Fault early warning method, training method, electronic equipment and storage medium
By using LSTM network model and knowledge graph, fault prediction model is trained based on vehicle data and fault data, the problem of inadequate warning in the existing technology is solved, accurate prediction and timely reminder of vehicle failures is achieved, and vehicle safety is improved .
Patent Information
- Application Number
- CN202310639677.9
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2023-05-29
- Publication Date
- 2024-11-29
AI Technical Summary
The prior art is difficult to provide early warning before a vehicle failure occurs. It can only infer the type of fault when a fault occurs, and cannot achieve early warning, which may lead to more serious consequences.
Long-term short-term memory (LSTM) network model and knowledge graph are used to acquire vehicle data and pre-trained fault data, train fault prediction models, and make fault prediction based on entity information and entity relationships, and early warning.
Timely reminders of potential vehicle failures are achieved, losses caused by failures are reduced, and the accuracy of fault prediction is improved by considering the mutual influence between vehicle components.
Smart Images

Figure CN119049148A_ABST
Abstract
Description
Technical Field
[0001] Embodiments of the present application relate to the field of vehicle technology, and more specifically, to a fault warning method, a training method, an electronic device, and a storage medium. Background Art
[0002] Most car owners do not have a deep understanding of the vehicle's structure, and they only realize that a fault has occurred when a fault code causes an alarm or the vehicle has obvious abnormalities. If the user does not pay attention to it or it is inconvenient to repair (such as on the highway), it may bring more serious consequences. In order to improve the safety of vehicle use, some technologies use some vehicle fault inference methods based on simple logical relationships to infer the type of vehicle fault.
[0003] However, the above inference method can often only infer the fault type when the fault occurs to assist in fault repair, but cannot achieve an early warning effect. Summary of the invention
[0004] The embodiments of the present application provide a fault warning method, training method, electronic device and storage medium that can at least partially solve the above-mentioned problems or other problems existing in the prior art.
[0005] An embodiment of the present application provides a fault warning method, including: acquiring vehicle data within a specified time step; determining a vehicle fault prediction result based on a pre-trained fault prediction model according to the vehicle data and a pre-trained knowledge graph of vehicle faults; and performing a warning operation in response to the fault prediction result indicating that the vehicle has a fault risk, wherein the network model of the fault prediction model is a long short-term memory (LSTM) network model.
[0006] An embodiment of the present application also provides a method for training a fault prediction model, including: obtaining a pre-trained knowledge graph of vehicle faults and vehicle fault data, wherein the vehicle fault data is at least distinguished and labeled as fault-free or faulty; and using the knowledge graph and vehicle fault data to train a long short-term memory (LSTM) network model to obtain a fault prediction model.
[0007] An embodiment of the present application also provides an electronic device, comprising: at least one processor and a memory, wherein the memory is communicatively connected to the at least one processor and stores instructions that can be executed by the at least one processor, and the instructions are executed by the at least one processor so that the at least one processor can execute the fault warning method mentioned in the above embodiment or the training method of the fault prediction model mentioned in the above embodiment.
[0008] The embodiments of the present application also provide a computer-readable storage medium storing a computer program. When the computer program is executed by a processor, the fault warning method mentioned in the above embodiments or the training method of the fault prediction model mentioned in the above embodiments is implemented.
[0009] According to some embodiments of the present application, the network model of the fault prediction model uses a long short-term memory network model, which can help the owner analyze the possibility of potential faults based on the vehicle data within a specified time step, and promptly remind the owner, thereby reducing the loss of the owner caused by the fault. In addition, the fault prediction model performs fault prediction based on the entity information indicated by the knowledge graph and the information on the relationship between the entities, fully considering the mutual influence between vehicle components, and can better predict vehicle faults compared to fault prediction based on separate signals. BRIEF DESCRIPTION OF THE DRAWINGS
[0010] Other features, objects and advantages of the present application will become more apparent by reading the detailed description of non-limiting embodiments made with reference to the following drawings, in which:
[0011] Figure 1 is a schematic block diagram of a vehicle fault warning system according to some embodiments of the present application;
[0012] Figure 2 is a flowchart of a method for training a fault prediction model according to some embodiments of the present application;
[0013] Figure 3 is a flowchart of a fault warning method according to some embodiments of the present application; and
[0014] Figure 4 is a schematic block diagram of an electronic device according to some embodiments of the present application. DETAILED DESCRIPTION
[0015] In order to better understand the present application, a more detailed description will be made of various aspects of the present application with reference to the accompanying drawings. It should be understood that these detailed descriptions are only descriptions of exemplary embodiments of the present application, and are not intended to limit the scope of the present application in any way. Throughout the specification, the same reference numerals refer to the same elements. The expression "and / or" includes any and all combinations of one or more of the associated listed items.
