Vehicle fault diagnosis method based on OBD data

Abstract

The invention belongs to the technical field of vehicle fault diagnosis, and discloses a vehicle fault diagnosis method based on OBD data. The method comprises the following steps that a cloud data center constructs an artificial intelligence model and deploys the artificial intelligence model to all OBD systems; the OBD system is used for performing data structure conversion on the real-time OBD data by using a data structure conversion model; the OBD system is used for performing knowledge mapping on the converted real-time OBD data by using a named entity and entity relationship recognition model according to a vehicle fault knowledge graph; the OBD system performs vehicle fault diagnosis by using a vehicle fault diagnosis model according to the mapped real-time OBD data; and the OBD system is used for generating a fault maintenance strategy by using a fault maintenance strategy generation model according to a real-time vehicle fault diagnosis result. According to the invention, the problems of poor universality, poor fault diagnosis accuracy and low intelligent degree in the prior art are solved.

Description

Technical Field

[0001] The present invention belongs to the technical field of vehicle fault diagnosis, and particularly relates to a vehicle fault diagnosis method based on OBD data. Background Art

[0002] With the rapid development of the automotive industry, the accuracy and efficiency of vehicle fault diagnosis have become the key to ensuring the safe operation of vehicles. The On-Board Diagnostics (OBD) system, which is used to monitor the vehicle status and control emissions, ensures the normal operation of the vehicle emission system by continuously monitoring key components such as the engine, catalytic converter, and oxygen sensor. Once an anomaly is detected, the system records the fault information and related codes and issues a warning through the fault light to alert the driver. OBD data is obtained through the vehicle's OBD interface, and these data not only include fault codes but also cover rich vehicle operating status information, including various operating condition data of the vehicle, such as fuel pressure, engine air flow, vehicle speed, etc.

[0003] Although the OBD system has been applied to a certain extent in vehicle fault diagnosis, the existing vehicle fault diagnosis technologies still have the following defects:

[0004] 1) Poor generality: The OBD data formats of different vehicle models and different manufacturers are different, resulting in inconsistent data structures. It is difficult for existing technologies to uniformly process these heterogeneous data, which limits the generality of the OBD system;

[0005] 2) Poor accuracy of fault diagnosis: Existing fault diagnoses often rely on a single algorithm or simple rules and are difficult to handle complex and changeable vehicle faults. The generalization ability and robustness of a single algorithm are insufficient, and misdiagnosis or missed diagnosis is likely to occur;

[0006] 3) Low level of intelligence: In existing technologies, the generation of fault repair strategies mainly relies on manual experience, lacking intelligence and automation. This leads to low efficiency in generating repair strategies and it is difficult to ensure the optimality of the strategies. Summary of the Invention

[0007] In order to solve the problems of poor generality, poor accuracy of fault diagnosis, and low level of intelligence existing in the prior art, the purpose of the present invention is to provide a vehicle fault diagnosis method based on OBD data.

[0008] The technical solution adopted by the present invention is as follows:

[0009] A vehicle fault diagnosis method based on OBD data, comprising the following steps:

[0010] The cloud data center constructs a data structure conversion model, a vehicle fault knowledge graph, a named entity and entity relationship recognition model, a vehicle fault diagnosis model, and a fault repair strategy generation model, and deploys them to all OBD systems;

[0011] The OBD system collects the real-time OBD data of the current vehicle, and uses the data structure conversion model to perform data structure conversion on the real-time OBD data to obtain the converted real-time OBD data;

[0012] The OBD system, according to the vehicle fault knowledge graph, uses the named entity and entity relationship recognition model to perform knowledge mapping on the converted real-time OBD data to obtain the mapped real-time OBD data;

[0013] The OBD system, according to the mapped real-time OBD data, uses the vehicle fault diagnosis model to perform vehicle fault diagnosis to obtain the real-time vehicle fault diagnosis result;

[0014] The OBD system, according to the real-time vehicle fault diagnosis result, uses the fault repair strategy generation model to generate a fault repair strategy to obtain the real-time fault repair strategy.

[0015] Furthermore, the cloud data center constructs a data structure conversion model, a vehicle fault knowledge graph, a named entity and entity relationship recognition model, a vehicle fault diagnosis model, and a fault repair strategy generation model, and deploys them to all OBD systems, including the following steps:

[0016] The cloud data center collects a number of historical OBD data and a number of historical vehicle fault knowledge, and performs preprocessing to obtain a number of preprocessed historical OBD data and a number of preprocessed vehicle fault knowledge;

[0017] Perform data structure conversion on a number of preprocessed historical OBD data to obtain a number of converted historical OBD data;

[0018] According to a number of preprocessed vehicle fault knowledge, use natural language processing algorithms to construct a named entity and entity relationship recognition model, and obtain a vehicle fault knowledge graph;

[0019] According to a number of preprocessed historical OBD data and the corresponding converted historical OBD data, use deep learning algorithms to construct a data structure conversion model;

[0020] According to a number of converted historical OBD data, use deep learning algorithms to construct a vehicle fault diagnosis model, and generate a number of historical vehicle fault diagnosis results;

[0021] According to a number of historical vehicle fault diagnosis results, use reinforcement learning algorithms to construct a fault repair strategy generation model, and generate a number of historical fault repair strategy generation experiences;

[0022] Deploy the data structure conversion model, vehicle fault knowledge graph, named entity and entity relationship recognition model, vehicle fault diagnosis model, and fault repair strategy generation model to all OBD systems.

[0023] Furthermore, the named entity and entity relationship recognition model is constructed based on the BERT-CRF-SVM algorithm, and the named entity and entity relationship recognition model includes a semantic feature extraction module constructed based on the BERT algorithm, a named entity recognition module constructed based on the CRF algorithm, and an entity relationship recognition module constructed based on the SVM algorithm, which are connected in sequence.

