Vehicle Fault Diagnosis Methods and Systems

Abstract

This invention provides a vehicle fault diagnosis method and system. The system includes: an on-board data acquisition terminal, an on-board edge computing platform, and a cloud platform. The on-board data acquisition terminal acquires vehicle bus data and vehicle noise and / or vibration data. The on-board edge computing platform synchronizes the vehicle bus data and noise and / or vibration data in time, and uses a fault identification model to identify faults in the time-synchronized data. When a fault is identified, the platform sends the identified fault data to the cloud platform. The cloud platform receives the fault data sent by the on-board edge computing platform, analyzes the fault data using a fault analysis model, obtains the vehicle fault analysis results, and sends them back to the on-board edge computing platform. This invention utilizes vehicle bus data and noise and / or vibration data to locate vehicle faults, improving the accuracy of fault location.

Description

Technical Field

[0001] This invention relates to the field of vehicle fault diagnosis technology, and more specifically, to a vehicle fault diagnosis method and system. Background Technology

[0002] With the development of Internet of Things (IoT) and cloud computing technologies, automakers are also undergoing digital and service-oriented transformation, striving to continuously improve product quality while seeking new ways to serve customers.

[0003] When a car malfunctions, the frequency and amplitude of mechanical vibrations differ significantly from normal conditions, resulting in abnormal vibrations and noises, particularly noticeable in the engine, transmission, and other power or transmission components. Currently, diagnosing car malfunctions mainly relies on the experience of repair personnel through abnormal noise, but most drivers lack this ability. This method lacks timeliness and universality, failing to allow drivers to promptly and accurately recognize when a vehicle malfunction has occurred.

[0004] To overcome the shortcomings of existing technologies, a device for real-time monitoring of abnormal vehicle noise has been proposed. This device is composed of a noise acquisition module, a signal processor, and a main controller, and is capable of automatically judging and identifying whether major components of a vehicle are malfunctioning, and can promptly issue warning signals and fault information. The device includes a noise acquisition module, a signal processing module, a main controller module, an alarm notification module, a remote network module, and a fault diagnosis and monitoring module. The data output of the noise acquisition module is connected to the data input of the signal processing module, which in turn is connected to the data input of the main controller module. The data output of the main controller module is connected to the data inputs of both the noise acquisition module and the alarm notification module. The data output of the alarm notification module is connected to the data input of the remote network module, which in turn is connected to the data input of the fault diagnosis and monitoring module. Based on the detected abnormal vibration and noise from parts such as the engine and transmission, this device determines the location of the malfunction, issues a mobile phone notification, and transmits information to the repair shop for real-time fault diagnosis and preventative monitoring.

[0005] In practical applications, due to the complexity of automotive fault types, the aforementioned devices cannot accurately locate faults using only single noise data. Therefore, there is an urgent need for a solution that can improve the accuracy of fault location. Summary of the Invention

[0006] The technical problem to be solved by the embodiments of the present invention is to provide a vehicle fault diagnosis method and system to improve the accuracy of vehicle fault location.

[0007] To address the aforementioned technical problems, this invention provides a vehicle fault diagnosis system, comprising: an onboard data acquisition terminal, an onboard edge computing platform, and a cloud platform; wherein,

[0008] The vehicle-mounted data acquisition terminal is used to collect vehicle bus data and vehicle noise and / or vibration data, and send the collected data to the vehicle-mounted edge computing platform.

[0009] The vehicle-mounted edge computing platform is used to synchronize the bus data and noise and / or vibration data of the vehicle in time, and to use a fault identification model to identify faults in the time-synchronized data, and when a fault is identified, to send the identified fault data to the cloud platform.

[0010] The cloud platform is used to receive fault data sent by the vehicle edge computing platform, analyze the fault data using a fault analysis model, obtain vehicle fault analysis results, and send them to the vehicle edge computing platform. The fault analysis results include fault type and / or fault handling suggestions.

[0011] Optionally, the vehicle-mounted data acquisition terminal includes a vehicle bus data acquisition module and a noise and vibration acquisition module, wherein,

[0012] The vehicle bus data acquisition module listens to or actively polls the vehicle bus interface to read the vehicle's bus data.

[0013] The noise and vibration acquisition module collects noise and / or vibration data inside and outside the vehicle through noise sensors and / or vibration sensors installed on the vehicle.

[0014] Optionally, the in-vehicle edge computing platform includes: a data synchronization and preprocessing module, a fault identification module, a data upload module, and a result display module; wherein,

[0015] The data synchronization and preprocessing module is used to synchronize the bus data and noise and / or vibration data of the vehicle in time, and to send the time-synchronized data to the fault identification module and the data upload module after performing a first filtering process.

[0016] The fault identification module is used to identify faults in the data sent by the data synchronization and preprocessing module using a fault identification model, and when a fault is identified, to send the identified fault data to the data upload module.

[0017] The data upload module is used to send fault data from the fault identification module to the cloud platform, and to send data from the data synchronization and preprocessing module to the cloud platform according to a preset data sending strategy.

[0018] The result display module is used to display the fault information identified by the fault identification module, and / or to display the fault analysis results sent by the cloud platform.

[0019] Optionally, the cloud platform includes: a cloud gateway, a third-party system integration gateway, a filtering and feature extraction module, a fault record analysis module, a data storage module, a model training module, a model management module, and a fault analysis module, wherein,

[0020] The cloud gateway is used to connect and manage multiple vehicle edge computing platforms, receive data sent by the vehicle edge computing platforms, and / or send the vehicle fault analysis results obtained by the fault analysis module to the vehicle edge computing platforms.

[0021] The filtering and feature extraction module is used to perform format conversion, filtering and feature extraction on the data from the vehicle edge computing platform, and transmit the extracted fault feature data to the data storage module.

[0022] The fault record analysis module is used to extract fault types from the received vehicle fault handling records and associate the extracted fault types with the vehicle's fault data records.

[0023] The data storage module is used to store the fault feature data extracted by the filtering and feature extraction module, the fault type extracted by the fault record analysis module, and the data received by the cloud gateway from the vehicle edge computing platform.

[0024] The model training module is used to add labels to the fault feature data of the data storage module according to the fault type extracted by the fault record analysis module, update the training data, and retrain the fault identification model and the fault analysis model using a preset training algorithm and the updated training data.

[0025] The model management module is used for version management, storage and publishing of fault identification models and fault analysis models. It publishes the trained fault identification model to the vehicle edge computing platform and the trained fault analysis model to the fault analysis module.

[0026] The fault analysis module is used to analyze the data stored in the data storage module using a fault analysis model to obtain the fault analysis results of the vehicle.

[0027] The third-party system integration gateway is used to receive vehicle fault handling records sent by the third-party system and send them to the fault record analysis module, and / or send the vehicle fault analysis results obtained by the fault analysis module to the third-party system.

[0028] Optionally, the fault record analysis module is further configured to extract fault handling suggestions corresponding to the fault type from the received vehicle fault handling records, and associate the extracted fault handling suggestions with the corresponding fault type and save them in the knowledge base of the data storage module.

[0029] The fault analysis module is also used to search the knowledge base after obtaining the fault type of the vehicle using the fault analysis model, and obtain fault handling suggestions corresponding to the fault type.

[0030] Optionally, the vehicle's bus data frame includes data of n different bus parameters polled at fixed time intervals Δt;

[0031] The time synchronization of the vehicle's bus data with noise and / or vibration data includes:

[0032] The collected noise and / or vibration data and their sampling time are recorded as a single data record. The data records are then sorted according to the sampling time to obtain a first data record table of the noise and / or vibration data.

[0033] The engine's first speed data is collected by a speed sensor connected to the crankshaft of the vehicle engine, and the corresponding sampling time is recorded. The collected first speed data and its sampling time are recorded as a data record, and the data records of the first speed data are sorted according to the sampling time to obtain the first data record table of the first speed data.