[0016] It should also be understood that expressions such as "include", "including", "have", "contain" and / or "comprising" are open rather than closed expressions in this specification, which indicate the presence of the stated features, elements and / or components, but do not exclude the presence of one or more other features, elements, components and / or combinations thereof. In addition, when expressions such as "at least one of..." appear after a list of listed features, they modify the entire list of features rather than just the individual elements in the list. In addition, when describing embodiments of the present application, "may" is used to mean "one or more embodiments of the present application". And, the term "exemplary" is intended to refer to an example or illustration.
[0017] Unless otherwise specified, all words (including engineering terms and scientific and technological terms) used in this article have the same meaning as those commonly understood by ordinary technicians in the field to which this application belongs. It should also be understood that, unless clearly stated in this application, words defined in common dictionaries should be interpreted as having the same meaning as their meaning in the context of the relevant technology, and should not be interpreted in an idealized or overly formal sense.
[0018] It should be noted that, in the absence of conflict, the embodiments and features in the embodiments of the present application can be combined with each other. In addition, unless explicitly limited or contradictory to the context, the specific steps included in the method described in the present application are not necessarily limited to the order described, but can be performed in any order or in parallel. The present application will be described in detail below with reference to the accompanying drawings and in conjunction with the embodiments.
[0019] Figure 1 is a schematic block diagram of a vehicle fault warning system according to some embodiments of the present application. Figure 1 As shown, the vehicle fault warning system 100 may include a plurality of vehicle components (eg, 110) and a warning device 120. Optionally, the vehicle component 110 may be communicatively connected with the warning device 120 to transmit signal data to the warning device 120.
[0020] It should be understood that, without departing from the teachings of this application, the early warning device 120 may obtain signal data associated with the vehicle component 110 from the vehicle component 110 , or may obtain signal data associated with the vehicle component 110 from detection sensors and other devices of the vehicle component 110 , and this application does not impose any restrictions on this.
[0021] In some embodiments of the present application, the early warning device 120 may be a vehicle host, or a server that is communicatively connected to the vehicle host, etc., and the present application does not impose any limitation on this.
[0022] In some embodiments of the present application, a fault prediction model is provided in the warning device 120, and a fault warning is performed through the fault prediction model. For ease of understanding, the process of training the fault prediction model and the process of performing a fault warning based on the fault prediction model are described below.
[0023] Figure 2 FIG. 1 is a flow chart of a method for training a fault prediction model according to some embodiments of the present application. Figure 2 As shown, the training method 200 of the fault prediction model may include the following steps:
[0024] S21, obtaining a pre-trained knowledge graph of vehicle faults and vehicle fault data, wherein the vehicle fault data is at least distinguished and marked as no fault or fault; and
[0025] S22, use the knowledge graph and vehicle fault data to train the long short-term memory network model to obtain a fault prediction model.
[0026] According to some embodiments of the present application, using the knowledge graph of vehicle faults to train the fault prediction model is conducive to the fault prediction model to perform fault prediction based on the entity information indicated by the knowledge graph and the information on the relationship between the entities, fully considering the mutual influence between vehicle components, and better predicting vehicle faults compared to fault prediction based on separate signals. In addition, training the fault prediction model based on the long short-term memory network model allows the fault prediction model to fully consider the cumulative impact of abnormal conditions of vehicle components and more accurately predict vehicle faults.
[0027] For ease of understanding, the following Figure 2 The training method 200 of the fault prediction model is illustrated as an example.
[0028] In some embodiments of the present application, the knowledge graph can be obtained by training vehicle fault data. For example, the vehicle fault data can be fault data uploaded by multiple vehicles stored in the cloud, or it can be fault data stored locally in a certain vehicle, and the present application does not limit this. Most fault phenomena are caused by the influence of multiple abnormal vehicle components, and the mutual influence between vehicle components can be regarded as prior knowledge. In order to learn prior knowledge, other network models such as Bayesian networks can be used in some technologies. However, Bayesian network models are relatively weak in semantic understanding, and there are many names of vehicle components. After learning prior knowledge using models such as Bayesian networks, each vehicle component will be regarded as an independent feature, which will increase the complexity of the model. In some embodiments of the present application, prior knowledge can be better learned and applied through knowledge graphs, semantic understanding can be better performed, and prior knowledge can be learned from historical vehicle fault data.