[0024] Furthermore, according to a number of preprocessed vehicle fault knowledge, use natural language processing algorithms to construct a named entity and entity relationship recognition model, and obtain a vehicle fault knowledge graph, including the following steps:

[0025] Use the BERT-CRF-SVM algorithm to construct an initial named entity and entity relationship recognition model;

[0026] According to a number of preprocessed vehicle fault knowledge, optimize and train the initial named entity and entity relationship recognition model to obtain a final named entity and entity relationship recognition model, and generate a number of knowledge named entities and a number of knowledge entity relationships;

[0027] According to a number of knowledge named entities and a number of knowledge entity relationships, conduct knowledge graph construction to obtain a vehicle fault knowledge graph.

[0028] Furthermore, the data structure conversion model is constructed based on the LSTM-cGAN algorithm, and the data structure conversion model includes a sequence feature extraction module constructed based on the LSTM algorithm and a data structure conversion module constructed based on the cGAN algorithm, which are connected in sequence. The data structure conversion module includes a generator and a discriminator connected in sequence, and both the generator and the discriminator are connected to the sequence feature extraction module.

[0029] Furthermore, according to a number of preprocessed historical OBD data and corresponding converted historical OBD data, use deep learning algorithms to construct a data structure conversion model, including the following steps:

[0030] Use the LSTM-cGAN algorithm to construct an initial data structure conversion model; the initial data structure conversion model includes an initial sequence feature extraction module and an initial data structure conversion module;

[0031] According to the first loss function of the initial generator and the second loss function of the initial discriminator in the initial data structure conversion module, obtain a comprehensive loss function;

[0032] Optimize the initial sequence feature extraction module based on several preprocessed historical OBD data to obtain the final sequence feature extraction module, and generate several historical OBD data sequence features;

[0033] Optimize the initial generator based on several historical OBD data sequence features to obtain the optimized generator, and generate several transformed generated OBD data;

[0034] Optimize the initial discriminator based on several transformed historical OBD data and the corresponding transformed generated OBD data sequences to obtain the optimized discriminator, and generate several historical discrimination results;

[0035] Based on the transformed generated OBD data of each historical OBD data and the corresponding historical discrimination results, use the comprehensive loss function to obtain the real-time comprehensive loss value during the training optimization process;

[0036] If the real-time comprehensive loss value is less than the loss value threshold, output the final generator and the final discriminator, and obtain the final data structure conversion module according to the final generator and the final discriminator;

[0037] Integrate the final sequence feature extraction module and the final data structure conversion module to obtain the final data structure conversion model.

[0038] Furthermore, the vehicle fault diagnosis model is constructed based on the GAT-Droput-MLP algorithm, and the vehicle fault diagnosis model includes a graph structure feature extraction module constructed based on the GAT-Droput algorithm and a vehicle fault diagnosis module constructed based on the MLP algorithm.

[0039] Furthermore, construct a vehicle fault diagnosis model according to several transformed historical OBD data using a deep learning algorithm, and generate several historical vehicle fault diagnosis results, including the following steps:

[0040] Use the GAT-Droput-MLP algorithm to construct an initial vehicle fault diagnosis model; the initial vehicle fault diagnosis model includes an initial graph structure feature extraction module and an initial vehicle fault diagnosis module;

[0041] Optimize the initial graph structure feature extraction module according to several transformed historical OBD data, and during the training optimization process, use the Droput algorithm to prune the initial graph structure feature extraction module to obtain the final graph structure feature extraction module, and generate several historical graph structure features;

[0042] Optimize the initial vehicle fault diagnosis module according to several historical graph structure features to obtain the final vehicle fault diagnosis module, and generate several historical vehicle fault diagnosis results;

[0043] Integrate the final graph structure feature extraction module and the final vehicle fault diagnosis module to obtain the final vehicle fault diagnosis model.

[0044] Furthermore, the fault repair strategy generation model is constructed based on the MOGRPO-FFWA algorithm, and the fault repair strategy generation model includes a fault repair strategy generation module constructed based on the MOGRPO algorithm and a network parameter optimization module constructed based on the FFWA algorithm connected in sequence. The fault repair strategy generation module includes a set of objective functions, an experience replay pool, a policy network, and an agent. The agent is respectively connected to the set of objective functions, the experience replay pool, and the policy network, and the policy network is connected to the network parameter optimization module.

[0045] Furthermore, according to several historical vehicle fault diagnosis results, use the reinforcement learning algorithm to construct a fault repair strategy generation model and generate several historical fault repair strategy generation experiences, including the following steps:

[0046] Use the MOGRPO-FFWA algorithm to construct an initial fault repair strategy generation model, and the initial fault repair strategy generation model includes an initial fault repair strategy generation module and a network parameter optimization module;

[0047] Taking minimizing the prediction error as the optimization goal, use the network parameter optimization module to optimize the initial network parameters of the initial policy network of the initial fault repair strategy generation module to obtain an optimized policy network;

[0048] Set a set of objective functions and an experience replay pool for the initial fault repair strategy generation module with the optimized policy network, set an action space and a state space for the agent, and use the fault repair strategy generation problem as the simulation environment of the initial fault repair strategy generation module;

[0049] Based on any objective function in the set of objective functions, optimize and train the initial fault repair strategy generation module according to the historical vehicle fault diagnosis results to obtain the final fault repair strategy generation module and generate several historical fault repair strategy generation experiences;

[0050] Store several historical fault repair strategy generation experiences in the experience replay pool of the final fault repair strategy generation module, and integrate the final fault repair strategy generation module and the network parameter optimization module to obtain the final fault repair strategy generation model.

