OTA upgrade failure intelligent diagnosis system and method based on TSP platform

Abstract

The application provides an OTA upgrade fault intelligent diagnosis system and method based on a TSP platform, and relates to the technical field of vehicle fault diagnosis. The method comprises the following steps: collecting OTA diagnosis data according to a preset data dimension when a vehicle is upgraded by OTA and triggers fault diagnosis; determining the fault type of the OTA upgrade fault, and constructing fault features and fault processing strategies for each fault type to obtain a fault classification library; matching the fault classification library with the OTA diagnosis data in terms of fault features, determining the fault type of the OTA diagnosis data according to the matching result if the matching is successful, and calling a machine learning model deployed in the TSP platform to identify the fault type of the OTA diagnosis data if the matching fails; and then generating a fault processing strategy. The application is used to solve the problems that the fault positioning efficiency is low during OTA diagnosis, the new energy vehicle scene is not well adapted, and the TSP platform resources are not fully utilized, so that the diagnosis accuracy is limited by the data dimension of a single vehicle.

Description

Technical Field

[0001] This invention relates to the field of vehicle fault diagnosis technology, and specifically to an intelligent fault diagnosis system and method for OTA upgrades based on the TSP platform. Background Technology

[0002] Currently, over-the-air (OTA) technology upgrades have become a core means for intelligent connected vehicles to achieve functional iteration, defect repair, and performance optimization. The upgrade process covers multiple key components, including cloud servers, communication networks, in-vehicle telematics boxes (T-BOX), multi-domain controllers (such as the electronic control unit (ECU) of the three-electric system, the autonomous driving domain controller, the central control entertainment ECU, and the chassis ECU), and in-vehicle execution modules.

[0003] Existing OTA fault diagnosis technologies for new energy vehicles mostly rely on manual analysis. They mainly collect on-board upgrade logs (such as ECU operation logs and CAN bus interaction logs), cloud traffic data, and vehicle information (model, production batch, controller configuration). Combined with engineers' experience, they locate the root cause of the problem. Some systems only realize basic anomaly detection (such as judging file corruption by upgrade package check value and identifying unauthorized operations based on a list of commonly used login devices).

[0004] However, the existing OTA fault diagnosis technologies have the following shortcomings. First, the fault location efficiency is low. OTA upgrades for new energy vehicles involve a large number of controllers (including dedicated controllers for the electric drive system). Due to the dispersion of vehicle data from various stages, such as cloud upgrade task data, on-board T-BOX logs, electric drive system status codes, and Controller Area Network (CAN) data, manual cross-platform integration and analysis of this data is required, resulting in a long location cycle and making it difficult to meet the rapid response requirements of automakers for batch upgrades. Second, there is insufficient adaptation to new energy vehicle scenarios. Existing diagnostic rules do not consider the unique scenarios of various new energy vehicle models. For example, transmission interruptions caused by weak networks in underground garages or tunnels, unstable power supply to the electric drive system triggered by upgrades while driving, unauthorized upgrades after vehicle transfer, and software conflicts caused by the collaborative work of the electric drive system and the autonomous driving domain controller are not covered by the diagnostics. Thirdly, the Telematics Service Provider (TSP) platform has not been fully utilized. As the core hub of "vehicle-cloud-human" collaboration, the TSP platform has the capabilities of vehicle data collection, remote communication and basic diagnostics. However, the existing technology only uses the TSP platform for data transmission during fault diagnosis, without utilizing its cross-vehicle data aggregation capabilities to integrate various vehicle fault data (such as the working status of the three-electric system, power battery voltage fluctuations, and motor controller temperature) for in-depth fault diagnosis. This results in the diagnostic accuracy being limited to the single-vehicle data dimension. Summary of the Invention

[0005] In view of this, it is necessary to provide an intelligent diagnostic system and method for OTA upgrade faults based on the TSP platform, in order to solve the problems of low fault location efficiency, insufficient adaptation to new energy vehicle scenarios, and failure to fully utilize TSP platform resources, which leads to diagnostic accuracy being limited to a single vehicle data dimension.