[0029] Optionally, the data type of the vehicle fault data may be time series data, i.e., a series of continuous fault data. In the above example, using time series data as training data to train the fault prediction model is helpful to discover the timing problem of vehicle faults, so as to try to provide fault warning earlier before the fault occurs.
[0030] In some embodiments of the present application, the model of the knowledge graph may be a TransGCN model. The TransGCN model is trained to obtain information about each entity in the triple of the knowledge graph of vehicle faults and information representing the relationship between entities. For example, the TransGCN model includes a TransR model and a graph convolutional network. For example, the TransR model includes entity information and information representing the relationship between entities. In the above example, the Trans model in the TransGCN model uses the TransR model, so that different relationships have different semantic spaces, and in their respective semantic spaces, the entity B vector can be obtained based on the mapping of entity A vector and relationship vector. Compared with models such as Bayesian network, the TransR model can better handle text semantic relationships. For example, given the triple (signal A, high, fault C), signal B is similar to signal A, and the TransR model can infer (signal B, high, fault C). Based on this, the TransR model is selected to train the knowledge graph, which can better understand the semantics of the training data and help the fault prediction model better understand the entity relationship.
[0031] It is understandable that, without departing from the teachings of the present application, the relationship between entities may also include other relationships, such as the relationship between vehicle signals and fault types, the relationship between fault phenomena and fault types, etc., and the present application does not impose any limitation on this.
[0032] It should be understood that, without departing from the teachings of this application, entity information may include information characterizing vehicle components, vehicle signals, fault phenomena, and fault types. There may be many one-to-many and many-to-one relationships between triples, and their specific relationships are obtained through training data. Information characterizing the relationship between entities may at least include: first relationship information indicating the relationship between vehicle signals, vehicle components, and fault phenomena, second relationship information between vehicle components and fault types, third relationship information between vehicle components, and fourth relationship information between vehicle signals. This application does not limit entity information and information characterizing the relationship between entities.
[0033] For example, vehicle components may refer to specific components of a vehicle, such as doors, windows, steering wheels, etc., fault phenomena may refer to "low water temperature, low oil volume", etc., vehicle signals may refer to sensor signals related to fault phenomena, and fault types may be predefined according to the type of vehicle fault currently found. In some examples, the relationship text between vehicle signals and fault types may be the value of the signal, such as signal A high leads to fault B, the triple is (signal A, high, fault B), the relationship text between fault phenomena and vehicle signals may also be the value, such as low water temperature leads to signal C, the triple is (water temperature, low, signal C), and the relationship text leading to the fault may use relationship words such as "lead to", for example, (signal A, lead to, fault A), which is not limited here. These triples may also have hidden relationships, for example, signal A-influence-signal B-lead-component C-occurrence-fault D, signal A-produce-phenomenon B-produce-signal C-lead-component D-occurrence-fault E, which is not limited here.
[0034] In some embodiments of the present application, after the knowledge graph is obtained through training, a long short-term memory network model can be trained based on the knowledge graph and vehicle fault data to obtain a vehicle fault prediction model.
[0035] As an option, vehicle fault data can be labeled as non-faulty or faulty as training data, so that the fault prediction model can subsequently directly predict whether the vehicle has a fault or not.
[0036] Optionally, the fault data of the vehicle with faults can be distinguished and labeled as at least one fault type. In other words, the vehicle fault data, as training data, can be labeled as no fault and a specific fault type, so that the fault prediction model can subsequently predict the possible fault type of the vehicle.
[0037] It should be understood that the labeling method of vehicle fault data can be adjusted as needed without departing from the teachings of this application, and this application does not impose any restrictions on this.
[0038] In some embodiments of the present application, the fault prediction model may perform fault prediction for the entire vehicle, or may perform fault prediction for each vehicle component.
[0039] For example, the fault prediction model may be a fault prediction model for the entire vehicle. For example, the fault data of all vehicle components and the knowledge graph may be used to train an initial long short-term memory network model to obtain a fault prediction model.