[0051] The beneficial effects of the present invention are:

[0052] The present invention discloses a vehicle fault diagnosis method based on OBD data. By constructing a data structure conversion model, it uniformly processes OBD data of different vehicle models and manufacturers, ensuring the consistency of the data structure, thereby improving the universality of the OBD system. It converts the OBD data in sequence structure into graph structure, enhancing the representation ability of OBD data for deep information. It uses a vehicle fault knowledge graph and a named entity and entity relationship recognition model for knowledge mapping, improving the comprehensiveness of OBD data information. And by constructing a vehicle fault diagnosis model, combining multiple algorithms and rules, it improves the generalization ability and robustness of the model, reduces misdiagnosis and missed diagnosis phenomena, and improves the accuracy of fault diagnosis. Through a fault repair strategy generation model, it realizes the automatic generation of repair strategies, improves the efficiency and optimality of strategy generation, and provides high-value references and tips for drivers or maintenance personnel.

[0053] Other beneficial effects of the present invention will be further described in the specific implementation manner. Brief Description of the Drawings

[0054] Figure 1 is a flowchart of the vehicle fault diagnosis method based on OBD data in the present invention. Specific Implementation Manner

[0055] The following further explains the present invention in conjunction with the drawings and specific embodiments.

[0056] Embodiment:

[0057] As Figure 1 shown, this embodiment provides a vehicle fault diagnosis method based on OBD data, including the following steps:

[0058] S1: In the cloud data center, construct a data structure conversion model, a vehicle fault knowledge graph, a named entity and entity relationship recognition model, a vehicle fault diagnosis model, and a fault repair strategy generation model, and deploy them to all OBD systems, including the following steps:

[0059] S1-1: In the cloud data center, collect a number of historical OBD data and a number of historical vehicle fault knowledge, and perform preprocessing to obtain a number of preprocessed historical OBD data and a number of preprocessed vehicle fault knowledge;

[0060] The preprocessing includes formatting processing to ensure that the data meets the input requirements of subsequent model training, and data cleaning to remove redundant and irrelevant information, improving the quality and training efficiency of the data;

[0061] S1-2: Perform data structure conversion on a number of preprocessed historical OBD data to obtain a number of converted historical OBD data;

[0062] Manually convert the historical OBD data in sequence format into the converted historical OBD data in graph structure to provide data support for subsequent model training;

[0063] S1-3: According to a number of preprocessed vehicle fault knowledge, use natural language processing algorithms to construct a named entity and entity relationship recognition model, and obtain a vehicle fault knowledge graph;

[0064] The named entity and entity relationship recognition model is constructed based on the Bidirectional Encoder Representations from Transformers (BERT)-Conditional Random Fields (CRF)-Support Vector Machine (SVM) algorithm from Transformers, and the named entity and entity relationship recognition model includes a semantic feature extraction module constructed based on the BERT algorithm, a named entity recognition module constructed based on the CRF algorithm, and an entity relationship recognition module constructed based on the SVM algorithm connected in sequence;

[0065] The BERT of the semantic feature extraction module can capture the deep semantic information in the knowledge text, which is very useful for identifying different types of knowledge named entities (such as clinical test terms, clinical test component names, etc.). The CRF of the named entity recognition module can consider the dependencies between adjacent knowledge named entity labels, which can help the model learn the sequence dependencies of entity labels, thereby improving the accuracy of named entity annotation and realizing the extraction of knowledge named entities; the SVM of the entity relationship recognition module is mainly used to classify the relationships of the extracted named entity pairs, judge whether there is a specific relationship between them, and the type of the relationship, and convert the named entity and its context information into high-dimensional feature vectors, which can effectively represent the relationships between entities. By combining the deep semantic information extracted by BERT and the sequence dependencies considered by CRF, complex entity relationships can be processed more effectively;

[0066] According to a number of preprocessed vehicle fault knowledge, use natural language processing algorithms to construct a named entity and entity relationship recognition model, and obtain a vehicle fault knowledge graph, including the following steps:

[0067] S1-3-1: Use the BERT-CRF-SVM algorithm to construct an initial named entity and entity relationship recognition model;

[0068] S1-3-2: Optimize and train the initial named entity and entity relationship recognition model based on a number of preprocessed vehicle fault knowledge to obtain the final named entity and entity relationship recognition model, and generate a number of knowledge named entities and a number of knowledge entity relationships, including the following steps:

[0069] S1-3-2-1: Use the initial semantic feature extraction module of the initial named entity and entity relationship recognition model to extract the historical semantic features of the preprocessed vehicle fault knowledge;

[0070] S1-3-2-2: Based on the historical semantic features, use the initial named entity recognition module to perform named entity recognition to obtain a number of knowledge named entities;

[0071] S1-3-2-3: Based on the historical semantic features of a number of knowledge named entities, use the initial entity relationship recognition module to perform entity relationship recognition to obtain the corresponding number of knowledge entity relationships;

[0072] S1-3-2-4: Traverse all the preprocessed vehicle fault knowledge to obtain the final semantic feature extraction module, the final named entity recognition module, and the final entity relationship recognition module;

[0073] S1-3-2-5: Integrate the final semantic feature extraction module, the final named entity recognition module, and the final entity relationship recognition module to obtain the final named entity and entity relationship recognition model, and generate a number of knowledge named entities and a number of knowledge entity relationships;

[0074] S1-3-3: Construct a knowledge graph based on a number of knowledge named entities and a number of knowledge entity relationships to obtain a vehicle fault knowledge graph;

[0075] S1-4: Based on a number of preprocessed historical OBD data and the corresponding converted historical OBD data, use a deep learning algorithm to construct a data structure conversion model;

[0076] The data structure conversion model is constructed based on the Long Short-Term Memory (LSTM)-Conditional Generative Adversarial Network (cGAN) algorithm, and the data structure conversion model includes a sequence feature extraction module constructed based on the LSTM algorithm and a data structure conversion module constructed based on the cGAN algorithm connected in sequence. The data structure conversion module includes a generator and a discriminator connected in sequence, and both the generator and the discriminator are connected to the sequence feature extraction module;