[0034] The system receives bus data frames from the vehicle, wherein the n different bus parameters include the engine's second speed data; based on the reception time of the bus data frame and the fixed time interval, it extracts the data of each bus parameter from the same bus data frame and determines the corresponding sampling time; it treats the data of the same bus parameter in multiple bus data frames and its sampling time as a data record, and sorts the data records of the same bus parameter according to the sampling time to obtain a second data record table for each bus parameter; and it resamples and interpolates the second data record table for each bus parameter according to the sampling time interval of the first speed data to obtain a third data record table for each bus parameter.

[0035] According to the sampling time, the rotational speed data in the first data record table of the first rotational speed data and the third data record table of the second rotational speed data are aligned, and the difference value between L consecutive aligned rotational speed data is calculated; and, each time, the data record in the third data record table of the second rotational speed data is shifted one position to the left, and the difference value between L consecutive aligned rotational speed data is recalculated.

[0036] The time deviation between the first data record table of the first rotational speed data and the third data record table of the second rotational speed data is determined based on the smallest of the difference values ​​obtained from multiple calculations.

[0037] Based on the time deviation, the sampling time in the third data record table of each bus parameter is updated, and based on the updated sampling time, the first data record table of the noise and / or vibration data is aligned with the third data record table of each bus parameter.

[0038] This invention also provides a vehicle fault diagnosis method, including:

[0039] The vehicle's bus data, noise and / or vibration data are collected through the vehicle-mounted acquisition terminal, and the collected data is sent to the vehicle-mounted edge computing platform.

[0040] The vehicle edge computing platform synchronizes the vehicle's bus data with noise and / or vibration data in time, and uses a fault identification model to identify faults in the time-synchronized data. When a fault is identified, the fault data is sent to the cloud platform.

[0041] The cloud platform receives fault data sent by the vehicle edge computing platform, analyzes the fault data using a fault analysis model, obtains vehicle fault analysis results, and sends them to the vehicle edge computing platform. The fault analysis results include fault type and / or fault handling suggestions.

[0042] Optionally, the vehicle fault diagnosis method further includes:

[0043] The vehicle bus data is read by listening to or actively polling the vehicle bus interface.

[0044] Noise and / or vibration data inside and outside the vehicle are collected using noise and / or vibration sensors installed on the vehicle.

[0045] Optionally, the in-vehicle edge computing platform includes: a data synchronization and preprocessing module, a fault identification module, a data upload module, and a result display module; the method further includes:

[0046] The data synchronization and preprocessing module synchronizes the vehicle's bus data with noise and / or vibration data in time, and then performs a first filtering process on the time-synchronized data before sending it to the fault identification module and the data upload module.

[0047] The fault identification module uses a fault identification model to identify faults in the data sent by the data synchronization and preprocessing module, and when a fault is identified, the fault data is sent to the data upload module.

[0048] The data upload module sends fault data from the fault identification module to the cloud platform, and sends data from the data synchronization and preprocessing module to the cloud platform according to a preset data sending strategy.

[0049] The result display module displays the fault information identified by the fault identification module, and / or displays the fault analysis results sent by the cloud platform.

[0050] Optionally, the cloud platform includes: a cloud gateway, a third-party system integration gateway, a filtering and feature extraction module, a fault record analysis module, a data storage module, a model training module, a model management module, and a fault analysis module; the method further includes:

[0051] The cloud gateway connects to and manages multiple vehicle edge computing platforms, receives data sent by the vehicle edge computing platforms, and / or sends the vehicle fault analysis results obtained by the fault analysis module to the vehicle edge computing platforms.

[0052] The filtering and feature extraction module converts, filters, and extracts features from the data from the vehicle edge computing platform, and transmits the extracted fault feature data to the data storage module.

[0053] The fault record analysis module extracts the fault type from the received vehicle fault handling records and associates the extracted fault type with the vehicle's fault data records.

[0054] The data storage module stores the fault feature data extracted by the filtering and feature extraction module, the fault type extracted by the fault record analysis module, and the data received by the cloud gateway from the vehicle edge computing platform.

[0055] The model training module adds labels to the fault feature data of the data storage module based on the fault type extracted by the fault record analysis module, and updates the training data; and retrains the fault identification model and the fault analysis model using a preset training algorithm and the updated training data.

[0056] The model management module manages, stores, and publishes versions of the fault identification model and the fault analysis model. The trained fault identification model is published to the vehicle edge computing platform, and the trained fault analysis model is published to the fault analysis module.

[0057] The fault analysis module uses a fault analysis model to analyze the data stored in the data storage module to obtain the vehicle's fault analysis results.

[0058] The third-party system integration gateway receives vehicle fault handling records sent by the third-party system and sends them to the fault record analysis module, and / or sends the vehicle fault analysis results obtained by the fault analysis module to the third-party system.

[0059] Optionally, the method further includes:

[0060] The fault record analysis module extracts fault handling suggestions corresponding to the fault type from the received vehicle fault handling records, and associates the extracted fault handling suggestions with the corresponding fault type and saves them in the knowledge base of the data storage module.

[0061] The fault analysis module obtains the vehicle's fault type using the fault analysis model, then searches the knowledge base to obtain fault handling suggestions corresponding to the fault type.

[0062] Optionally, the vehicle's bus data frame includes data of n different bus parameters polled at fixed time intervals Δt;

[0063] The time synchronization of the vehicle's bus data with noise and / or vibration data includes:

[0064] The collected noise and / or vibration data and their sampling time are recorded as a single data record. The data records are then sorted according to the sampling time to obtain a first data record table of the noise and / or vibration data.

[0065] The engine's first speed data is collected by a speed sensor connected to the crankshaft of the vehicle engine, and the corresponding sampling time is recorded. The collected first speed data and its sampling time are recorded as a data record, and the data records of the first speed data are sorted according to the sampling time to obtain the first data record table of the first speed data.

[0066] The system receives bus data frames from the vehicle, wherein the n different bus parameters include the engine's second speed data; based on the reception time of the bus data frame and the fixed time interval, it extracts the data of each bus parameter from the same bus data frame and determines the corresponding sampling time; it treats the data of the same bus parameter in multiple bus data frames and its sampling time as a data record, and sorts the data records of the same bus parameter according to the sampling time to obtain a second data record table for each bus parameter; and it resamples and interpolates the second data record table for each bus parameter according to the sampling time interval of the first speed data to obtain a third data record table for each bus parameter.

[0067] According to the sampling time, the rotational speed data in the first data record table of the first rotational speed data and the third data record table of the second rotational speed data are aligned, and the difference value between L consecutive aligned rotational speed data is calculated; and, each time, the data record in the third data record table of the second rotational speed data is shifted one position to the left, and the difference value between L consecutive aligned rotational speed data is recalculated.

[0068] The time deviation between the first data record table of the first rotational speed data and the third data record table of the second rotational speed data is determined based on the smallest of the difference values ​​obtained from multiple calculations.

[0069] Based on the time deviation, the sampling time in the third data record table of each bus parameter is updated, and based on the updated sampling time, the first data record table of the noise and / or vibration data is aligned with the third data record table of each bus parameter.

[0070] Compared with existing technologies, the vehicle fault diagnosis method and system provided in this invention utilize vehicle bus data and noise and / or vibration data to locate vehicle faults, thereby improving the accuracy of fault location. Furthermore, this invention also sends the identified fault data to a cloud platform for processing, reducing the processing performance requirements of the in-vehicle edge computing platform and alleviating its computational load. Manufacturers and service providers can use the vehicle fault diagnosis method and system of this invention to provide services to customers, and it can also be used to analyze product quality issues and enhance product competitiveness. Attached Figure Description

[0071] To more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings used in the description of the embodiments of the present invention will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0072] Figure 1 This is a schematic diagram of the structure of a vehicle fault diagnosis system provided in an embodiment of the present invention;

[0073] Figure 2 This is an example diagram illustrating the acquisition of rotational speed data using a rotational speed sensor, as per an embodiment of the present invention.