[0006] To address the aforementioned problems, this invention provides an intelligent fault diagnosis system for OTA upgrades based on the TSP platform, comprising: The vehicle-mounted multi-source data synchronous acquisition module is used to collect OTA diagnostic data according to preset data dimensions when the vehicle is upgraded over-the-air (OTA) and fault diagnosis is triggered. The data dimensions include: cloud download dimension, vehicle dimension, historical fault dimension, vehicle operating status dimension, vehicle operation dimension, and vehicle model dimension. The fault classification library construction module is used to determine the fault type of OTA upgrade failure, and to construct fault features and fault handling strategies for each fault type to obtain the fault classification library. The fault type includes a first fault type that corresponds one-to-one with the data dimension and a second fault type that is divided according to the first fault type under the vehicle model. The fault intelligent diagnosis module is used to match the fault classification library with the OTA diagnostic data for fault features. If the feature matching is successful, the fault type of the OTA diagnostic data is determined based on the fault feature matching result. If the feature matching fails, the machine learning model deployed in the TSP platform is called to identify the fault in the OTA diagnostic data and obtain the fault type of the OTA diagnostic data. The fault diagnosis result output module is used to generate corresponding fault handling strategies based on the fault type of the OTA diagnostic data.

[0007] In one possible implementation, the vehicle-mounted multi-source data synchronous acquisition module is also used to trigger fault diagnosis when the vehicle OTA upgrade meets any of the following fault conditions; The status code for a single vehicle's OTA upgrade indicates that the OTA upgrade has failed. The TSP platform detected an error code from the vehicle's embedded data points during an OTA upgrade of a single vehicle. When performing OTA upgrades on the same vehicle model in batches, the percentage of vehicles with failed OTA upgrades or error codes detected by the TSP platform at the vehicle mounting points reaches a preset threshold.

[0008] In one possible implementation, the vehicle-mounted multi-source data synchronous acquisition module is further configured to establish a communication link from the vehicle to the cloud after triggering fault diagnosis, and the communication link is used to transmit the acquired OTA raw data. The original OTA data is timestamped, and data features associated with OTA upgrade failures are extracted from the original OTA data. Each data feature is then defined as a data dimension. The raw OTA data is divided according to the data dimensions to obtain OTA diagnostic data.

[0009] In one possible implementation, the fault intelligent diagnosis module is further configured to perform fault feature matching in the following manner; Based on the data dimensions of the OTA diagnostic data, preset rules are retrieved from the fault classification library; The corresponding fault characteristics are determined from the preset rules retrieved; The fault type corresponding to the fault feature is determined as the fault type of the OTA diagnostic data; If no corresponding preset rule is found in the fault classification library based on the data dimensions of the OTA diagnostic data, then feature matching is determined to have failed.

[0010] In one possible implementation, the fault intelligent diagnosis module is further configured to train a machine learning model in the following manner: Fault data samples of various vehicle models during OTA upgrades were obtained from the historical database of the TSP platform. The faulty data sample is cleaned to obtain a clean sample; Extract derived features of the corresponding vehicle models from the clean samples, and calculate the correlation coefficient between the derived features and the fault type of the OTA upgrade failure; Remove derived features with correlation coefficients less than a threshold, and train a machine learning model based on the remaining derived features.

[0011] In one possible implementation, the fault intelligent diagnosis module is further configured to obtain feedback results of the fault handling strategy in handling OTA upgrade faults; The feedback results are stored as new fault data in the historical database of the TSP platform to obtain the updated historical database. When the incremental value of newly added fault data in the updated historical database is greater than a preset incremental threshold, the machine learning model is updated based on the updated historical database.

[0012] In one possible implementation, the fault diagnosis result output module is further configured to construct a diagnostic report text based on the matching results of the fault features, the fault type of the OTA diagnostic data, and the fault handling strategy. The diagnostic report text is pushed through multiple output carriers, including engineer's terminal, car owner's terminal, and car manufacturer's server terminal.