[0077] The LSTM in the sequence feature extraction module is used to process and predict long-term dependencies in time series data. In vehicle fault diagnosis, the original OBD data is often time series data. LSTM can effectively capture the temporal features in this data and extract useful sequence features from the original OBD data. These features may include various time-varying parameters such as engine operating status, vehicle speed, fuel consumption, etc.; the task of the generator in the data structure conversion module is to generate graph-structured data that conforms to the target data structure based on the sequence features extracted from the LSTM. The task of the discriminator is to determine whether the data output by the generator is real, that is, whether it conforms to the real data distribution. Through this adversarial training, the generator continuously optimizes its generation ability until it can generate high-quality data, enabling the converted OBD data to better adapt to subsequent vehicle fault diagnosis;

[0078] According to a number of preprocessed historical OBD data and the corresponding converted historical OBD data, use deep learning algorithms to construct a data structure conversion model, including the following steps:

[0079] S1-4-1: Use the LSTM-cGAN algorithm to construct an initial data structure conversion model; the initial data structure conversion model includes an initial sequence feature extraction module and an initial data structure conversion module;

[0080] S1-4-2: According to the first loss function of the initial generator and the second loss function of the initial discriminator in the initial data structure conversion module, obtain the comprehensive loss function;

[0081] S1-4-3: According to a number of preprocessed historical OBD data, train and optimize the initial sequence feature extraction module to obtain the final sequence feature extraction module, and generate a number of historical OBD data sequence features;

[0082] S1-4-4: According to a number of historical OBD data sequence features, train and optimize the initial generator to obtain an optimized generator, and generate a number of converted generated OBD data;

[0083] S1-4-5: According to a number of converted historical OBD data and the corresponding converted generated OBD data sequences, train and optimize the initial discriminator to obtain an optimized discriminator, and generate a number of historical discrimination results;

[0084] S1-4-6: According to the converted generated OBD data and the corresponding historical discrimination results of each historical OBD data, use the comprehensive loss function to obtain the real-time comprehensive loss value during the training and optimization process;

[0085] S1-4-7: If the real-time comprehensive loss value is less than the loss value threshold, output the final generator and the final discriminator, and based on the final generator and the final discriminator, obtain the final data structure conversion module;

[0086] S1-4-8: Integrate the final sequence feature extraction module and the final data structure conversion module to obtain the final data structure conversion model;

[0087] S1-5: Based on a number of converted historical OBD data, use deep learning algorithms to construct a vehicle fault diagnosis model and generate a number of historical vehicle fault diagnosis results;

[0088] The vehicle fault diagnosis model is constructed based on the Graph Attention Network (GAT)-Droput-Multi-Layer Perceptron (MLP) algorithm, and the vehicle fault diagnosis model includes a graph structure feature extraction module constructed based on the GAT-Droput algorithm and a vehicle fault diagnosis module constructed based on the MLP algorithm;

[0089] The GAT of the graph structure feature extraction module can learn the features of different nodes and the complex interaction features between different nodes in the converted historical OBD data, use the attention mechanism to assign different weights to the node features and edge features of different nodes, so as to highlight the parameter nodes and edge relationships that are more important for fault diagnosis. The Dropout technique is used to reduce the risk of model overfitting and improve the generalization ability of the model; the vehicle fault diagnosis module is used to fuse the features obtained by the graph structure feature extraction module, and perform classification or regression analysis on the fused features to identify whether the vehicle has a fault and may further determine the type and severity of the fault, etc.;

[0090] Based on a number of converted historical OBD data, use deep learning algorithms to construct a vehicle fault diagnosis model and generate a number of historical vehicle fault diagnosis results, including the following steps:

[0091] S1-5-1: Use the GAT-Droput-MLP algorithm to construct an initial vehicle fault diagnosis model; the initial vehicle fault diagnosis model includes an initial graph structure feature extraction module and an initial vehicle fault diagnosis module;

[0092] S1-5-2: According to a number of converted historical OBD data, train and optimize the initial graph structure feature extraction module, and during the training and optimization process, use the Droput algorithm to perform pruning processing on the initial graph structure feature extraction module to obtain the final graph structure feature extraction module and generate a number of historical graph structure features;

[0093] S1-5-3: Optimize the initial vehicle fault diagnosis module according to several historical graph structure features to obtain the final vehicle fault diagnosis module, and generate several historical vehicle fault diagnosis results;

[0094] S1-5-4: Integrate the final graph structure feature extraction module and the final vehicle fault diagnosis module to obtain the final vehicle fault diagnosis model;

[0095] S1-6: According to several historical vehicle fault diagnosis results, use the reinforcement learning algorithm to construct a fault repair strategy generation model, and generate several historical fault repair strategy generation experiences;

[0096] The fault repair strategy generation model is constructed based on the Multi-Objective GroupRelative Policy Optimization (MOGRPO)-Fast Fireworks OptimizationAlgorithm (FFWA). The fault repair strategy generation model includes a fault repair strategy generation module constructed based on the MOGRPO algorithm and a network parameter optimization module constructed based on the FFWA algorithm connected in sequence. The fault repair strategy generation module includes a set of objective functions, an experience replay pool, a policy network, and an agent. The agent is respectively connected to the set of objective functions, the experience replay pool, and the policy network. The policy network is connected to the network parameter optimization module;