[0074] Figure 3 This is a schematic diagram of the data structure of noise / vibration data and bus data stream in an embodiment of the present invention;

[0075] Figure 4 This is an example diagram of a first data recording table of rotational speed data from a rotational speed sensor and a second data recording table of rotational speed from a vehicle bus, according to an embodiment of the present invention.

[0076] Figure 5 This is an example diagram illustrating the calculation of time deviations between data recording tables in an embodiment of the present invention;

[0077] Figure 6 This is an example diagram of the third data record table for the second rotational speed data after updating the sampling time, according to an embodiment of the present invention.

[0078] Figure 7 An example diagram illustrating the updating of the sampling time corresponding to each bus parameter in each bus data frame according to an embodiment of the present invention;

[0079] Figure 8 This is a schematic flowchart of a vehicle fault diagnosis method according to an embodiment of the present invention;

[0080] Figure 9 This is a schematic diagram illustrating the data processing flow of the vehicle-mounted signal acquisition terminal and the edge computing platform according to an embodiment of the present invention;

[0081] Figure 10 This is a schematic diagram illustrating the process of cloud platform processing IoT data and diagnosing faults according to an embodiment of the present invention;

[0082] Figure 11 This is a flowchart illustrating the interaction between a cloud platform and a third-party system according to an embodiment of the present invention.

[0083] Figure 12 This is another interaction flowchart between the cloud platform and the third-party system according to an embodiment of the present invention;

[0084] Figure 13 This is a flowchart illustrating the model training and deployment process according to an embodiment of the present invention;

[0085] Figure 14 Example diagrams of noise or vibration signal data from two channels CH1 and CH2 and data from the vehicle bus in an embodiment of the present invention. Detailed Implementation

[0086] To make the technical problems, technical solutions, and advantages of this invention clearer, a detailed description will be provided below in conjunction with the accompanying drawings and specific embodiments. In the following description, specific details such as particular configurations and components are provided merely to aid in a comprehensive understanding of the embodiments of this invention. Therefore, those skilled in the art should understand that various changes and modifications can be made to the embodiments described herein without departing from the scope and spirit of this invention. Furthermore, for clarity and brevity, descriptions of known functions and structures have been omitted.

[0087] It should be understood that the phrase "one embodiment" or "an embodiment" throughout the specification means that a specific feature, structure, or characteristic related to the embodiment is included in at least one embodiment of the invention. Therefore, "in one embodiment" or "in an embodiment" appearing throughout the specification do not necessarily refer to the same embodiment. Furthermore, these specific features, structures, or characteristics can be combined in any suitable manner in one or more embodiments.

[0088] In various embodiments of the present invention, it should be understood that the sequence number of each process described below does not imply the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the present invention.

[0089] As described in the background section, some existing vehicle fault diagnosis methods suffer from problems such as low fault location accuracy. To address at least one of these problems, this invention provides a vehicle fault diagnosis system that uses vehicle bus data and noise and / or vibration data to locate vehicle faults, thereby improving the accuracy of fault location.

[0090] Please refer to Figure 1 This invention provides a vehicle fault diagnosis system, including a cloud platform, an in-vehicle data acquisition terminal, an in-vehicle edge computing platform, and a cloud platform. Among these,

[0091] The vehicle-mounted data acquisition terminal is used to collect vehicle bus data and vehicle noise and / or vibration data, and send the collected data to the vehicle-mounted edge computing platform.

[0092] The vehicle-mounted edge computing platform is used to synchronize the bus data and noise and / or vibration data of the vehicle in time, and to use a fault identification model to identify faults in the time-synchronized data, and when a fault is identified, to send the identified fault data to the cloud platform.

[0093] The cloud platform is used to receive fault data sent by the vehicle edge computing platform, analyze the fault data using a fault analysis model, obtain vehicle fault analysis results, and send them to the vehicle edge computing platform. The fault analysis results include fault type and / or fault handling suggestions.

[0094] Here, the fault identification model can be pre-trained by the cloud platform and published to the vehicle edge computing platform. The fault analysis model can also be pre-trained by the cloud platform. Furthermore, in practical applications, the cloud platform can update the training sets of the aforementioned fault identification model and fault analysis model based on collected vehicle maintenance data and vehicle data sent by the vehicle edge computing platform, and train and update the aforementioned models based on the updated training sets, and then publish the updated models.

[0095] Considering the computing power of the in-vehicle edge computing platform and reducing its computational load, this embodiment of the invention utilizes the aforementioned fault identification model at the in-vehicle edge computing platform to identify the presence of a fault. Then, if a fault is identified, the fault data, specifically data within a preset time period before and after the fault identification (including noise and / or vibration data, and vehicle bus time), is sent to the cloud platform. The cloud platform then uses a fault analysis model to analyze this data, identify the specific fault type, and can further match fault handling suggestions corresponding to the fault type. It should be noted that the aforementioned fault identification model and fault analysis model can be trained using training data through machine learning algorithms such as clustering.

[0096] like Figure 1 As shown, the vehicle-mounted data acquisition terminal is typically installed in a vehicle and may specifically include a vehicle bus data acquisition module and a noise and vibration acquisition module, wherein:

[0097] The vehicle bus data acquisition module reads vehicle bus data by listening to or actively polling the vehicle bus interface. For example, the vehicle bus data acquisition module can connect to the vehicle bus through an on-board diagnostics (OBD) interface or other bus interfaces, and read vehicle bus data by listening to or actively polling.

[0098] The noise and vibration acquisition module collects noise and / or vibration data inside and outside the vehicle through noise sensors and / or vibration sensors installed on the vehicle.

[0099] Here, the noise and vibration acquisition module can be connected to multiple sensors to collect noise and vibration signals inside or outside the vehicle without damage.

[0100] like Figure 1 As shown, the vehicle-mounted edge computing platform is responsible for preprocessing the data uploaded by the vehicle-mounted data acquisition terminal and uploading the data to the cloud platform in batch or real-time data stream mode. It can also display the analysis results or raw data. The edge computing platform is typically installed inside the vehicle and may include: a data synchronization and preprocessing module, a fault identification module, a data upload module, and a result display module, wherein:

[0101] The data synchronization and preprocessing module is used to synchronize the vehicle's bus data with noise and / or vibration data in time, and to send the time-synchronized data to the fault identification module and the data upload module after performing a first filtering process.

[0102] Here, the data synchronization and preprocessing module can perform basic filtering on the data signal, such as finite impulse response (FIR) filtering, order filtering, etc.

[0103] The fault identification module is used to identify faults in the data sent by the data synchronization and preprocessing module using a fault identification model, and when a fault is identified, the fault data is sent to the data upload module.

[0104] Here, the fault identification module can identify data streams containing fault modes in a timely manner by monitoring the data stream in real time, and transmit them to the cloud platform in real time.

[0105] The data upload module is used to send fault data from the fault identification module to the cloud platform, and to send data from the data synchronization and preprocessing module to the cloud platform according to a preset data sending strategy.

[0106] Here, the data upload module supports both asynchronous batch upload and real-time data stream upload. Since lossless noise and vibration data signals consume a large amount of bandwidth, to conserve transmission resources, as one implementation method, the fault identification module only transmits faulty data streams to the cloud platform in real time. Normal data streams can be uploaded to the cloud platform in batches when bandwidth is available, or, according to settings, only necessary data can be transmitted to save bandwidth and storage space.

[0107] The result display module is used to display the fault information identified by the fault identification module, and / or to display the fault analysis results sent by the cloud platform.

[0108] Here, the results display module can provide real-time display of data streams in the form of charts or graphs, and can also display the analysis results fed back by the cloud platform to facilitate operation by fault diagnosis personnel in the vehicle.