[0013] This invention also provides an intelligent fault diagnosis method for OTA upgrades based on a TSP platform, comprising: When a vehicle undergoes an OTA upgrade and triggers fault diagnosis, OTA diagnostic data is collected according to preset data dimensions, including: cloud download dimension, vehicle dimension, historical fault dimension, vehicle operating status dimension, vehicle operation dimension, and vehicle model dimension. The fault types of OTA upgrade failures are determined, and fault characteristics and fault handling strategies are constructed for each fault type to obtain a fault classification library. The fault types include a first fault type that corresponds one-to-one with the data dimensions and a second fault type that is divided under the vehicle model. The fault classification library is matched with the OTA diagnostic data for fault features. If the feature matching is successful, the fault type of the OTA diagnostic data is determined based on the fault feature matching result. If the feature matching fails, the machine learning model deployed in the TSP platform is called to identify faults in the OTA diagnostic data to obtain the fault type of the OTA diagnostic data. A corresponding fault handling strategy is generated based on the fault type of the OTA diagnostic data.

[0014] The present invention also provides an electronic device, including a memory and a processor, wherein the memory is used to store a program; the processor is coupled to the memory and is used to execute the program stored in the memory to implement the steps of the above-described intelligent diagnosis method for OTA upgrade faults based on the TSP platform.

[0015] The present invention also provides a non-transitory computer-readable storage medium having a computer program stored thereon, wherein the computer program, when executed by a processor, implements the steps of the above-described intelligent diagnosis method for OTA upgrade faults based on the TSP platform.

[0016] The beneficial effects of adopting the above implementation method are as follows: The OTA upgrade fault intelligent diagnosis system and method based on the TSP platform provided by this invention first designs a system processing flow from multi-dimensional diagnostic data collection, to fault classification library construction, to dual diagnosis of feature matching and model fault identification, and finally outputs diagnostic results. This processing flow does not require cross-platform data integration and fault analysis, which can significantly shorten the OTA fault location time for various types of new energy vehicles and meet the rapid response requirements of car manufacturers' batch operations. When constructing the fault classification library, a second fault type is further subdivided according to the vehicle model under the first fault type corresponding to the data dimension, which can cover the unique and complex fault scenarios of various types of new energy vehicles, enhance the OTA fault identification effect of new energy vehicle models, and improve the adaptability of new energy vehicle scenarios. In addition, deploying the machine learning model on the TSP platform can effectively utilize the cross-vehicle data aggregation capability of the TSP platform, accelerate the model calculation speed, and perform unified diagnostic prediction by integrating OTA diagnostic data from multiple data dimensions, thereby improving diagnostic accuracy and supporting the large-scale parallel diagnostic needs of new energy vehicles. Attached Figure Description

[0017] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying 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.

[0018] Figure 1 A schematic diagram of the framework of the OTA upgrade fault intelligent diagnosis system based on the TSP platform provided by the present invention; Figure 2 A flowchart illustrating the intelligent fault diagnosis method for OTA upgrades based on the TSP platform provided by this invention. Figure 3 A schematic diagram of an embodiment of the electronic device provided by the present invention. Detailed Implementation

[0019] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of the present invention, and not all of them. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without creative effort are within the scope of protection of the present invention.

[0020] In the description of the embodiments of this application, unless otherwise stated, "a plurality of" means two or more.

[0021] In this embodiment of the invention, the terms "comprising" and "having" and any variations thereof are intended to cover non-exclusive inclusion, for example, a process, method, apparatus, product or device that includes a series of steps or modules is not necessarily limited to those steps or modules that are explicitly listed, but may include other steps or modules that are not explicitly listed or that are inherent to such process, method, product or device.

[0022] The naming or numbering of steps in the embodiments of the present invention does not mean that the steps in the method flow must be executed in the time / logical order indicated by the naming or numbering. The execution order of the named or numbered process steps can be changed according to the technical purpose to be achieved, as long as the same or similar technical effect can be achieved.

[0023] In this document, the term "embodiment" means that a particular feature, structure, or characteristic described in connection with an embodiment may be included in at least one embodiment of the invention. The appearance of this phrase in various places throughout the specification does not necessarily refer to the same embodiment, nor is it a mutually exclusive, independent, or alternative embodiment. It will be explicitly and implicitly understood by those skilled in the art that the embodiments described herein can be combined with other embodiments.