[0040] As another example, the fault prediction model may be a fault prediction model for vehicle components. For example, a long short-term memory network model may be established for each vehicle component. The process of training the long short-term memory network model using the knowledge graph and vehicle fault data to obtain a fault prediction model may include: using the knowledge graph and the associated data related to each vehicle component in the vehicle fault data to train different long short-term memory network models respectively to obtain a fault prediction model for each vehicle component. For example, the input data of the long short-term memory network model may include all signal data for each time step, the corresponding entity vector, and the vector related to the vehicle component fault in the knowledge graph. The output of the long short-term memory network model may be multi-classified, and the predicted values are no fault and multiple specific fault types.
[0041] It should be understood that, without departing from the teachings of this application, the model architecture of the long short-term memory network model (such as the number of convolutional layers, etc.) can be determined based on one or more factors such as the volume of vehicle data, the project's requirements for vehicle fault prediction accuracy, and the computing power of the early warning device 120 using the fault prediction model, and this application does not impose any restrictions on this.
[0042] In the above example, the knowledge graph is used to train the fault prediction model, and the correspondence between vehicle faults and related signals, the whole vehicle system and each subsystem, and each subsystem and various vehicle faults is applied, so that the fault prediction model can better predict the failure of vehicle components and concurrent failures of the vehicle components.
[0043] After completing the exemplary description of the training process of the fault prediction model, the following exemplary description is given of a method for performing fault warning based on the fault prediction model.
[0044] Figure 3 FIG. 1 is a flow chart of a fault warning method according to some embodiments of the present application. Figure 3 As shown, the fault warning method 300 may include the following steps:
[0045] S31, obtaining vehicle data within a specified time step;
[0046] S32, based on the pre-trained fault prediction model, determining a vehicle fault prediction result according to the vehicle data and the pre-trained vehicle fault knowledge graph; and
[0047] S33, in response to the fault prediction result indicating that the vehicle has a fault risk, performing a warning operation, wherein the network model of the fault prediction model is a long short-term memory (LSTM) network model.
[0048] According to some embodiments of the present application, the network model of the fault prediction model uses a long short-term memory network model, which can help the owner analyze the possibility of potential faults based on the vehicle data within a specified time step, and promptly remind the owner, thereby reducing the loss of the owner caused by the fault. In addition, the fault prediction model performs fault prediction based on the entity information indicated by the knowledge graph and the information on the relationship between the entities, fully considering the mutual influence between vehicle components, and can better predict vehicle faults compared to fault prediction based on separate signals.
[0049] For ease of understanding, the fault warning method 300 is exemplarily described below.
[0050] In some embodiments of the present application, the above-mentioned fault warning method 300 may be executed by the warning device 120 of the vehicle fault warning system 100. The knowledge graph and fault prediction model in the warning device 120 may be obtained by training the warning device 120 itself, or may be obtained by training other devices and then imported into the warning device 120, and the present application does not limit this.
[0051] In some embodiments of the present application, the early warning device 120 may periodically or continuously collect vehicle data. The vehicle data may include vehicle signals. The vehicle signals may include sensor signals related to vehicle components. For example, signals of vehicle components themselves or signals for detecting some performance parameters of vehicle components, etc., which are not limited in the present application.
[0052] In some embodiments of the present application, the early warning device 120 may perform fault prediction based on the vehicle data within the time step after collecting the vehicle data of the specified time step. For example, a timer may be configured in the early warning device 120. The timing duration of the timer may be a specified time step. The early warning device 120 may perform fault prediction based on the vehicle data collected within the time period after the timer reaches its timing duration.
[0053] It should be understood that, without departing from the teachings of this application, the early warning device 120 may start data collection and fault warning and other related operations after determining that the vehicle is started, or may start data collection and fault warning and other related operations after determining that the vehicle starts to travel, so as to detect vehicle faults early during vehicle travel and reduce the situation where faults during vehicle travel cause harm to its driving safety. This application does not limit the timing of triggering data collection and fault warning and other related operations by the early warning device 120.
[0054] In some embodiments of the present application, the model of the knowledge graph is a TransGCN model, and the information of the entities in the triples of the knowledge graph of vehicle faults and the information representing the relationship between the entities are obtained through the TransGCN model. Among them, the entities may include vehicle parts, vehicle signals, fault phenomena and fault types, etc., which are not limited here, and the relationships can be obtained based on vehicle fault data training, which are not listed here one by one.