[0109] like Figure 1 As shown, the cloud platform supports functions such as big data storage, model training, and fault cause analysis. The cloud platform can be a public cloud, a private cloud, or a hybrid cloud. Specifically, the cloud platform may include: a cloud gateway, a third-party system integration gateway, a filtering and feature extraction module, a fault record analysis module, a data storage module, a model training module, a model management module, and a fault analysis module.

[0110] The cloud gateway is used to connect and manage multiple vehicle edge computing platforms, receive data sent by the vehicle edge computing platforms, and / or send the vehicle fault analysis results obtained by the fault analysis module to the vehicle edge computing platforms.

[0111] The filtering and feature extraction module is used to perform format conversion, filtering and feature extraction on the data from the vehicle edge computing platform, and transmit the extracted fault feature data to the data storage module.

[0112] Here, the filtering and feature extraction module performs format conversion, advanced filtering, feature extraction, and other processing on the data from the edge computing platform, and then transmits it to the corresponding data storage module.

[0113] The fault record analysis module is used to extract fault types from the received vehicle fault handling records and associate the extracted fault types with the vehicle's fault data records.

[0114] The data storage module supports both structured and unstructured data storage. It is used to store fault feature data extracted by the filtering and feature extraction module, fault types extracted by the fault record analysis module, and data received by the cloud gateway from the vehicle edge computing platform.

[0115] The model training module is used to add labels to the fault feature data of the data storage module according to the fault type extracted by the fault record analysis module, update the training data, and retrain the fault identification model and the fault analysis model using a preset training algorithm and the updated training data.

[0116] The model management module is used for version management, storage and publishing of fault identification models and fault analysis models. It publishes the trained fault identification model to the vehicle edge computing platform and the trained fault analysis model to the fault analysis module.

[0117] Here, the model training module supports automatically labeling fault feature data from the data storage according to the identified fault type, and using a preset algorithm to retrain the model based on the latest data. It also supports manual-assisted model training.

[0118] The fault analysis module is used to analyze the data stored in the data storage module using a fault analysis model, thereby discovering the cause of the fault, such as the fault type, based on the fault mode, and obtaining the fault analysis results of the vehicle.

[0119] The third-party system integration gateway can perform data transmission and data format conversion between the cloud platform and third-party systems (such as the vehicle maintenance system of a car dealer). Specifically, it can be used to receive vehicle fault handling records sent by the third-party system and send them to the fault record analysis module, and / or send the vehicle fault analysis results obtained by the fault analysis module to the third-party system.

[0120] Furthermore, in the cloud platform, the fault record analysis module is also used to extract fault handling suggestions corresponding to the fault type from the received vehicle fault handling records, and associate the extracted fault handling suggestions with the corresponding fault type and store them in the knowledge base of the data storage module. The fault analysis module is also used to, after obtaining the vehicle's fault type using the fault analysis model, search the knowledge base to obtain the fault handling suggestions corresponding to the fault type. In this way, the fault analysis module can not only analyze the specific fault type, but also provide the corresponding fault handling suggestions, and send the fault type and fault handling suggestions to the result display module of the in-vehicle edge computing platform and / or to a third-party system, such as a 4S shop's maintenance system, through a third-party system integration gateway.

[0121] As can be seen, the vehicle fault diagnosis system of this invention collects noise and / or vibration data inside and outside the vehicle through noise sensors and / or vibration sensors installed on the vehicle, and combines this data with the vehicle's bus data to identify vehicle faults. Through the above modules, this invention fully utilizes noise and / or vibration data inside and outside the vehicle, and combines this data with the vehicle's bus data for fault identification, thereby improving the accuracy of vehicle fault location.

[0122] In this embodiment of the invention, noise data is collected using a noise sensor installed on the vehicle, and vibration data is collected using a vibration sensor installed on the vehicle. Of course, the vibration sensor and noise sensor can also be integrated into the same device.

[0123] The noise / vibration data stream is acquired in real time using a high-speed A / D chip in the sensor, and a GPS time stamp is added simultaneously during acquisition. The vehicle bus data stream, however, originates from polling the vehicle bus. Due to delays in the polling process, vehicle bus response time, and data transmission, there is a time discrepancy between the vehicle bus data stream and the noise / vibration signal. To improve the accuracy of fault analysis, this embodiment of the invention uses a synchronization mechanism to eliminate or reduce this time discrepancy, ensuring that the bus data is time-consistent or nearly consistent with the noise and / or vibration data.

[0124] To achieve the aforementioned time synchronization, embodiments of the present invention provide a specific method for time synchronization. Specifically, as follows... Figure 2 As shown, in this embodiment of the invention, a speed sensor is installed on a sensor on the crankshaft of the vehicle engine, and an on-board data acquisition terminal collects the engine speed data. The speed sensor collects speed data at a frequency similar to that of the aforementioned noise / vibration sensor; both use dedicated sensors to collect relevant parameters in real time, and their collection frequency is typically much higher than that of the vehicle bus data. Therefore, the speed data collected by the speed sensor and the speed data contained in the vehicle bus data can be synchronized, ensuring that the data collected by the speed sensor is synchronized with the bus data time, which in turn ensures that the data collected by the noise / vibration sensor is synchronized with the bus data time.

[0125] Figure 3 A schematic diagram of the data structure for noise / vibration data and the bus data stream is provided. It is assumed that one noise or vibration signal is acquired through channel 1, and another noise or vibration signal is acquired through channel 2. These two noise or vibration signals are synchronized, either according to the reception time or the time stamp in the data. However, the vehicle bus data stream has a certain time deviation from the aforementioned noise or vibration signals, typically exhibiting a delay relative to them. Figure 3 The document describes a structure for a bus data frame, which includes data for n different bus parameters (such as PID1, PID2, PID3, ..., PIDn) that are polled at fixed time intervals Δt. That is, assuming the sampling time of PID1 (or the reception time of the data frame) is t, then the sampling time of the i-th bus parameter PIDi in the data frame (or the reception time of the data frame) is: t + (i-1)Δt.

[0126] One implementation method for the data synchronization and preprocessing module to synchronize the vehicle's bus data with noise and / or vibration data in time is as follows:

[0127] A) Collect the noise and / or vibration data and their sampling time as a data record, and sort the data records of the noise and / or vibration data according to the sampling time to obtain the first data record table of the noise and / or vibration data.

[0128] B) The engine's first speed data is collected by a speed sensor connected to the crankshaft of the vehicle engine, and the corresponding sampling time is recorded. The collected first speed data and its sampling time are recorded as a data record, and the data records of the first speed data are sorted according to the sampling time to obtain the first data record table of the first speed data. Figure 4 An example of a first data record table of rotational speed data from a rotational speed sensor is given, wherein each data record includes rotational speed data (e.g., RPM1) and the sampling time corresponding to that rotational speed data (e.g., time1).

[0129] C) Receive the bus data frame of the vehicle, wherein the n different bus parameters include the second speed data of the engine; extract the data of each bus parameter from the same bus data frame according to the reception time of the bus data frame and the fixed time interval Δt, and determine the corresponding sampling time; take the data of the same bus parameter in multiple bus data frames and its sampling time as a data record, and sort the data records of the same bus parameter according to the sampling time to obtain the second data record table of each bus parameter; and resample and interpolate the second data record table of each bus parameter according to the sampling time interval s1 of the first speed data to obtain the third data record table of each bus parameter.

[0130] Figure 4 An example of a second data record table for the rotational speed data in a bus data frame is also provided, where each data record includes the rotational speed data (e.g., rpm1) and the corresponding sampling time (e.g., time1). It should be noted that... Figure 4 The same time 'i' in the first and second data recording tables is not necessarily the same. Since the sampling frequency of the second data recording table is usually lower than that of the first data recording table, that is, there is more rotational speed data in the first data recording table within the same time period, in order to facilitate comparison, this embodiment of the invention resamples and interpolates the second data recording tables of each bus parameter so that the sampling frequency of the second data recording tables of each bus parameter after interpolation is the same as that of the first data recording table, so as to facilitate subsequent calculation and processing.