[0024] The OTA upgrade fault intelligent diagnosis system based on the TSP platform provided by this invention can be applied to OTA upgrade scenarios for single vehicles or batches of vehicles. The executing entity can be the vehicle's in-vehicle system, the cloud platform on which the vehicle relies for OTA upgrades, or various terminals or servers communicating with the in-vehicle system. When a vehicle OTA upgrade is initiated, the OTA upgrade fault intelligent diagnosis system based on the TSP platform of this invention is invoked to trigger fault diagnosis during the vehicle OTA upgrade and ultimately output the corresponding fault handling strategy, serving as the theoretical basis for subsequent automatic or manual handling of vehicle OTA upgrade faults.

[0025] The following section details the intelligent OTA upgrade fault diagnosis system based on the TSP platform provided by this invention. Unless otherwise stated, the term "system" as used below refers to the intelligent OTA upgrade fault diagnosis system based on the TSP platform.

[0026] Figure 1 This is a schematic diagram of the framework of the OTA upgrade fault intelligent diagnosis system based on the TSP platform provided by the present invention, as shown below. Figure 1 As shown, the OTA upgrade fault intelligent diagnosis system based on the TSP platform specifically includes: an on-board multi-source data synchronous acquisition module, a fault classification library construction module, a fault intelligent diagnosis module, and a fault diagnosis result output module, which will be described one by one below.

[0027] The vehicle-mounted multi-source data synchronous acquisition module is used to collect OTA diagnostic data according to preset data dimensions when the vehicle is upgraded via OTA and fault diagnosis is triggered.

[0028] When a vehicle undergoes an OTA upgrade and triggers fault diagnosis, the onboard multi-source data synchronous acquisition module can clearly define the scope of OTA diagnostic data collection based on preset data dimensions. These data dimensions are determined according to the data sources of each key stage involved in the vehicle OTA upgrade, and mainly include: cloud download dimension, onboard dimension, historical fault dimension, vehicle operating status dimension, vehicle operation dimension, and vehicle model dimension.

[0029] Among them, the OTA diagnostic data in the cloud download dimension mainly includes: OTA upgrade package size, version, upgrade task distribution strategy and time, vehicle binding information (such as vehicle identification number, vehicle component serial number, etc.).

[0030] The OTA diagnostic data for vehicle-mounted systems mainly includes: ECU of the three-electric system, domain controller model or status, network speed, power battery charge, CAN bus interaction log, T-BOX communication status, battery management system (BMS) parameters, motor controller temperature, etc.

[0031] Historical fault-related OTA diagnostic data mainly includes: historical upgrade success rate for the same vehicle model, past fault solutions, and a list of ECU compatibility for the three-electric system.

[0032] OTA diagnostic data for vehicle operating status mainly includes: vehicle ignition status, gear information, driving speed, remaining range, and charging status.

[0033] OTA diagnostic data for vehicle operation mainly includes: the driver's list of in-vehicle application-bound devices, historical upgrade time preferences, and vehicle authorization records.

[0034] The OTA diagnostic data for vehicle models mainly includes: (new energy vehicles) vehicle model, production batch, and controller configuration list (including the adaptation status of 800V high-voltage platform, three-electric system controller, and high-level autonomous driving domain controller).

[0035] During the actual data collection process, the data is collected in parallel according to the above six data dimensions, and then the OTA diagnostic data collected from each data dimension is uploaded to the vehicle multi-source data synchronous collection module.

[0036] In one possible implementation, the on-board multi-source data synchronous acquisition module is also used to trigger fault diagnosis when the vehicle OTA upgrade meets any of the following fault conditions; Fault Condition 1: The status code for OTA upgrade of a single vehicle indicates that the OTA upgrade of the vehicle has failed.

[0037] Fault Condition 2: When performing an OTA upgrade on a single vehicle, the TSP platform detects an error code from the vehicle's embedded points.

[0038] Fault Condition 3: When performing OTA upgrades on the same vehicle model in batches, the percentage of vehicles with failed OTA upgrades or error codes detected by the TSP platform's onboard data points reaches a preset threshold.

[0039] Since vehicle OTA upgrades may be performed on a single vehicle or in batches on multiple vehicles of the same model, the conditions for triggering fault diagnosis need to be set separately. Fault conditions 1 and 2 mentioned above are for a single vehicle. Specifically, the system monitors the vehicle OTA upgrade process and obtains the status code for each vehicle's OTA upgrade. If this status code indicates that the vehicle OTA upgrade has failed, an OTA upgrade fault can be determined. In addition, the system also identifies error codes detected by the TSP platform at the vehicle's data entry points in real time. When an error code is detected by the TSP platform at the vehicle's data entry points, an OTA upgrade fault can also be determined, requiring the triggering of fault diagnosis.