[0097] The set of objective functions of the fault repair strategy generation module can handle multiple conflicting objectives, such as repair cost, time, efficiency, etc., and generate repair strategies that balance these objectives. The agent learns historical repair strategies through the experience replay pool, continuously optimizing its own strategy generation ability. The agent updates the policy network according to the learned experience to generate more effective repair strategies. The design of the experience replay pool and the agent enables the model to continuously learn and optimize, improving the quality of strategy generation. Since the fault repair strategy generation module adopts a group exploration method, it can avoid falling into local optimal solutions to a certain extent. The policy network: outputs the distribution probability of actions in a given state. The fault repair strategy generation module directly updates the policy network through gradients, eliminating the Critic model in traditional reinforcement learning, making the algorithm structure more concise; The network parameter optimization module uses the FFWA algorithm to optimize the initial network parameters of the policy network, improving the accuracy and efficiency of fault repair strategy generation;

[0098] According to several historical vehicle fault diagnosis results, use the reinforcement learning algorithm to construct a fault repair strategy generation model, and generate several historical fault repair strategy generation experiences, including the following steps:

[0099] S1-6-1: Use the MOGRPO-FFWA algorithm to construct an initial fault repair strategy generation model, which includes an initial fault repair strategy generation module and a network parameter optimization module;

[0100] S1-6-2: With the goal of minimizing the prediction error, use the network parameter optimization module to optimize the initial network parameters of the initial policy network of the initial fault repair strategy generation module to obtain an optimized policy network, including the following steps:

[0101] S1-6-2-1: Encode the initial network parameters of the initial policy network into the individual vector of the health assessment result optimization model, and use the Circle chaotic mapping sequence for initialization according to the individual vector to obtain an initial FFWA population, which includes several initial FFWA individuals;

[0102] The formula is:

[0103]

[0104] In the formula, q l is the initial FFWA individual of the Circle chaotic mapping; is a randomly generated initial FFWA individual; l is the FFWA individual indicator; mod(*) is the remainder function;

[0105] S1-6-2-2: With the goal of minimizing the prediction error, and according to the optimization goal, set the fitness function;

[0106] The formula is:

[0107] Fit(x) = minMSN

[0108] In the formula, Fit(x) is the fitness function; MSN is the prediction error value;

[0109] S1-6-2-3: According to the fitness function, obtain the fitness values of the initial FFWA individuals in the initial FFWA population, and according to the fitness values, obtain the explosion radius and the number of sparks of the initial FFWA individuals;

[0110] The formula is:

[0111]

[0112] In the formula, S l' is the number of sparks of the initial FFWA individual q l' ; M' is the number constant; f max is the maximum fitness value in the initialized FFWA population; f(q l') is the initial FFWA individual q l' ; τ is an infinitesimal constant; a" is the convergence factor; σ is a non-zero positive real number;

[0113]

[0114] In the formula, R l' is the initial FFWA individual q l' 's explosion radius; is the explosion radius adjustment constant; f min is the minimum fitness value in the initialized FFWA population; L is the total number of FFWA individuals;

[0115]

[0116] In the formula, a" is the convergence factor; tanh(.) is the hyperbolic tangent function; t is the iteration indicator; t max is the maximum number of iterations; a max and a min are the maximum and minimum values of the convergence factor respectively; λ is the decreasing rate parameter, k" is the decreasing period parameter, λ = -2π, k' = π;

[0117] The number of sparks determines the number of sub-fireworks generated after each firework explosion. The explosion radius determines the distribution range of the sparks generated after the firework explosion in the solution space. In the early stage of iteration, the value of a" is large, the number of sparks of the FFWA individual is small, and the explosion radius is large, which helps to reduce the computational burden and is more widely distributed, helping to explore more solution spaces. In the later stage of iteration, a smaller explosion radius helps to perform fine search in the local area, and a larger number of sparks helps to increase the diversity of the search;

[0118] S1-6-2-4: According to the explosion radius, the number of sparks, and the fitness value, perform firework explosion to obtain a number of updated FFWA individuals of the updated FFWA population;

[0119] The formula is:

[0120] q' l' = q l' + S l' × rand(-1, 1)

[0121] In the formula, q' l' is the updated FFWA individual; rand(-1, 1) is a random number from -1 to 1;

[0122] S1-6-2-5: Use the Gaussian mutation algorithm to perform Gaussian mutation on the initialized FFWA population to generate a number of Gaussian mutated FFWA individuals of the Gaussian mutated FFWA population;

[0123] The formula is:

[0124] q" l' = q l' + S l' × G(1, 1)

[0125] In the formula, q" l' is the FFWA individual after Gaussian mutation; G(1, 1) is a random number of a Gaussian distribution with both mean and variance of 1;

[0126] S1-6-2-6: Use the dynamic reverse learning algorithm to perform dynamic reverse learning on the initialized FFWA population to obtain a number of reverse FFWA individuals of the reverse FFWA population;

[0127] The formula is:

[0128] q''' l' = γ × (L max + L min ) - q l'

[0129] In the formula, q''' l' is the reverse FFWA individual; γ is the decreasing inertia coefficient; L max , L min are the maximum and minimum values of the vector space respectively;

[0130] S1-6-2-7: Use the fitness function to obtain the fitness values of each updated FFWA individual, the FFWA individual after Gaussian mutation, and the reverse FFWA individual, and take the FFWA individual with the minimum fitness value as the optimal individual;

[0131] S1-6-2-8: If the number of iterations is greater than or equal to the iteration number threshold or the fitness value of the optimal individual is less than the fitness threshold, decode the individual vector of the optimal individual to obtain the optimal initial network parameters;

[0132] S1-6-2-9: Optimize the initial policy network of the initial fault repair strategy generation module according to the optimal initial network parameters to obtain an optimized policy network;

[0133] S1-6-3: Set a set of objective functions and an experience replay pool for the initial fault repair strategy generation module with the optimized policy network, set an action space and a state space for the intelligent agent, and use the fault repair strategy generation problem as the simulation environment of the initial fault repair strategy generation module;

