, the weight of the tire pressure sensor will be reduced accordingly.

[0053] Next, a Kalman filter is used to process the rapidly changing sensor data. The state transfer matrix A and the observation matrix H are obtained by analyzing historical data. The process noise covariance matrix Q and the observation noise covariance matrix R are determined according to the sensor specifications and actual test results. Through iterative optimization, a stable state estimate is finally obtained.

[0054] In terms of deep learning model construction, a two-stream neural network architecture is adopted. The 3D-CNN stream uses four 3D convolutional layers, each followed by BatchNormalization and ReLU activation functions, and finally a fully connected layer.

[0055] The LSTM stream contains 2 layers of bidirectional LSTM with 128 units each, followed by an attention layer. The outputs of the two streams are fused through a multi-layer perceptron (MLP) with 3 hidden layers with 256, 128, and 64 neurons respectively.

[0056] To evaluate the model performance, 100 vehicles were randomly divided into a training set (80 vehicles) and a test set (20 vehicles), and an experiment was conducted for 6 months.

[0057] During this period, all actual faults were recorded and compared with the model's predictions. At the same time, the results of traditional rule-based diagnostic methods and single data source machine learning methods were recorded as a control, as follows: Table 1 Experimental data table Vehicle number Actual failure Model predictions Rule-based predictions Single-source machine learning predictions Prediction accuracy Reduced maintenance costs Reduced downtime V001 3 3 2 2 100% 25% 30% V002 2 2 1 2 100% 20% 25% V003 4 4 3 3 100% 30% 35% V004 1 1 1 0 100% 15% 20% V005 3 3 2 2 100% 22% 28% V006 2 2 1 1 100% 18% 23% By analyzing the above table data, it can be clearly seen that the vehicle fault diagnosis and maintenance assessment method based on artificial intelligence proposed in the present invention has significant advantages over the traditional method: First, in terms of fault prediction accuracy, the model of the present invention showed excellent performance. For all six test vehicles, the model was able to accurately predict the number of actual faults, achieving a prediction accuracy of 100%. In contrast, traditional rule-based methods and machine learning methods based on a single data source have varying degrees of prediction bias in most cases.

[0058] For example, for vehicle V003, four failures actually occurred, while the traditional method and the single data source method only predicted three failures respectively, showing obvious shortcomings.

[0059] Secondly, the present invention performs well in reducing maintenance costs. Data shows that the use of models for fault diagnosis and maintenance assessment can reduce maintenance costs by an average of 15% to 30%. This significant cost saving is mainly due to the accurate prediction ability of the model, which enables maintenance work to be carried out more targeted and avoids unnecessary inspections and repairs. For example, for vehicle V003, the maintenance cost reduction reached 30%, which is a considerable figure in industrial applications.

[0060] Furthermore, the present invention also demonstrates obvious advantages in reducing vehicle downtime: data show that the use of the model can reduce downtime by an average of 20% to 35%. This improvement is of great significance for improving vehicle utilization efficiency and reducing operational losses. In particular, for vehicles V003 and V005, the downtime was reduced by 35% and 28% respectively, which can bring significant economic benefits to commercial fleets or public transportation systems.

[0061] In addition, by comparing the data of different vehicles, it can be found that the model of the present invention has shown stable high performance when dealing with situations with different fault complexities. Whether it is vehicle V004 with only one fault or vehicle V003 with four faults, the model can accurately predict, which shows the robustness and adaptability of the method.

[0062] In summary, the artificial intelligence-based vehicle fault diagnosis and maintenance assessment method proposed in the present invention has demonstrated significant technical advantages and economic benefits in practical applications. It not only greatly improves the accuracy of fault diagnosis, but also reduces maintenance costs and reduces vehicle downtime through precise maintenance assessment. This innovative method provides a new solution for vehicle preventive maintenance and intelligent management, and is expected to have a broad and far-reaching impact in the automotive industry and transportation field.

[0063] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical solutions of the present invention, which should all be included in the scope of the claims of the present invention.

Claims

1. A vehicle fault diagnosis and maintenance assessment method based on artificial intelligence, characterized in that: include: Deploy a vehicle-mounted sensor network, collect multiple heterogeneous data, and process the heterogeneous data; Building a fault detection model based on the processed heterogeneous data, and performing fault judgment based on the results of the fault detection model; A vehicle knowledge graph is constructed, and a maintenance assessment plan is generated based on the results of the fault detection model.

2. The vehicle fault diagnosis and maintenance assessment method based on artificial intelligence as claimed in claim 1, characterized in that: Performing multi-source heterogeneous data processing on the heterogeneous data according to the evidence theory framework and filters; The theoretical framework of evidence is as follows: ； in, and are the weights of the first and second data sources at time t, respectively. m1 and m2 are the trust functions of the two independent evidence sources on the hypothesis, respectively. A is a set of possible hypotheses. B and C are subsets of the basic probability distribution, representing different hypotheses on the proposition.