[0040] The aforementioned fault condition 3 applies to the batch upgrade process of vehicles of the same model. The system will simultaneously monitor all vehicles of the same model undergoing batch OTA upgrades, and count the number of vehicles experiencing OTA upgrade failures or error codes detected by the TSP platform's onboard data points, and calculate the percentage of such vehicles. If this percentage exceeds a preset threshold, such as exceeding 5%, it indicates that some vehicles have experienced OTA upgrade failures during the current batch OTA upgrade, and fault diagnosis needs to be triggered.

[0041] In this embodiment of the invention, fixed fault conditions are set as the start switch for fault diagnosis, thereby automatically triggering fault diagnosis during vehicle OTA upgrades and quickly triggering fault diagnosis during batch OTA upgrades, without having to check each vehicle with OTA upgrade failure individually, thus ensuring the timeliness of fault monitoring.

[0042] In one possible implementation, the vehicle-mounted multi-source data synchronous acquisition module is also used to establish a communication link from the vehicle to the cloud after triggering fault diagnosis.

[0043] Here, after triggering fault diagnosis, the onboard multi-source data synchronous acquisition module begins collecting OTA diagnostic data. First, the MQTT protocol is used to establish a communication link between the cloud and the vehicle, i.e., to build a real-time transmission channel. This real-time transmission channel enables bidirectional communication between the vehicle and the cloud for transmitting the collected raw OTA data. During link configuration, a strategy of "onboard T-BOX edge computing combined with cloud collaboration" is adopted. This means that high-frequency data (interaction logs of the three-electric system and CAN bus logs) is compressed and processed locally on the onboard T-BOX, while local caching and breakpoint resumption are activated in weak network environments.

[0044] Furthermore, the original OTA data is timestamped, and data features associated with OTA upgrade failures are extracted from the original OTA data. Each data feature is then defined as a data dimension.

[0045] Here, the OTA raw data collected via the communication link can be time-aligned according to the collection timestamp of each data type to ensure the time sequence synchronization of data collected across different data dimensions. Simultaneously, data features with strong correlations to OTA upgrade faults are extracted from the OTA raw data. These data features can be data type, data source (e.g., cloud download, vehicle-mounted, historical faults, vehicle operation, vehicle control, and vehicle model). Then, each data feature is defined as a data dimension: cloud download dimension, vehicle-mounted dimension, historical fault dimension, vehicle operating status dimension, vehicle control dimension, and vehicle model dimension.

[0046] Finally, the raw OTA data is divided according to data dimensions to obtain OTA diagnostic data.

[0047] The raw OTA data is divided according to the data dimensions determined by data characteristics, thereby obtaining the OTA diagnostic data required for system fault diagnosis. This transforms various unstructured raw OTA data into structured OTA diagnostic data, serving as the basis for subsequent fault diagnosis.

[0048] In this embodiment of the invention, when fault diagnosis is triggered, the transmission and preprocessing of OTA diagnostic data are achieved by establishing a real-time transmission channel, aligning timestamps, and dividing data dimensions. This ensures the real-time nature and integrity of the OTA diagnostic data and improves the inference accuracy and response speed of subsequent fault diagnosis.

[0049] See also Figure 1 The fault classification library construction module is used to determine the fault type of OTA upgrade failure, and to construct fault characteristics and fault handling strategies for each fault type to obtain the fault classification library.

[0050] Here, a fault classification library is constructed in the fault classification library construction module. The fault classification library stores multiple preset rules. The preset rules are stored according to the data-fault-countermeasure mapping relationship template. Each preset rule records a case of handling a vehicle OTA upgrade fault.

[0051] In the specific construction process, the first step is to determine the fault types of OTA upgrade failures, which must cover all possible types of OTA upgrade failures, thereby defining a standardized fault classification system. The fault types include a first fault type that corresponds one-to-one with the data dimensions, and a second fault type that is further divided into two categories based on the first fault type under the vehicle model.