[0134] S1-6-4: Based on any one of the objective function sets, optimize and train the initial fault repair strategy generation module according to the historical vehicle fault diagnosis results to obtain the final fault repair strategy generation module, and generate several historical fault repair strategy generation experiences;

[0135] S1-6-5: Store several historical fault repair strategy generation experiences in the experience replay pool of the final fault repair strategy generation module, and integrate the final fault repair strategy generation module and the network parameter optimization module to obtain the final fault repair strategy generation model;

[0136] S1-7: Deploy the data structure conversion model, vehicle fault knowledge graph, named entity and entity relationship recognition model, vehicle fault diagnosis model, and fault repair strategy generation model to all OBD systems, including the following steps:

[0137] S1-7-1: Extract the metadata of the data structure conversion model, vehicle fault knowledge graph, named entity and entity relationship recognition model, vehicle fault diagnosis model, and fault repair strategy generation model;

[0138] S1-7-2: Send the metadata to all OBD systems connected to the cloud data center;

[0139] S1-7-3: Based on the OBD system, perform reconstruction according to the metadata to obtain the reconstructed data structure conversion model, reconstructed vehicle fault knowledge graph, reconstructed named entity and entity relationship recognition model, reconstructed vehicle fault diagnosis model, and reconstructed fault repair strategy generation model;

[0140] S2: The OBD system collects the real-time OBD data of the current vehicle and uses the data structure conversion model to perform data structure conversion on the real-time OBD data to obtain the converted real-time OBD data, including the following steps:

[0141] S2-1: The OBD system collects the real-time OBD data of the current vehicle and inputs the real-time OBD data into the data structure conversion model;

[0142] S2-2: Use the sequence feature extraction module of the data structure conversion model to extract the real-time OBD data sequence features of the real-time OBD data;

[0143] S2-3: According to the real-time OBD data sequence features, use the generator of the data structure conversion module of the data structure conversion model to generate the corresponding converted real-time OBD data;

[0144] The converted real-time OBD data is graph-structured data, including several nodes and the edges between the nodes. Among them, the nodes include parameter nodes of engine-related data (engine speed, engine temperature, fuel pressure, air flow, ignition timing, knock sensor data), parameter nodes of emission control system data (oxygen sensor data, nitrogen oxide sensor data, catalytic converter efficiency, particulate trap status, exhaust gas recirculation system status), and parameter nodes of vehicle status data (vehicle speed, wheel speed, vehicle acceleration, braking system status). The attribute of the node is the corresponding parameter value, and the edge is the relationship between each node. For example, the engine speed - engine temperature edge, the engine-related data - fuel pressure edge, etc.;

[0145] S3: The OBD system, according to the vehicle fault knowledge graph, uses the named entity and entity relationship recognition model to perform knowledge mapping on the converted real-time OBD data to obtain the mapped real-time OBD data, including the following steps:

[0146] S3-1: The OBD system uses the semantic feature extraction module of the named entity and entity relationship recognition model to extract the real-time semantic features of the converted real-time OBD data;

[0147] The real-time semantic feature is the semantic feature of the parameter node, generally the semantic feature of the parameter name;

[0148] S3-2: According to the real-time semantic features of the converted real-time OBD data, use the named entity recognition module of the named entity and entity relationship recognition model to perform named entity recognition to obtain several data named entities;

[0149] S3-3: Obtain the similarity between each data named entity and several knowledge named entities in the vehicle fault knowledge graph, and use the knowledge named entity with the highest similarity as the first target knowledge named entity of the data named entity;

[0150] S3-4: Use several knowledge entity relationships of the first target knowledge named entity as several target knowledge entity relationships of the data named entity, and use the knowledge named entity on the other side of the target knowledge entity relationship as the second target knowledge named entity of the data named entity;

[0151] S3-5: According to the first target knowledge named entity, several target knowledge entity relationships, and several second target knowledge named entities of each data named entity, perform knowledge mapping on the real-time OBD data to obtain the mapped real-time OBD data;

[0152] S4: The OBD system, according to the mapped real-time OBD data, uses the vehicle fault diagnosis model to perform vehicle fault diagnosis to obtain the real-time vehicle fault diagnosis result, including the following steps:

[0153] S4-1: The OBD system inputs the real-time OBD data after mapping into the vehicle fault diagnosis model;

[0154] S4-2: Using the graph structure feature extraction module of the vehicle fault diagnosis model, extract the real-time graph structure features of the real-time OBD data after mapping;

[0155] S4-3: According to the real-time graph structure features, use the vehicle fault diagnosis module of the vehicle fault diagnosis model to conduct vehicle fault diagnosis and obtain the real-time vehicle fault diagnosis result;

[0156] The real-time vehicle fault diagnosis result includes real-time fault codes (when there is a problem with the vehicle, the electronic control unit will record corresponding fault codes, which are usually presented in the form of a combination of letters and numbers), real-time fault description information (each fault code has a corresponding detailed description to explain the specific content of the fault. For example, P0123 may indicate an engine coolant temperature sensor circuit range / performance problem), real-time fault severity prediction result (the fault code will indicate its severity. Some minor problems may only cause a warning light to come on, while other critical errors, such as engine overheating, may directly affect driving safety), real-time fault prediction location (the fault diagnosis result will point out the specific location where the fault occurs, such as the engine, transmission, sensor, or actuator, etc.), real-time warning and alarm information (the OBD system can also provide warning and alarm information, such as battery high temperature alarm, single cell high voltage alarm, insulation alarm, etc.);

[0157] S5: The OBD system, according to the real-time vehicle fault diagnosis result, uses the fault repair strategy generation model to generate a fault repair strategy and obtains the real-time fault repair strategy, including the following steps:

[0158] S5-1: Analyze the real-time vehicle fault diagnosis result to obtain several real-time vehicle fault states, and according to the several real-time vehicle fault states, update the state space of the agent in the fault repair strategy generation module of the fault repair strategy generation model to obtain the updated state space;