[0055] Optionally, the entity information and the information representing the relationship between the entities can be represented in the form of vectors. Representing the entity information and the information representing the relationship between the entities in the form of vectors can better reflect the relationship between the entities and is more conducive to discovering hidden relationships.
[0056] It should be understood that, without departing from the teachings of the present application, entity information and information characterizing the relationship between entities may also be represented in other forms, and the present application does not impose any limitation on this.
[0057] Optionally, the TransGCN model includes the TransR model and the graph convolutional network. The Trans model in the TransGCN model uses the TransR model, so that different relationships have different semantic spaces, and in their respective semantic spaces, the entity B vector can be obtained based on the mapping of entity A vector and relationship vector. Compared with models such as the Bayesian network, the TransR model can better handle text semantic relationships. In the above example, the TransR model is selected to train the knowledge graph, which can better understand the semantics of the training data and help the fault prediction model better understand the entity relationship.
[0058] Optionally, the TransR model includes entity information and information characterizing the relationship between entities, wherein the entity information includes information characterizing vehicle components, vehicle signals, fault phenomena and fault types, and the relationship information includes at least: first relationship information indicating the relationship between vehicle signals, vehicle components and fault phenomena, second relationship information between vehicle components and fault types, third relationship information between vehicle components, and fourth relationship information between vehicle signals.
[0059] For example, vehicle components may refer to specific components of a vehicle, such as doors, windows, steering wheels, etc., fault phenomena may refer to "low water temperature, low oil volume", etc., vehicle signals may refer to sensor signals related to fault phenomena, and fault types may be predefined according to the type of vehicle fault currently found. In some examples, the relationship text between vehicle signals and fault types may be the value of the signal, such as signal A high leads to fault B, the triple is (signal A, high, fault B), the relationship text between fault phenomena and vehicle signals may also be the value, such as low water temperature leads to signal C, the triple is (water temperature, low, signal C), and the relationship text leading to the fault may use relationship words such as "lead to", for example, (signal A, lead to, fault A), which is not limited here. These triples may also have hidden relationships, for example, signal A-influence-signal B-lead-component C-occurrence-fault D, signal A-produce-phenomenon B-produce-signal C-lead-component D-occurrence-fault E, which is not limited here.
[0060] In some embodiments of the present application, the process of determining the vehicle fault prediction result may be determined according to the type of the fault prediction model.
[0061] For example, the fault prediction model is a fault prediction model for the entire vehicle. Based on the pre-trained fault prediction model, the process of determining the vehicle fault prediction result according to the vehicle data and the pre-trained vehicle fault knowledge graph may include: inputting the vehicle data within a specified time step and the relationship information and entity information indicated by the knowledge graph into the fault prediction model to predict whether the vehicle has a fault. For example, during the use of the vehicle, the early warning device 120 can collect vehicle data of the vehicle in real time. When the collected data reaches the specified time step, the data of the previous time step can be combined with the TransGCN model and the fault prediction model to predict whether the vehicle is about to fail.
[0062] For another example, the fault prediction model is a fault prediction model for vehicle parts. Based on the pre-trained fault prediction model, the process of determining the vehicle fault prediction result according to the vehicle data and the pre-trained vehicle fault knowledge graph may include: inputting the associated data related to the vehicle parts in the vehicle data, the entities and relationships related to the vehicle parts in the knowledge graph into the fault prediction model to obtain the fault prediction result for the vehicle parts. Among them, the vehicle data may include vehicle signals. For example, the entity information and the information representing the relationship between the entities in the knowledge graph exist in the form of vectors. The vehicle can collect vehicle data in real time during use. When the collected vehicle data reaches a specified time step, the associated data related to the vehicle part in the vehicle data of the previous time step, the entity vector corresponding to the vehicle part in the knowledge graph, and the relationship vector related to the vehicle part in the knowledge graph are input into the fault prediction model to predict in real time whether the vehicle part is faulty.
[0063] It should be understood that, without departing from the teachings of the present application, the fault prediction model may also be a fault prediction model for each system of the vehicle, for example, a braking system, etc., and the present application does not impose any limitation on this.