[0052] Based on the above six data dimensions, this embodiment of the invention first defines a primary classification in the standardized fault classification system, dividing the fault types into six categories and determining them as the first fault types that correspond one-to-one with the data dimensions. Specifically, these include six fault types: upgrade package abnormality, vehicle hardware abnormality, software conflict, environmental and driving status interference, operational violations, and network failure.

[0053] Furthermore, this embodiment of the invention also implements a secondary classification under the first fault type, further dividing the first fault type into a second fault type based on the vehicle model. This secondary subdivision mainly targets various models of new energy vehicles. For example, for a certain model of new energy vehicle, its software conflict faults can be further subdivided into "incompatibility between ECUs in the three-electric system" and "conflict between the 800V high-voltage platform and the cross-domain controller," etc.

[0054] After determining the fault types of OTA upgrade failures, fault characteristics and fault handling strategies are constructed for each fault type to obtain a fault classification library.

[0055] For each fault type, the fault data can be determined first, and then a unique feature label can be constructed to determine the fault characteristics of the fault data. For example, if the fault data for a certain fault type is "motor controller overheating", this situation may be caused by the motor controller temperature exceeding the safety threshold when the OTA upgrade is interrupted. Therefore, its fault characteristic can be determined as "motor controller temperature exceeds the safety threshold when the upgrade is interrupted".

[0056] Furthermore, fault handling strategies are constructed for fault data under corresponding fault types. These fault handling strategies can be collected from historical OTA upgrade fault handling cases. Based on the fault of the corresponding fault type, the corresponding fault handling strategy can be queried from the historical fault case library. For example, based on the fault data "BMS conflicts with autonomous driving domain controller", the fault handling strategy can be determined by querying to be "first upgrade the BMS to a compatible version, then trigger the TSP platform verification, and finally distribute the autonomous driving domain upgrade package in a targeted manner".

[0057] Therefore, by using the above method, fault characteristics and fault handling strategies are constructed one by one under the fault data corresponding to the fault type, and a mapping relationship is established between the three to form a preset rule. Through fault data of various fault types, multiple preset rules can be established, ultimately forming a fault classification library.

[0058] This invention constructs a fault classification library as a diagnostic database to provide a basis for subsequent fault diagnosis. The classification according to fault type ensures accurate matching of fault type and fault characteristics during subsequent fault diagnosis, improving the efficiency of fault diagnosis. Furthermore, in data type classification, a second fault type is further subdivided according to vehicle model under the first fault type corresponding to the data dimension. This covers various unique and complex faults of new energy vehicles of different models, enhancing the recognition effect and improving adaptability in new energy vehicle scenarios.

[0059] See also Figure 1 The fault intelligent diagnosis module is used to match fault features between the fault classification library and OTA diagnostic data. If the feature matching is successful, the fault type of the OTA diagnostic data is determined based on the matching result. If the feature matching fails, the machine learning model deployed in the TSP platform is called to identify faults in the OTA diagnostic data and obtain the fault type of the OTA diagnostic data.

[0060] Here, the system implements a dual diagnostic mechanism utilizing a fault classification library and a machine learning model through an intelligent fault diagnosis module. The first stage of diagnosis is achieved using the fault classification library. By matching fault characteristics between the library and OTA diagnostic data, and identifying the fault type in the OTA diagnostic data, fault diagnosis for vehicle OTA upgrades is realized.

[0061] In one possible implementation, the fault intelligent diagnosis module is also used to perform fault feature matching in the following manner, which is explained in detail below.

[0062] First, based on the data dimensions of the OTA diagnostic data, preset rules in the fault classification library are retrieved. Then, the corresponding fault characteristics are determined from the retrieved preset rules. Finally, the fault type corresponding to the fault characteristics is determined as the fault type of the OTA diagnostic data.

[0063] Here, during the specific retrieval, the fault classification library is searched through the OTA diagnostic data data dimension. This allows for querying the corresponding preset rules for the data dimension, thereby determining whether the fault features in the preset rules match the OTA diagnostic data. The match is determined by calculating the semantic similarity between the fault features and the OTA diagnostic data. If the calculated semantic similarity is greater than a threshold (e.g., 95%), the fault features are considered successfully matched. Furthermore, based on the successfully matched preset rules, the fault type in the OTA diagnostic data can be determined, thus achieving fault diagnosis for vehicle OTA upgrades.