3. The vehicle fault diagnosis and maintenance assessment method based on artificial intelligence as claimed in claim 2, characterized in that: The filter fuses the rapidly changing sensor data as follows: ； in, is the predicted value of the system state at time k, The estimated value of the system state at time k-1, is the dynamic weight, is the estimated value of the system state at time k-1, is the error covariance matrix, is the estimated value of the covariance matrix of the state estimation error at time k−1, is the Kalman gain, is the observed value at time k, The mapping of the predicted value at time k in the measurement space, A, B, H, Q, R, I are the state transfer matrix, control input matrix, observation matrix, process noise covariance matrix, observation noise covariance matrix and identity matrix respectively. is the state update result of the improved Kalman filter, is the updated error covariance matrix.

4. The vehicle fault diagnosis and maintenance assessment method based on artificial intelligence as claimed in claim 3, characterized in that: Processing the heterogeneous data further includes: According to historical data and current environmental conditions, the preliminary weight of each data source is calculated according to the dynamic adjustment weight formula, and filters are applied to each data source separately to obtain preliminary state predictions and update results. According to the preliminary results of the filter, the weight of each data source is adjusted to obtain dynamic weights, and the dynamic weights are used to combine evidence. The improved evidence combination results are used to further optimize the data source state transfer matrix, control input matrix, observation matrix, process noise covariance matrix and observation noise covariance matrix. The final fusion result is obtained by combining the data source states and weights after multiple optimizations.

5. The vehicle fault diagnosis and maintenance assessment method based on artificial intelligence as claimed in claim 4, characterized in that: The fault detection model is a two-stream neural network architecture; The first-stream neural network processes temporally and spatially correlated sensor data, while the second-stream neural network processes data with strong sequentiality; The outputs of the first-stream neural network and the second-stream neural network are dynamically fused through an attention mechanism.

6. The vehicle fault diagnosis and maintenance assessment method based on artificial intelligence as claimed in claim 5, characterized in that: The fault judgment includes: If the predicted probability of the first fault is greater than the high probability threshold and the uncertainty is less than the low uncertainty threshold LU, the fault is directly reported, and it is recommended to carry out the corresponding maintenance immediately, and a detailed description of the fault and the possible impact are provided; if the uncertainty is greater than or equal to the low uncertainty threshold but less than the medium uncertainty threshold, the fault is reported and it is recommended to carry out maintenance as soon as possible; if there is a second fault and the probability is greater than the high probability threshold, and the probability of the first fault minus the probability of the second fault is less than the probability difference threshold, the second fault is detected, and if the probability of the first fault minus the probability of the second fault is greater than the probability difference threshold, the second fault cannot be determined; If there is no fault with a probability greater than the high probability threshold, but there are multiple faults greater than the medium probability threshold, if multiple faults greater than the medium probability threshold are less than the probability difference threshold or the uncertainty is greater than the medium uncertainty threshold, multiple fault types are reported. If the uncertainty is greater than the high uncertainty threshold, a warning is given that multiple faults exist. Otherwise, a detailed description of the fault type is provided and each fault is checked one by one according to the probability size.

7. The vehicle fault diagnosis and maintenance assessment method based on artificial intelligence as claimed in claim 6, characterized in that: The vehicle knowledge graph includes: When a specific fault symptom is input, the message is propagated in the vehicle knowledge graph, the importance of different nodes and edges is dynamically adjusted using the attention mechanism, and the most likely fault cause is located through multi-hop reasoning; After the cause of the fault is located, similar situations are retrieved based on the case-based reasoning engine. If similar situations are found, similar cases are used to provide reference for current maintenance decisions. If no similar situations are found, a maintenance plan is generated based on various factors of the vehicle.

8. A vehicle fault diagnosis and maintenance assessment system based on artificial intelligence, based on the vehicle fault diagnosis and maintenance assessment method based on artificial intelligence according to any one of claims 1 to 7, characterized in that: include: A data processing module, used to deploy a vehicle-mounted sensor network, collect multiple heterogeneous data, and process the heterogeneous data; A fault judgment module is used to build a fault detection model based on the processed heterogeneous data and perform fault judgment based on the results of the fault detection model; The maintenance assessment module is used to construct a vehicle knowledge graph and generate a maintenance assessment plan based on the results of the fault detection model.

9. A computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the steps of the vehicle fault diagnosis and maintenance assessment method based on artificial intelligence according to any one of claims 1 to 7 are implemented.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the vehicle fault diagnosis and maintenance assessment method based on artificial intelligence according to any one of claims 1 to 7 are implemented.

Images