[0131] D) According to the sampling time, align the speed data in the first data record table of the first speed data and the third data record table of the second speed data, that is, align the data records with the same or similar sampling time. Then, calculate the difference value between L consecutive aligned speed data. Also, each time, shift the data record in the third data record table of the second speed data one position to the left and recalculate the difference value between L consecutive aligned speed data.

[0132] Here, L is a preset integer value, such as 5 or 10. The difference value can be represented by the mean square error between the L consecutive aligned speed data points, or by the root mean square error between the L consecutive aligned speed data points, or by summing the absolute values ​​of the differences between the L consecutive aligned speed data points and using the sum as the difference value. As one implementation, the difference value can also be calculated according to the following formula:

[0133]

[0134] Here, err represents the difference value, rpm i represents the i-th first speed data among the L corresponding speed data, and RPM i represents the i-th second speed data among the L corresponding speed data.

[0135] After shifting the data record in the third data record table of the second rotational speed data to the left by q positions, the deviation value can be calculated according to the following formula:

[0136]

[0137] E) Determine the time deviation between the first data record table of the first rotational speed data and the third data record table of the second rotational speed data based on the smallest of the differences obtained from multiple calculations.

[0138] Assuming that the sampling time in the above data recording table is gradually delayed from left to right, and the sampling time in the third data recording table is delayed, it is necessary to shift it to the left, and after each shift by one bit, recalculate the above difference value to find the minimum difference value, and take the time deviation between the first data recording table of the first speed data and the third data recording table of the second speed data at the minimum difference value as the time delay amount of the bus parameter.

[0139] F) Update the sampling time in the third data record table of each bus parameter according to the time deviation, and align the first data record table of the noise and / or vibration data with the third data record table of each bus parameter according to the updated sampling time.

[0140] Assuming the m-th bus parameter in the bus data frame is the second rotational speed data, and the deviation value calculated after shifting the data record in the third data record table of the second rotational speed data to the left by q positions is the minimum, that is, the time deviation corresponding to the minimum deviation value is: q*s1, where s1 represents the sampling time interval of the first rotational speed data, such as... Figure 5 As shown.

[0141] Therefore, the sampling time of each data record in the third data record table of the second rotational speed data needs to be subtracted from the above-mentioned time deviation in order to update its sampling time. Figure 6 An example of a third data record table for the second rotational speed data after updating the sampling time is given. The sampling time for other bus parameter point data record tables can also be updated in the same way. Then, the data records in the third data record table with the same or closest sampling times to the first data record table can be aligned to complete the time synchronization process.

[0142] As another implementation, embodiments of the present invention can also recalculate and update the sampling time corresponding to each bus parameter in each bus data frame based on the aforementioned time deviation, such as... Figure 7 As shown, assuming the sampling time of PID1 is t, and Δt is the sampling time interval between two adjacent bus parameters, the data of each bus parameter and the updated sampling time are extracted from the updated bus data frame, thereby updating the second data record table of each bus parameter. Then, according to the sampling time interval s1 of the first rotational speed data, the updated second data record table of each bus parameter is resampled and interpolated to obtain the updated third data record table of each bus parameter. Then, the data record table can be directly synchronized with the first data record table of the noise and / or vibration data based on the time in the data record table; that is, the data records with the same or closest sampling times are aligned.

[0143] After the data records are aligned, the sampling time can be deleted, or it can be retained and sent along with the data to the fault identification module or cloud platform for identification and analysis.

[0144] Based on the above vehicle fault diagnosis system, this invention also provides a vehicle fault diagnosis method, such as... Figure 8 As shown, the method includes:

[0145] Step 81: Collect vehicle bus data and vehicle noise and / or vibration data through the vehicle-mounted acquisition terminal, and send the collected data to the vehicle-mounted edge computing platform.

[0146] Here, in this embodiment of the invention, the vehicle-mounted data acquisition terminal can listen to or actively poll the vehicle bus interface to read the vehicle's bus data; and collect noise data and / or vibration data inside and outside the vehicle through noise sensors and / or vibration sensors installed on the vehicle.

[0147] Step 82: Through the vehicle edge computing platform, the bus data and noise and / or vibration data of the vehicle are synchronized in time. The fault identification model is used to identify faults in the time-synchronized data. When a fault is identified, the fault data is sent to the cloud platform.

[0148] Step 83: Receive fault data sent by the vehicle edge computing platform through the cloud platform, analyze the fault data using a fault analysis model, obtain the vehicle fault analysis results, and send them to the vehicle edge computing platform. The fault analysis results include fault type and / or fault handling suggestions.

[0149] Through the above steps, this embodiment of the invention utilizes vehicle bus data, noise and / or vibration data to locate vehicle faults, thereby improving the accuracy of fault location. Furthermore, this embodiment of the invention also sends the identified fault data to the cloud platform for processing, which reduces the processing performance requirements of the in-vehicle edge computing platform and alleviates its computational load.

[0150] Here, the vehicle's bus data frame includes data from n different bus parameters polled at fixed time intervals Δt. Synchronizing the vehicle's bus data with noise and / or vibration data may specifically include the following steps:

[0151] Step a: Collect the noise and / or vibration data and their sampling time as a data record, and sort the data records of the noise and / or vibration data according to the sampling time to obtain the first data record table of the noise and / or vibration data.

[0152] Step b: Collect the first engine speed data and record the corresponding sampling time by using a speed sensor connected to the crankshaft of the vehicle engine; take the collected first speed data and its sampling time as a data record, and sort the data records of the first speed data according to the sampling time to obtain the first data record table of the first speed data;

[0153] Step c: Receive the bus data frame of the vehicle, wherein the n different bus parameters include the second speed data of the engine; extract the data of each bus parameter from the same bus data frame and determine the corresponding sampling time according to the reception time of the bus data frame and the fixed time interval; take the data of the same bus parameter in multiple bus data frames and its sampling time as a data record, and sort the data records of the same bus parameter according to the sampling time to obtain the second data record table of each bus parameter; and resample and interpolate the second data record table of each bus parameter according to the sampling time interval of the first speed data to obtain the third data record table of each bus parameter.

[0154] Step d: According to the sampling time, align the rotational speed data in the first data record table of the first rotational speed data and the third data record table of the second rotational speed data, and calculate the difference value between L consecutive aligned rotational speed data; and each time, shift the data record in the third data record table of the second rotational speed data one position to the left, and recalculate the difference value between L consecutive aligned rotational speed data.

[0155] Step e: Determine the time deviation between the first data recording table of the first rotational speed data and the third data recording table of the second rotational speed data based on the smallest of the differences obtained from multiple calculations.

[0156] Step f: Based on the time deviation, update the sampling time in the third data recording table of each bus parameter, and align the first data recording table of the noise and / or vibration data with the third data recording table of each bus parameter based on the updated sampling time.

[0157] Figure 9 The data processing flow for the vehicle-mounted signal acquisition terminal and edge computing platform is given, including:

[0158] Step 91: Connect one or more noise and vibration sensors to the vehicle-mounted acquisition terminal to collect noise and vibration signals inside and outside the vehicle.

[0159] Step 92: Collect bus data of the vehicle through the vehicle-mounted acquisition terminal.

[0160] Step 93: The data synchronization and preprocessing module synchronizes multiple noise and vibration signals, as well as bus data from the vehicle bus, with time stamps accurate to the millisecond level.

[0161] Step 94: The data synchronization and preprocessing module supports filtering the data using built-in algorithms, and extracting data features from the synchronized data using basic algorithms such as frequency domain analysis.

[0162] Step 95: The fault identification module identifies data features based on the model updated from the cloud platform, initially identifying data records that may contain faults. Due to the computing power limitations of edge computing and the requirements of real-time processing, this process only performs preliminary fault identification and cannot determine the specific fault type or cause.