[0064] For example, if the OTA diagnostic data retrieves a fault characteristic of "login device is not a commonly used device, not a vehicle authorized user ID" according to a preset rule, then the fault type of the OTA diagnostic data is directly determined to be "operation violation" based on this preset rule. Similarly, if the data retrieves "driving status, power battery voltage fluctuation exceeds reasonable range", the fault type is directly determined to be "environmental and driving status interference". If the data retrieves "motor controller temperature exceeds safety threshold, upgrade interrupted", the fault type can be directly determined to be "onboard hardware abnormality, motor controller overheating".

[0065] If no corresponding preset rule is found in the fault classification library based on the data dimensions of the OTA diagnostic data, then feature matching is determined to have failed.

[0066] If the fault features in the preset rules do not match the OTA diagnostic data, that is, the calculated semantic similarity is less than or equal to the threshold (95%), it means that the current fault classification library does not record fault features or fault types similar to the OTA diagnostic data, and the fault feature matching is determined to be unsuccessful.

[0067] In this embodiment of the invention, rule matching between the fault classification library and OTA diagnostic data is achieved through fault features. This enables the rapid identification of most simple faults in complex OTA upgrade failure scenarios without the need for complex screening and calculation processes, thereby improving the efficiency of fault diagnosis.

[0068] When feature matching fails, it indicates that the fault classification library does not cover the complex faults represented by the current OTA diagnostic data. Therefore, the second diagnostic mechanism is initiated, which calls the machine learning model deployed in the TSP platform to identify faults in the OTA diagnostic data and obtain the fault type of the OTA diagnostic data.

[0069] Here, considering the TSP platform's ability to aggregate multi-source data, this embodiment of the invention deploys the fault prediction model within the TSP platform, specifically on its edge nodes. This accelerates the calculation speed of the fault prediction model and improves prediction efficiency. The machine learning model identifies faults in the OTA diagnostic data and outputs probability distributions for various fault types. The fault type with the highest probability value in the probability distribution is then selected as the fault type in the OTA diagnostic data.

[0070] In one possible implementation, the fault intelligent diagnosis module is also used to train a machine learning model in the following manner, as detailed below.

[0071] First, fault data samples of various vehicle models during OTA upgrades are obtained from the historical database of the TSP platform. These samples are then cleaned to obtain clean samples. Next, derived features for the corresponding vehicle models are extracted from the clean samples, and the correlation coefficient between these derived features and the first fault type is calculated. Finally, derived features with correlation coefficients less than a threshold are removed, and a machine learning model is trained based on the remaining derived features.

[0072] Here, fault data samples from various vehicle models during OTA upgrades are obtained from the historical database of the TSP platform. The TSP platform's historical database records logs from various vehicle models during OTA upgrades, including OTA upgrade fault handling logs for various new energy vehicle models not covered by the fault classification library. These logs can serve as training samples for machine learning models, enabling the machine learning model to diagnose and predict OTA upgrade faults for these new energy vehicle models when the fault classification library cannot diagnose them.

[0073] The acquired fault data samples are filtered according to three criteria: vehicle model coverage, controller type coverage, and scenario completeness (pure electric or plug-in hybrid, intelligent connected, driving or parking, charging). Invalid samples are removed, such as samples from abnormal OTA upgrade failures caused by user-initiated upgrade termination, vehicle collisions leading to upgrade interruptions, and abnormal data collection. Valid samples are retained. Further, missing value and outlier handling are performed. For samples with missing data reported by users or missing data from some sensors, a filling algorithm based on the similarity of features within the same vehicle model, upgrade task, and fault scenario is used to fill in the missing fields. For continuous features, the interquartile range (IMR) is used to identify extreme values, removing abnormal samples that exceed a reasonable range and retaining clean samples.

[0074] Since the OTA upgrade faults recorded in these clean samples are not covered in the fault classification library, the fault features in the preset rules recorded in the fault classification library are no longer applicable. Therefore, this embodiment of the invention extracts derived features of the vehicle model from the clean samples. The vehicle model here is generally a new energy vehicle model, and the derived features are exclusive to new energy vehicles, such as: charging status, voltage fluctuation coupling features, 800V platform operating status identifier, and ECU interaction delay difference of the three-electric system, etc.