[0159] S5-2: Randomly extract several historical fault repair strategy generation experiences from the experience replay pool of the fault repair strategy generation module, and generate several possible fault repair actions according to the several historical fault repair strategy generation experiences;

[0160] S5-3: According to the several possible fault repair actions, update the action space of the agent in the fault repair strategy generation module to obtain the updated action space;

[0161] S5-4: Select the real-time objective function from the set of objective functions of the fault repair strategy generation module, and based on the real-time objective function, use an agent to control the policy network to generate the probability distribution of all possible fault repair actions corresponding to each real-time vehicle fault state in the updated state space;

[0162] S5-5: Take the possible fault repair action with the highest probability distribution in the updated action space as the executed fault repair action corresponding to the real-time vehicle fault state;

[0163] S5-6: Integrate the executed fault repair actions of all real-time vehicle fault states in the updated state space to obtain the real-time fault repair strategy;

[0164] The real-time fault repair strategy includes real-time fault isolation decision (when a fault is detected, the OBD system will first try to isolate the fault to prevent it from affecting other systems or components, which may include shutting down certain functions or subsystems to avoid the spread of the fault), real-time fault repair decision (according to the fault diagnosis result, the OBD system will provide specific repair decisions, which may include replacing the faulty component, adjusting system parameters or performing other necessary repair operations), and real-time alarm and notification decision (the OBD system will notify the driver or maintenance personnel of the fault diagnosis and repair suggestions in real time so that they can take actions in a timely manner. For example, display warning information on the vehicle dashboard, send repair suggestions to the driver's mobile phone, etc.).

[0165] The present invention discloses a vehicle fault diagnosis method based on OBD data. By constructing a data structure conversion model, it uniformly processes OBD data of different vehicle models and manufacturers to ensure the consistency of the data structure, thereby improving the universality of the OBD system; converting the OBD data in sequence structure into graph structure enhances the representation ability of OBD data for deep information. Using a vehicle fault knowledge graph and a named entity and entity relationship recognition model for knowledge mapping improves the comprehensiveness of OBD data information. And by constructing a vehicle fault diagnosis model and combining multiple algorithms and rules, it improves the generalization ability and robustness of the model, reduces misdiagnosis and missed diagnosis phenomena, and improves the accuracy of fault diagnosis; through the fault repair strategy generation model, it realizes the automatic generation of repair strategies, improves the efficiency and optimality of strategy generation, and provides high-value references and tips for drivers or maintenance personnel.

[0166] The present invention is not limited to the above optional embodiments, and anyone can obtain other various forms of products under the inspiration of the present invention. The above specific embodiments should not be construed as limiting the protection scope of the present invention. The protection scope of the present invention should be defined by the claims, and the description can be used to interpret the claims.

Claims

1. A vehicle fault diagnosis method based on OBD data, characterized in that: The steps include: Cloud data center: build data structure conversion model, vehicle fault knowledge graph, named entity and entity relationship recognition model, vehicle fault diagnosis model and fault repair strategy generation model, and deploy them to all OBD systems; The OBD system collects the real-time OBD data of the current vehicle, and uses the data structure conversion model to convert the real-time OBD data into a data structure, thereby obtaining the converted real-time OBD data; The OBD system, based on the vehicle fault knowledge graph, uses the named entity and entity relationship recognition model to perform knowledge mapping on the converted real-time OBD data to obtain the mapped real-time OBD data; The OBD system uses the vehicle fault diagnosis model to perform vehicle fault diagnosis based on the mapped real-time OBD data and obtains real-time vehicle fault diagnosis results; The OBD system generates a fault repair strategy based on the real-time vehicle fault diagnosis result using a fault repair strategy generation model to obtain a real-time fault repair strategy.

2. A vehicle fault diagnosis method based on OBD data according to claim 1, characterized in that: The cloud data center builds a data structure conversion model, a vehicle fault knowledge graph, a named entity and entity relationship recognition model, a vehicle fault diagnosis model, and a fault repair strategy generation model, and deploys them to all OBD systems, including the following steps: The cloud data center collects some historical OBD data and some historical vehicle fault knowledge, and performs preprocessing to obtain some preprocessed historical OBD data and some preprocessed vehicle fault knowledge; Performing data structure conversion on a number of pre-processed historical OBD data to obtain a number of converted historical OBD data; Based on some pre-processed vehicle fault knowledge, a named entity and entity relationship recognition model is constructed using a natural language processing algorithm, and a vehicle fault knowledge graph is obtained; Based on a number of pre-processed historical OBD data and the corresponding converted historical OBD data, a data structure conversion model is constructed using a deep learning algorithm; Based on some converted historical OBD data, a vehicle fault diagnosis model is constructed using a deep learning algorithm, and some historical vehicle fault diagnosis results are generated; Based on several historical vehicle fault diagnosis results, a reinforcement learning algorithm is used to build a fault repair strategy generation model and generate several historical fault repair strategy generation experiences; The data structure conversion model, vehicle fault knowledge graph, named entity and entity relationship recognition model, vehicle fault diagnosis model and fault repair strategy generation model are deployed to all OBD systems.

3. A vehicle fault diagnosis method based on OBD data according to claim 2, characterized in that: The named entity and entity relationship recognition model is constructed based on the BERT-CRF-SVM algorithm, and the named entity and entity relationship recognition model includes a semantic feature extraction module constructed based on the BERT algorithm, a named entity recognition module constructed based on the CRF algorithm, and an entity relationship recognition module constructed based on the SVM algorithm, which are connected in sequence.