[0163] Step 96: Determine whether a data record containing a fault has been identified. If yes, proceed to step 97; otherwise, proceed to step 98.

[0164] Step 97: Data streams that may contain faults, identified in step 95, will be transmitted to the cloud platform in real time.

[0165] Step 98: If no normal data stream containing faults is identified in step 95, then upload the data in batches when the system bandwidth is idle, or upload only some types of data according to the settings to save bandwidth and storage.

[0166] Step 99: The results display module shows the real-time data stream and analysis feedback results from the cloud platform, allowing in-vehicle testers to easily conduct tests. This results display module can also be easily ported to smart handheld terminals such as smartphones.

[0167] As can be seen, the above method of the present invention can synchronize the bus data and noise and / or vibration data of the vehicle in time through the data synchronization and preprocessing module, and send the time-synchronized data to the fault identification module and the data upload module after performing a first filtering process.

[0168] The fault identification module uses a fault identification model to identify faults in the data sent by the data synchronization and preprocessing module, and when a fault is identified, the fault data is sent to the data upload module.

[0169] The data upload module sends fault data from the fault identification module to the cloud platform, and sends data from the data synchronization and preprocessing module to the cloud platform according to a preset data sending strategy.

[0170] The result display module displays the fault information identified by the fault identification module, and / or displays the fault analysis results sent by the cloud platform.

[0171] Figure 10 The process for cloud platforms to process IoT data and diagnose faults is also provided, including:

[0172] In steps 101-102, after receiving the data, the cloud IoT gateway first determines whether the data is a data stream already identified as containing a fault on the edge computing platform or normal batch archived data. If it is a data stream containing a fault, proceed to step 105; otherwise, proceed to steps 103-104.

[0173] In steps 103-104, if it is normal batch archived data, the filtering and feature extraction module only performs basic preprocessing, format conversion, etc. on the data and sends it to the data storage module.

[0174] Step 105: The filtering and feature extraction module first performs advanced filtering on the fault data stream. Unlike the basic filtering on the edge computing platform, the filtering on the cloud platform can perform more complex functions, such as filtering out human voices and wind noise, to achieve more accurate fault feature extraction and fault identification.

[0175] In steps 106-107, the data stream after advanced filtering is further filtered by the filtering and feature extraction module to extract fault features according to a preset algorithm, and then converted into structured data stored in the data storage module.

[0176] Step 108: The fault cause analysis module uses the built-in analysis model to obtain feature data from the data storage module in a timed or triggered manner, and uses machine learning algorithms such as clustering to locate the fault.

[0177] Step 109: If the fault can be located, i.e. the fault type can be predicted, proceed to step 110; otherwise, proceed to step 111.

[0178] Step 110: After predicting the fault type, historical processing methods for the same or similar fault types can be extracted from the knowledge base of the data storage module as fault processing suggestions, and sent together with the fault type to a third-party system (such as the maintenance system of a 4S store) and the vehicle edge computing platform.

[0179] Step 111: If the fault cannot be located, the process can be switched to manual handling for further confirmation.

[0180] As mentioned above, the cloud platform includes: a cloud gateway, a third-party system integration gateway, a filtering and feature extraction module, a fault record analysis module, a data storage module, a model training module, a model management module, and a fault analysis module. In this embodiment of the invention, the cloud gateway can connect to and manage multiple in-vehicle edge computing platforms, receive data sent by the in-vehicle edge computing platforms, and / or send the vehicle fault analysis results obtained by the fault analysis module to the in-vehicle edge computing platforms.

[0181] The filtering and feature extraction module converts, filters, and extracts features from the data from the vehicle edge computing platform, and transmits the extracted fault feature data to the data storage module.

[0182] The fault record analysis module extracts the fault type from the received vehicle fault handling records and associates the extracted fault type with the vehicle's fault data records.

[0183] The data storage module stores the fault feature data extracted by the filtering and feature extraction module, the fault type extracted by the fault record analysis module, and the data received by the cloud gateway from the vehicle edge computing platform.

[0184] The model training module adds labels to the fault feature data of the data storage module based on the fault type extracted by the fault record analysis module, and updates the training data; and retrains the fault identification model and the fault analysis model using a preset training algorithm and the updated training data.

[0185] The model management module manages, stores, and publishes versions of the fault identification model and the fault analysis model. The trained fault identification model is published to the vehicle edge computing platform, and the trained fault analysis model is published to the fault analysis module.

[0186] The fault analysis module uses a fault analysis model to analyze the data stored in the data storage module to obtain the vehicle's fault analysis results.

[0187] The third-party system integration gateway receives vehicle fault handling records sent by the third-party system and sends them to the fault record analysis module, and / or sends the vehicle fault analysis results obtained by the fault analysis module to the third-party system.

[0188] In this embodiment of the invention, the fault record analysis module can also extract fault handling suggestions corresponding to the fault type from the received vehicle fault handling records, associate the extracted fault handling suggestions with the corresponding fault type and store them in the knowledge base of the data storage module; and through the fault analysis module, after obtaining the vehicle fault type using the fault analysis model, the knowledge base is searched to obtain the fault handling suggestions corresponding to the fault type.

[0189] Figures 11-12 It also outlines the interaction process between the cloud platform and third-party systems, including, for example... Figure 11 As shown, it includes:

[0190] Steps 1101-1102: The third-party system integration gateway obtains relevant maintenance data from external systems (such as the maintenance system of a 4S store) by subscription or active query. The data includes the corresponding vehicle identification number, time, fault type, handling method, etc., and is stored after preliminary format conversion.

[0191] In steps 1103-1104, the fault type identification module processes the received data, identifies the fault category, categorizes and labels it, and stores it in a structured manner. Simultaneously, the fault type identification module extracts features from the description of the fault handling process and stores the extracted feature information as a knowledge base in the data storage module. Subsequently, if a similar fault type is diagnosed, the data stored in the knowledge base will be automatically associated as maintenance suggestions.

[0192] like Figure 12 As shown, it includes:

[0193] In steps 1201-1202, the third-party system integration gateway receives the analysis result data from the cloud platform's fault cause analysis module and converts it into a data format that the third-party system can recognize.

[0194] Step 1203: Send the data to a third-party system (such as the 4S store's maintenance system) for the 4S store staff to refer to as a fault indication and fault diagnosis aid.

[0195] Figure 13 The training and deployment process of the relevant models in this embodiment of the invention is given, with the model training module executed periodically to update the model. Two types of models are included: a fault identification module for in-vehicle edge computing platforms and a fault cause analysis module for cloud platforms. Due to the limited computing power of edge computing platforms and the high real-time requirements, only a small amount of data features can be extracted. The fault identification module can make a preliminary judgment on the existence of a fault based on these limited data features. Because it relies on fewer data features, the model parameters are adjusted to include as much data as possible that could indicate a fault. The fault cause analysis model for cloud platforms relies on more data features and the powerful computing power of the cloud platform, focusing on accurately identifying the type of fault. The training of both types of models follows the same process, specifically including:

[0196] In steps 1301-1302, the model training module first retrieves the identified fault feature dataset and fault type dataset from the data storage module, and then matches the two based on the vehicle's unique code and time. The model training module automatically labels the fault types and attaches the labels to the corresponding fault feature data.

[0197] In steps 1303-1304, the model training module trains the model parameters according to the preset algorithm, and publishes the model to the fault identification module of the vehicle edge computing platform or the fault analysis module of the cloud platform according to the type of the trained model.

[0198] The following embodiment of the present invention provides a specific example to further illustrate the above method in detail.

[0199] Figure 14 The data presented includes noise or vibration signal data from two channels, CH1 and CH2, and data from the vehicle bus. Since the noise and vibration signals originate from different signal sources, time synchronization is required before analysis. Because both data packets have time stamps, the data synchronization and preprocessing module synchronizes the two types of data based on the time stamps and the calculated transmission delay. Before analysis, the noise and vibration signals can be filtered using a combined digital filter to remove irrelevant frequency bands or orders.