[0064] In some embodiments of the present application, the fault prediction model is at least used to determine the probability of a fault in a vehicle component represented by the associated data. The early warning device 120 may determine that the fault prediction result indicates that the vehicle has a fault risk in response to the fault prediction result indicating that the probability of a fault in the vehicle component is greater than the early warning threshold. For example, if the vehicle fault data used to train the fault prediction model is labeled as faulty and non-faulty, the fault prediction model may determine the probability of a fault in the vehicle component based on the input data. If the fault prediction model determines that the probability of a fault in the vehicle component is greater than the early warning threshold based on the associated data of the vehicle component within the input time step, and the entities and relationships related to the vehicle component, it may be determined that the fault prediction result indicates that the vehicle has a fault risk, triggering an early warning operation.
[0065] Optionally, the fault prediction model is at least used to determine the probability of occurrence of each fault type of the vehicle component represented by the associated data. In response to the fault prediction result indicating that the probability of a vehicle component having a fault is greater than the warning threshold, the early warning device 120 determines that the fault prediction result indicates that the vehicle has a fault risk, and the process may include: in response to the fault prediction result indicating that the probability of occurrence of any fault type of the vehicle component is greater than the warning threshold of the fault type, determining that the fault prediction result indicates that the vehicle has a fault risk. For example, if the vehicle fault data used to train the fault prediction model is labeled as a specific fault type and no fault, the fault prediction model can determine the probability of the vehicle component having faults corresponding to each fault type based on the input data. If the fault prediction model determines that the probability of occurrence of any fault type of the vehicle component is greater than the warning threshold of the fault type based on the associated data of the vehicle component within the input time step, and the entities and relationships related to the vehicle component, it can be determined that the fault prediction result indicates that the vehicle has a fault risk, and the early warning operation is triggered.
[0066] In some embodiments of the present application, the warning operation performed by the warning device 120 may include but is not limited to at least one of the following: playing warning voice through the vehicle speaker, displaying warning information through the vehicle's central control display screen, projecting and displaying warning information through the vehicle's head-up display device (Head-Up Display, HUD) and sending warning information to the owner's mobile terminal.
[0067] Optionally, the early warning operation may include: feedback of the fault type of the vehicle component whose probability of occurrence is greater than the early warning threshold. In the above example, the early warning device 120 can timely feedback the possible fault type of the vehicle, which is convenient for the owner and maintenance personnel to understand the vehicle condition so that they can perform vehicle maintenance based on the vehicle fault type.
[0068] It should be understood that, without departing from the teachings of this application, the warning operations executable by the warning device 120 can be set as needed, and this application does not impose any restrictions on this.
[0069] According to some embodiments of the present application, the network model of the fault prediction model uses a long short-term memory network model, which can help the owner analyze the possibility of potential faults based on the vehicle data within a specified time step, and promptly remind the owner, thereby reducing the loss of the owner caused by the fault. In addition, the fault prediction model performs fault prediction based on the entity information indicated by the knowledge graph and the information on the relationship between the entities, fully considering the mutual influence between vehicle components, and can better predict vehicle faults compared to fault prediction based on separate signals.
[0070] In addition, in some implementations, the TransR model is used in the knowledge graph to apply text semantics to the knowledge graph, helping the model to better understand entity relationships.
[0071] It should be noted that in the technical solution of the present disclosure, the acquisition, storage and application of the user's personal information involved are in compliance with the provisions of relevant laws and regulations and do not violate public order and good morals. It should also be noted that the vehicle data in this embodiment is obtained after the user's authorization (i.e., with the user's consent). In addition, the information / data obtained is not intended to characterize a specific type of user, and therefore cannot reflect the personal information of a specific type of user.
[0072] The step division of the above methods is only for the purpose of clear description. When implemented, they can be combined into one step or some steps can be split and decomposed into multiple steps. As long as they include the same logical relationship, they are all within the scope of protection of this patent; adding insignificant modifications to the algorithm or process or introducing insignificant designs without changing the core design of the algorithm and process are all within the scope of protection of this patent.
[0073] The embodiment of the present application also provides an electronic device, such as Figure 4 As shown, the electronic device 400 may include: at least one processor and a memory, the memory is communicatively connected to the at least one processor and stores instructions that can be executed by the at least one processor, and the instructions are executed by the at least one processor so that the at least one processor can execute the above-mentioned fault prediction model training method 200 or fault warning method 300.
[0074] For example, the electronic device 400 may be the warning device 120 in the vehicle fault warning system 100 . The warning device 120 may be implemented as a vehicle host or a server in communication with the vehicle host, and this application does not impose any limitation on this.
[0075] An embodiment of the present application further provides a computer-readable storage medium, which stores a computer program. When the computer program is executed by a processor, the fault prediction model training method 200 or the fault warning method 300 is implemented.