4. A vehicle fault diagnosis method based on OBD data according to claim 3, characterized in that: Based on some pre-processed vehicle fault knowledge, a named entity and entity relationship recognition model is constructed using a natural language processing algorithm, and a vehicle fault knowledge graph is obtained, including the following steps: Use the BERT-CRF-SVM algorithm to build an initial named entity and entity relationship recognition model; According to some pre-processed vehicle fault knowledge, the initial named entity and entity relationship recognition model is optimized and trained to obtain the final named entity and entity relationship recognition model, and generate some knowledge named entities and some knowledge entity relationships; Based on several knowledge named entities and several knowledge entity relationships, a knowledge graph is constructed to obtain a vehicle fault knowledge graph.

5. A vehicle fault diagnosis method based on OBD data according to claim 2, characterized in that: The data structure conversion model is constructed based on the LSTM-cGAN algorithm, and the data structure conversion model includes a sequence feature extraction module constructed based on the LSTM algorithm and a data structure conversion module constructed based on the cGAN algorithm, which are connected in sequence. The data structure conversion module includes a generator and a discriminator connected in sequence, and the generator and the discriminator are both connected to the sequence feature extraction module.

6. A vehicle fault diagnosis method based on OBD data according to claim 5, characterized in that: Based on a number of pre-processed historical OBD data and the corresponding converted historical OBD data, a data structure conversion model is constructed using a deep learning algorithm, including the following steps: Using the LSTM-cGAN algorithm, an initial data structure conversion model is constructed; the initial data structure conversion model includes an initial sequence feature extraction module and an initial data structure conversion module; According to the first loss function of the initial generator and the second loss function of the initial discriminator in the initial data structure conversion module, a comprehensive loss function is obtained; According to a number of pre-processed historical OBD data, the initial sequence feature extraction module is trained and optimized to obtain a final sequence feature extraction module, and a number of historical OBD data sequence features are generated; According to several historical OBD data sequence characteristics, the initial generator is trained and optimized to obtain an optimized generator, and several converted OBD data are generated; According to a number of converted historical OBD data and corresponding converted generated OBD data sequences, an initial discriminator is trained and optimized to obtain an optimized discriminator, and a number of historical discrimination results are generated; According to the conversion of each historical OBD data, OBD data and corresponding historical discrimination results are generated, and the comprehensive loss function is used to obtain the real-time comprehensive loss value in the training optimization process; If the real-time comprehensive loss value is less than the loss value threshold, the final generator and the final discriminator are output, and the final data structure conversion module is obtained according to the final generator and the final discriminator; The final sequence feature extraction module and the final data structure conversion module are integrated to obtain the final data structure conversion model.

7. A vehicle fault diagnosis method based on OBD data according to claim 2, characterized in that: The vehicle fault diagnosis model is constructed based on the GAT-Droput-MLP algorithm, and the vehicle fault diagnosis model includes a graph structure feature extraction module constructed based on the GAT-Droput algorithm and a vehicle fault diagnosis module constructed based on the MLP algorithm.

8. A vehicle fault diagnosis method based on OBD data according to claim 7, characterized in that: Based on several converted historical OBD data, a vehicle fault diagnosis model is constructed using a deep learning algorithm, and several historical vehicle fault diagnosis results are generated, including the following steps: Using the GAT-Droput-MLP algorithm, an initial vehicle fault diagnosis model is constructed; the initial vehicle fault diagnosis model includes an initial graph structure feature extraction module and an initial vehicle fault diagnosis module; According to several converted historical OBD data, the initial graph structure feature extraction module is trained and optimized. During the training and optimization process, the Droput algorithm is used to prune the initial graph structure feature extraction module to obtain the final graph structure feature extraction module and generate several historical graph structure features. According to several historical graph structural features, the initial vehicle fault diagnosis module is trained and optimized to obtain the final vehicle fault diagnosis module, and several historical vehicle fault diagnosis results are generated; The final graph structure feature extraction module and the final vehicle fault diagnosis module are integrated to obtain the final vehicle fault diagnosis model.

9. A vehicle fault diagnosis method based on OBD data according to claim 2, characterized in that: The fault repair strategy generation model is constructed based on the MOGRPO-FFWA algorithm, and the fault repair strategy generation model includes a fault repair strategy generation module constructed based on the MOGRPO algorithm and a network parameter optimization module constructed based on the FFWA algorithm, which are connected in sequence. The fault repair strategy generation module includes an objective function set, an experience replay pool, a strategy network and an intelligent agent. The intelligent agent is respectively connected to the objective function set, the experience replay pool and the strategy network, and the strategy network is connected to the network parameter optimization module.

10. The vehicle fault diagnosis method based on OBD data according to claim 9, characterized in that: Based on several historical vehicle fault diagnosis results, a reinforcement learning algorithm is used to build a fault repair strategy generation model and generate several historical fault repair strategy generation experiences, including the following steps: Using the MOGRPO-FFWA algorithm, an initial fault troubleshooting strategy generation model is constructed, wherein the initial fault troubleshooting strategy generation model includes an initial fault troubleshooting strategy generation module and a network parameter optimization module; Taking minimizing the prediction error as the optimization goal, the network parameter optimization module is used to optimize the initial network parameters of the initial strategy network of the initial fault repair strategy generation module to obtain an optimized strategy network; Setting an objective function set and an experience replay pool for an initial fault repair strategy generation module with an optimized strategy network, setting an action space and a state space for an intelligent agent, and using a fault repair strategy generation problem as a simulation environment for the initial fault repair strategy generation module; Based on any objective function in the objective function set and according to historical vehicle fault diagnosis results, the initial fault troubleshooting strategy generation module is optimized and trained to obtain the final fault troubleshooting strategy generation module, and a number of historical fault troubleshooting strategy generation experiences are generated; Several historical fault troubleshooting strategy generation experiences are stored in the experience replay pool of the final fault troubleshooting strategy generation module, and the final fault troubleshooting strategy generation module and the network parameter optimization module are integrated to obtain the final fault troubleshooting strategy generation model.

Images