[0200] The fault identification module performs preliminary analysis of noise and data signals to extract preliminary fault characteristics. For example, it converts the time-domain signal into a frequency-domain signal in real time using Fast Fourier Transform and then transforms it into structured data, as shown in Table 1.

[0201]

[0202] Table 1

[0203] The fault identification module classifies the data based on model parameters sent from the cloud platform. There are only two classification labels: faulty and fault-free. Faulty data streams are uploaded to the cloud platform in real time for further analysis. Fault-free data is stored locally and transmitted in batches only when the transmission queue is idle, or, according to settings, only the automotive bus data stream is uploaded, excluding noise and vibration signals to save bandwidth and data usage.

[0204] After receiving the fault data stream, the cloud platform first performs advanced filtering on the data stream according to the algorithms preset in the filtering and feature extraction modules. This includes identifying wind noise and road noise patterns and filtering them out of the original data. Then, feature quantities are extracted again. In addition to the features mentioned above, the signal is resampled based on the vehicle speed for order analysis, and the sound pressure level corresponding to each order is extracted. Furthermore, more data from the vehicle are combined for comprehensive analysis, as shown in Table 2.

[0205]

[0206] Table 2

[0207] The fault analysis module uses a machine learning classification algorithm on the aforementioned feature data based on pre-set model algorithms and parameters. Specifically, in addition to "faulty" and "no fault," the classification labels can be further subdivided into fault type labels. After obtaining the fault labels, the module extracts specific fault descriptions, possible causes, handling suggestions, and similar case analyses from the fault knowledge base. Finally, the fault analysis results, including the access URL of the original data records, are sent to a third-party system, such as a 4S store's maintenance system, via a third-party system integration gateway. Alternatively, the analysis results can be directly sent to an in-vehicle edge computing platform.

[0208] After receiving a maintenance request from the platform, the 4S dealership retrieves data from the platform's data storage module. This data includes possible fault label classifications, time-domain and frequency-domain graphs of noise and vibration signals, the original noise and vibration signals, the vehicle bus data stream, and processing suggestions provided by the platform. After analysis, the 4S dealership staff sends a service reminder to the user. Once the user receives the service request, the 4S dealership staff performs the repairs and records the repair details in the maintenance system, as shown in Table 3.

[0209]

[0210] Table 3

[0211] Maintenance records are simultaneously sent to the cloud platform via a third-party system integration gateway. The cloud platform's fault type identification module first extracts the fault tags from the records. If no fault tags are entered in the original fault records, the fault type identification module extracts keywords from the maintenance records, automatically generates associated tags, and re-marks the associated fault data records based on vehicle identification numbers, time, and other information. Simultaneously, it generates an associated knowledge base. The knowledge base includes possible causes of the fault, handling suggestions, historical cases, etc.

[0212] As the recorded fault type labels are constantly updated through maintenance records, in order to ensure the accuracy of the model, the model training module periodically learns from the updated fault data records, recalculates the module parameters, and publishes the updated parameters to the fault cause analysis module on the cloud platform and the fault identification module on the vehicle edge computing platform.

[0213] Those skilled in the art will recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementations should not be considered beyond the scope of this invention.

[0214] Those skilled in the art will understand that, for the sake of convenience and brevity, the specific working processes of the systems, devices, and units described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here.

[0215] In the embodiments provided in this application, it should be understood that the disclosed apparatus and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative. For instance, the division of units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces; the indirect coupling or communication connection between apparatuses or units may be electrical, mechanical, or other forms.

[0216] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of the embodiments of the present invention, depending on actual needs.

[0217] In addition, the functional units in the various embodiments of the present invention can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit.

[0218] If the aforementioned functions are implemented as software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this invention, essentially, or the part that contributes to the prior art, or a portion of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0219] The above description is merely a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any person skilled in the art can easily conceive of various equivalent modifications or substitutions within the technical scope disclosed in the present invention, and these modifications or substitutions should all be covered within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be determined by the scope of the claims.

Claims

1. A vehicle fault diagnosis system, characterized in that, include: Vehicle-mounted data acquisition terminal, vehicle-mounted edge computing platform, and cloud platform; among them, The vehicle-mounted data acquisition terminal is used to collect vehicle bus data and vehicle noise and / or vibration data, and send the collected data to the vehicle-mounted edge computing platform. The vehicle-mounted edge computing platform is used to synchronize the bus data and noise and / or vibration data of the vehicle in time, and to use a fault identification model to identify faults in the time-synchronized data, and when a fault is identified, to send the identified fault data to the cloud platform. The cloud platform is used to receive fault data sent by the vehicle edge computing platform, analyze the fault data using a fault analysis model, obtain vehicle fault analysis results, and send them to the vehicle edge computing platform. The fault analysis results include fault type and / or fault handling suggestions. The vehicle's bus data frame includes data of n different bus parameters polled at fixed time intervals Δt. The time synchronization of the vehicle's bus data with noise and / or vibration data includes: The collected noise and / or vibration data and their sampling time are recorded as a data record. The data records of the noise and / or vibration data are sorted according to the sampling time to obtain the first data record table of the noise and / or vibration data. The engine's first speed data is collected by a speed sensor connected to the crankshaft of the vehicle engine, and the corresponding sampling time is recorded. The collected first speed data and its sampling time are recorded as a data record, and the data records of the first speed data are sorted according to the sampling time to obtain the first data record table of the first speed data. The system receives bus data frames from the vehicle, wherein the n different bus parameters include the engine's second speed data; based on the reception time of the bus data frame and the fixed time interval, it extracts the data of each bus parameter from the same bus data frame and determines the corresponding sampling time; it treats the data of the same bus parameter in multiple bus data frames and its sampling time as a data record, and sorts the data records of the same bus parameter according to the sampling time to obtain a second data record table for each bus parameter; and it resamples and interpolates the second data record table for each bus parameter according to the sampling time interval of the first speed data to obtain a third data record table for each bus parameter. According to the sampling time, the rotational speed data in the first data record table of the first rotational speed data and the third data record table of the second rotational speed data are aligned, and the difference value between L consecutive aligned rotational speed data is calculated; and, each time, the data record in the third data record table of the second rotational speed data is shifted one position to the left, and the difference value between L consecutive aligned rotational speed data is recalculated. The time deviation between the first data record table of the first rotational speed data and the third data record table of the second rotational speed data is determined based on the smallest of the difference values ​​obtained from multiple calculations. Based on the time deviation, the sampling time in the third data record table of each bus parameter is updated, and based on the updated sampling time, the first data record table of the noise and / or vibration data is aligned with the third data record table of each bus parameter.

2. The vehicle fault diagnosis system as described in claim 1, characterized in that, The vehicle-mounted data acquisition terminal includes a vehicle bus data acquisition module and a noise and vibration acquisition module, wherein... The vehicle bus data acquisition module listens to or actively polls the vehicle bus interface to read the vehicle's bus data. The noise and vibration acquisition module collects noise and / or vibration data inside and outside the vehicle through noise sensors and / or vibration sensors installed on the vehicle.

3. The vehicle fault diagnosis system as described in claim 1, characterized in that, The in-vehicle edge computing platform includes: a data synchronization and preprocessing module, a fault identification module, a data upload module, and a result display module; wherein... The data synchronization and preprocessing module is used to synchronize the bus data and noise and / or vibration data of the vehicle in time, and to send the time-synchronized data to the fault identification module and the data upload module after performing a first filtering process. The fault identification module is used to identify faults in the data sent by the data synchronization and preprocessing module using a fault identification model, and when a fault is identified, to send the identified fault data to the data upload module. The data upload module is used to send fault data from the fault identification module to the cloud platform, and to send data from the data synchronization and preprocessing module to the cloud platform according to a preset data sending strategy. The result display module is used to display the fault information identified by the fault identification module, and / or to display the fault analysis results sent by the cloud platform.