[0076] Figure 4 4 is a schematic block diagram of an electronic device 400 according to some embodiments of the present application. Figure 4 As shown, the electronic device 400 includes a processor 401, which can perform various appropriate actions and processes according to a computer program stored in a read-only memory (ROM) 402 or a computer program loaded from a memory 408 into a random access memory (RAM) 403. In the RAM 403, various programs and data required for the operation of the electronic device 400 can also be stored. The processor 401, the ROM 402, and the RAM 403 are connected to each other via a bus 404. An input / output (I / O) interface 405 is also connected to the bus 404.
[0077] Multiple components in the electronic device 400 are connected to the I / O interface 405, including: an input unit 406, such as buttons and touch screens of the vehicle computer; an output unit 407, connected to, for example, various types of displays, speakers, etc., to output various forms of signals; a memory 408, including any medium for storing computer executable programs; and a communication unit 409, such as a network card, a modem, a wireless communication transceiver, etc. The communication unit 409 allows the electronic device 400 to exchange information / data with other devices through a local area network or other wireless communication network.
[0078] The processor 401 may be a variety of general and / or special processing components with processing and computing capabilities. Some examples of the processor 401 include, but are not limited to, a central processing unit (CPU), a graphics processing unit (GPU), various dedicated artificial intelligence (AI) computing chips, various processors running machine learning model algorithms, digital signal processors (DSPs), and any appropriate processors, controllers, microcontrollers, etc. The processor 401 performs the various methods and processes described above, such as the training method 200 of the fault prediction model or the fault early warning method 300. For example, in some embodiments, the training method 200 of the fault prediction model or the fault early warning method 300 may be implemented as a computer software program, which is tangibly contained in a computer-readable storage medium, such as a memory 408. In some embodiments, part or all of the computer program may be loaded and / or installed on the electronic device 400 via the ROM 402 and / or the communication unit 409. When the computer program is loaded into the RAM 403 and executed by the processor 401, one or more steps of the training method 200 of the fault prediction model or the fault early warning method 300 described above may be performed. Alternatively, in other embodiments, the processor 401 may be configured to execute the fault prediction model training method 200 or the fault warning method 300 in any other appropriate manner (eg, by means of firmware).
[0079] Various aspects of the present application are described herein with reference to the flowcharts and / or timing diagrams of the methods, devices (systems), and computer program products according to the exemplary embodiments of the present application. It should be understood that each step of the flowchart and / or timing diagram and the combination of the steps in the flowchart and / or timing diagram can be implemented by computer-readable program instructions.
[0080] These computer-readable program instructions can be provided to a processor, a general-purpose computer, a special-purpose computer, or a processing unit of other programmable data processing devices in the electronic device 400, thereby producing a machine, so that when these instructions are executed by the processing unit of the computer or other programmable data processing device, a device that implements the functions / steps specified in one or more steps in the flowchart and / or timing diagram is generated. These computer-readable program instructions can also be stored in a computer-readable storage medium, and these instructions cause the computer, programmable data processing device, and / or other equipment to work in a specific manner, so that the computer-readable medium storing the instructions includes a manufactured product, which includes instructions for implementing various aspects of the functions / steps specified in one or more steps in the flowchart and / or timing diagram.
[0081] Computer-readable program instructions may also be loaded onto a computer, other programmable data processing apparatus, or other device so that a series of operational steps are performed on the computer, other programmable data processing apparatus, or other device to produce a computer-implemented process, thereby causing the instructions executed on the computer, other programmable data processing apparatus, or other device to implement the functions / steps specified in one or more steps in the flowchart and / or timing diagram.
[0082] The flowchart and timing diagram in the accompanying drawings show the possible architecture, functions and operations of the equipment, methods and computer program products according to multiple embodiments of the present application. In this regard, each step in the flowchart or timing diagram can represent a part of a module, a program segment or an instruction, and a part of the module, a program segment or an instruction contains one or more executable instructions for realizing the specified logical function. In some alternative embodiments, the functions marked in the steps can also occur in an order different from that marked in the accompanying drawings. For example, two consecutive steps can actually be executed substantially in parallel, and they can sometimes be executed in the opposite order, depending on the functions involved. It should also be noted that each step in the timing diagram and / or the flowchart, and the combination of the steps in the timing diagram and / or the flowchart, can be implemented with a dedicated hardware-based system that performs a specified function or action, or can be implemented with a combination of dedicated hardware and computer instructions.
[0083] The above description is only an implementation method of the present application and an explanation of the technical principles used. Those skilled in the art should understand that the scope of protection involved in the present application is not limited to the technical solution formed by a specific combination of the above technical features, but should also cover other technical solutions formed by any combination of the above technical features or their equivalent features without departing from the technical concept. For example, the above features are replaced with (but not limited to) technical features with similar functions disclosed in this application.
Claims
1. A fault warning method, characterized in that: include: Get vehicle data within the specified time step; Based on the pre-trained fault prediction model, determine the vehicle fault prediction result according to the vehicle data and the pre-trained vehicle fault knowledge graph; as well as In response to the fault prediction result indicating that the vehicle has a fault risk, performing a warning operation, Among them, the network model of the fault prediction model is a long short-term memory (LSTM) network model.
2. The method according to claim 1, wherein: The model of the knowledge graph is a TransGCN model, and the TransGCN model includes a TransR model and a graph convolutional network.
3. The method according to claim 2, wherein: The TransR model includes entity information and information representing the relationship between entities. The entity information includes information characterizing vehicle components, vehicle signals, fault phenomena and fault types. The relationship information includes at least: first relationship information indicating the relationship between the vehicle signal, the vehicle component and the fault phenomenon, second relationship information indicating the relationship between the vehicle component and the fault type, third relationship information between the vehicle components, and fourth relationship information between the vehicle signals.
4. The method according to any one of claims 1 to 3, wherein: The fault prediction model is a fault prediction model for the vehicle component. The pre-trained fault prediction model determines the vehicle fault prediction result according to the vehicle data and the pre-trained vehicle fault knowledge graph, including: The associated data related to the vehicle component in the vehicle data and the entities and relationships related to the vehicle component in the knowledge graph are input into the fault prediction model to obtain the fault prediction result of the vehicle component.
5. The method according to claim 4, wherein: The fault prediction model is at least used to determine the probability that the vehicle component represented by the associated data has a fault, and the method further includes: In response to the fault prediction result indicating that the probability that the vehicle component has a fault is greater than a warning threshold, it is determined that the fault prediction result indicates that the vehicle has a fault risk.
6. The method according to claim 5, wherein: The fault prediction model is at least used to determine the probability of occurrence of each fault type of the vehicle component represented by the associated data, and in response to the fault prediction result indicating that the probability of the vehicle component having a fault is greater than a warning threshold, determining that the fault prediction result indicates that the vehicle has a fault risk includes: In response to the fault prediction result indicating that the probability of occurrence of any of the fault types of the vehicle component is greater than a warning threshold of the fault type, it is determined that the fault prediction result indicates that the vehicle has a fault risk.
7. The method according to claim 6, wherein: The warning operation includes: Feedback is given of the fault type of the vehicle component whose occurrence probability is greater than the warning threshold.
8. A method for training a fault prediction model, characterized in that: include: Obtaining a pre-trained knowledge graph of vehicle faults and vehicle fault data, wherein the vehicle fault data is at least distinguished and marked as no fault or fault; as well as The knowledge graph and the vehicle fault data are used to train a long short-term memory (LSTM) network model to obtain a fault prediction model.
9. The method according to claim 8, wherein: The data of the fault data of the vehicle having a fault are differentiated and marked as at least one fault type.
10. The method according to claim 8 or 9, wherein: The using the knowledge graph and the vehicle fault data to train the long short-term memory (LSTM) network model to obtain a fault prediction model includes: Different long short-term memory (LSTM) network models are trained using the knowledge graph and associated data related to each vehicle component in the vehicle fault data to obtain a fault prediction model for each vehicle component.
11. An electronic device, characterized in that: include: at least one processor; as well as, A memory communicatively connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor, the instructions being executed by the at least one processor so that the at least one processor can execute the fault warning method as described in any one of claims 1 to 7 or the training method of the fault prediction model as described in any one of claims 8 to 10.
12. A computer-readable storage medium storing a computer program, characterized in that: When the computer program is executed by a processor, the fault warning method according to any one of claims 1 to 7 or the training method of the fault prediction model according to any one of claims 8 to 10 is implemented.
Citation Information
Cited By
Equipment fault early warning method and device, equipment and storage medium
CN119249285A
Fault identification method and device based on special vehicle control system
CN119292244A