4. The vehicle fault diagnosis system as described in claim 1, characterized in that, The cloud platform includes: a cloud gateway, a third-party system integration gateway, a filtering and feature extraction module, a fault record analysis module, a data storage module, a model training module, a model management module, and a fault analysis module. The cloud gateway is used to connect and manage multiple vehicle edge computing platforms, receive data sent by the vehicle edge computing platforms, and / or send the vehicle fault analysis results obtained by the fault analysis module to the vehicle edge computing platforms. The filtering and feature extraction module is used to perform format conversion, filtering and feature extraction on the data from the vehicle edge computing platform, and transmit the extracted fault feature data to the data storage module. The fault record analysis module is used to extract fault types from the received vehicle fault handling records and associate the extracted fault types with the vehicle's fault data records. The data storage module is used to store the fault feature data extracted by the filtering and feature extraction module, the fault type extracted by the fault record analysis module, and the data received by the cloud gateway from the vehicle edge computing platform. The model training module is used to add labels to the fault feature data of the data storage module according to the fault type extracted by the fault record analysis module, update the training data, and retrain the fault identification model and the fault analysis model using a preset training algorithm and the updated training data. The model management module is used for version management, storage and publishing of fault identification models and fault analysis models. It publishes the trained fault identification model to the vehicle edge computing platform and the trained fault analysis model to the fault analysis module. The fault analysis module is used to analyze the data stored in the data storage module using a fault analysis model to obtain the fault analysis results of the vehicle. The third-party system integration gateway is used to receive vehicle fault handling records sent by the third-party system and send them to the fault record analysis module, and / or send the vehicle fault analysis results obtained by the fault analysis module to the third-party system.

5. The vehicle fault diagnosis system as described in claim 4, characterized in that, The fault record analysis module is also used to extract fault handling suggestions corresponding to the fault type from the received vehicle fault handling records, and associate the extracted fault handling suggestions with the corresponding fault type and save them in the knowledge base of the data storage module. The fault analysis module is also used to search the knowledge base after obtaining the fault type of the vehicle using the fault analysis model, and obtain fault handling suggestions corresponding to the fault type.

6. A vehicle fault diagnosis method, characterized in that, include: The vehicle-mounted data acquisition terminal collects bus data and noise and / or vibration data of the vehicle, and sends the collected data to the vehicle-mounted edge computing platform. The vehicle edge computing platform synchronizes the vehicle's bus data with noise and / or vibration data in time, and uses a fault identification model to identify faults in the time-synchronized data. When a fault is identified, the fault data is sent to the cloud platform. The cloud platform receives fault data sent by the vehicle edge computing platform, analyzes the fault data using a fault analysis model, obtains vehicle fault analysis results, and sends them to the vehicle edge computing platform. The fault analysis results include fault type and / or fault handling suggestions. The vehicle's bus data frame includes data of n different bus parameters polled at fixed time intervals Δt. The time synchronization of the vehicle's bus data with noise and / or vibration data includes: The collected noise and / or vibration data and their sampling time are recorded as a data record. The data records of the noise and / or vibration data are sorted according to the sampling time to obtain the first data record table of the noise and / or vibration data. The engine's first speed data is collected by a speed sensor connected to the crankshaft of the vehicle engine, and the corresponding sampling time is recorded. The collected first speed data and its sampling time are recorded as a data record, and the data records of the first speed data are sorted according to the sampling time to obtain the first data record table of the first speed data. The system receives bus data frames from the vehicle, wherein the n different bus parameters include the engine's second speed data; based on the reception time of the bus data frame and the fixed time interval, it extracts the data of each bus parameter from the same bus data frame and determines the corresponding sampling time; it treats the data of the same bus parameter in multiple bus data frames and its sampling time as a data record, and sorts the data records of the same bus parameter according to the sampling time to obtain a second data record table for each bus parameter; and it resamples and interpolates the second data record table for each bus parameter according to the sampling time interval of the first speed data to obtain a third data record table for each bus parameter. According to the sampling time, the rotational speed data in the first data record table of the first rotational speed data and the third data record table of the second rotational speed data are aligned, and the difference value between L consecutive aligned rotational speed data is calculated; and, each time, the data record in the third data record table of the second rotational speed data is shifted one position to the left, and the difference value between L consecutive aligned rotational speed data is recalculated. The time deviation between the first data record table of the first rotational speed data and the third data record table of the second rotational speed data is determined based on the smallest of the difference values ​​obtained from multiple calculations. Based on the time deviation, the sampling time in the third data record table of each bus parameter is updated, and based on the updated sampling time, the first data record table of the noise and / or vibration data is aligned with the third data record table of each bus parameter.

7. The vehicle fault diagnosis method as described in claim 6, characterized in that, Also includes: The vehicle bus data is read by listening to or actively polling the vehicle bus interface. Noise and / or vibration data inside and outside the vehicle are collected using noise and / or vibration sensors installed on the vehicle.

8. The vehicle fault diagnosis method as described in claim 6, characterized in that, The vehicle-mounted edge computing platform includes: a data synchronization and preprocessing module, a fault identification module, a data upload module, and a result display module; the method further includes: The data synchronization and preprocessing module synchronizes the vehicle's bus data with noise and / or vibration data in time, and then performs a first filtering process on the time-synchronized data before sending it to the fault identification module and the data upload module. The fault identification module uses a fault identification model to identify faults in the data sent by the data synchronization and preprocessing module, and when a fault is identified, the fault data is sent to the data upload module. The data upload module sends fault data from the fault identification module to the cloud platform, and sends data from the data synchronization and preprocessing module to the cloud platform according to a preset data sending strategy. The result display module displays the fault information identified by the fault identification module, and / or displays the fault analysis results sent by the cloud platform.

9. The vehicle fault diagnosis method as described in claim 6, characterized in that, The cloud platform includes: a cloud gateway, a third-party system integration gateway, a filtering and feature extraction module, a fault record analysis module, a data storage module, a model training module, a model management module, and a fault analysis module. The method further includes: The cloud gateway connects to and manages multiple vehicle edge computing platforms, receives data sent by the vehicle edge computing platforms, and / or sends the vehicle fault analysis results obtained by the fault analysis module to the vehicle edge computing platforms. The filtering and feature extraction module converts, filters, and extracts features from the data from the vehicle edge computing platform, and transmits the extracted fault feature data to the data storage module. The fault record analysis module extracts the fault type from the received vehicle fault handling records and associates the extracted fault type with the vehicle's fault data records. The data storage module stores the fault feature data extracted by the filtering and feature extraction module, the fault type extracted by the fault record analysis module, and the data received by the cloud gateway from the vehicle edge computing platform. The model training module adds labels to the fault feature data of the data storage module based on the fault type extracted by the fault record analysis module, and updates the training data; and retrains the fault identification model and the fault analysis model using a preset training algorithm and the updated training data. The model management module manages, stores, and publishes versions of the fault identification model and the fault analysis model. The trained fault identification model is published to the vehicle edge computing platform, and the trained fault analysis model is published to the fault analysis module. The fault analysis module uses a fault analysis model to analyze the data stored in the data storage module to obtain the vehicle's fault analysis results. The third-party system integration gateway receives vehicle fault handling records sent by the third-party system and sends them to the fault record analysis module, and / or sends the vehicle fault analysis results obtained by the fault analysis module to the third-party system.

10. The vehicle fault diagnosis method as described in claim 9, characterized in that, Also includes: The fault record analysis module extracts fault handling suggestions corresponding to the fault type from the received vehicle fault handling records, and associates the extracted fault handling suggestions with the corresponding fault type and saves them in the knowledge base of the data storage module. The fault analysis module obtains the vehicle's fault type using the fault analysis model, then searches the knowledge base to obtain fault handling suggestions corresponding to the fault type.

Images