[0075] Furthermore, the correlation coefficients between derived features and the first fault type are calculated, along with the correlation coefficients between the OTA upgrade fault type and the fault type. For example, semantic similarity can be calculated; the higher the similarity, the higher the correlation coefficient. In specific calculations, the correlation coefficient between each derived feature and the first fault type corresponding to each data dimension can be calculated one by one. Then, the correlation coefficient between each derived feature and the second fault type classified according to the first fault type under the vehicle model can be further calculated. Finally, the highest correlation coefficient for each derived feature is selected.

[0076] Finally, derived features with correlation coefficients less than the coefficient threshold are removed, and the machine learning model is trained based on the remaining derived features.

[0077] Here, a coefficient threshold is used to filter derived features. If the correlation coefficient is less than the threshold, it indicates that the derived feature is not strongly associated with OTA upgrade failures, and it is removed. Conversely, if the correlation coefficient is greater than the threshold, it indicates that the derived feature is strongly associated with OTA upgrade failures, and it is retained. Finally, the derived features that are retained and not removed are used as data samples to train the machine learning model.

[0078] Machine learning models can employ decision trees and random forests, or deep neural network models with classification capabilities can be designed. During training, data samples are divided into training, validation, and test sets according to industry-standard proportions for machine learning. This ensures that the training set covers fault samples across all scenarios, and that the vehicle models and fault types distributed in the validation and test sets are consistent with those in the training set, thus avoiding data bias.

[0079] During training, the optimal hyperparameter combination, validated through experiments, was used to train the model on the training set. An early stopping mechanism was also implemented (training automatically stopped when the validation set accuracy showed no improvement after multiple training rounds) to prevent overfitting. After training, the model performance was validated using a test set to ensure it met diagnostic accuracy requirements. By comparing the performance (accuracy, recall, and inference speed) of mainstream classification models, the model that performed best in identifying complex faults (such as conflicts between an 800V high-voltage platform and a cross-domain controller) was selected as the core diagnostic model.

[0080] The trained core diagnostic model can extract corresponding derived features from OTA diagnostic data for classification, and finally output the probability distribution of various fault types, thereby realizing the diagnosis and identification of OTA upgrade faults.

[0081] When performing OTA upgrades on multiple vehicles of the same model in batches, common faults are first screened through a fault classification library, and then machine learning models are used to locate the individual faults of the corresponding models. For example, a voltage fluctuation may occur when a car owner upgrades while charging.

[0082] In this embodiment of the invention, a machine learning model is trained using derived features not covered by the fault classification library as a second-layer fault diagnosis mechanism. This can quickly and accurately locate complex faults that may exist in various types of new energy vehicles, fill the gap in the existing technology's insufficient coverage of fault diagnosis scenarios for OTA upgrades of new energy vehicles, support large-scale parallel diagnosis of new energy vehicles, significantly shorten the adaptation cycle for new energy vehicle models, and support automakers to quickly promote the iteration of OTA functions for new models.

[0083] See also Figure 1 The fault diagnosis result output module is used to generate corresponding fault handling strategies based on the fault type of the OTA diagnostic data.

[0084] Here, the fault diagnosis result output module is responsible for outputting diagnostic results. After determining the fault type in the OTA diagnostic data, it can generate a corresponding fault handling strategy according to the template of preset rules in the fault classification library. If the fault type is determined by fault feature matching based on the fault classification library, the preset rules are retrieved from the fault classification library according to the fault type, and the corresponding fault handling strategy is obtained from the preset rules. If the fault is predicted by a machine learning model, the corresponding fault cases are queried from the historical database stored in the TSP platform to determine the corresponding fault handling strategy. The fault handling strategy specifically includes root cause analysis of OTA upgrade faults and targeted solutions, which can be generated by referring to the fault diagnosis cases recorded in the preset rules of the fault classification library. Thus, through the above modules of the system, an integrated diagnostic processing flow from diagnostic data collection, fault classification library construction, dual intelligent diagnosis, and diagnostic result output is realized.

[0085] See also Figure 1 In one possible implementation, the fault diagnosis result output module is also used to construct a diagnostic report text based on the matching results of fault characteristics, the fault type of OTA diagnostic data, and the fault handling strategy, and push the diagnostic report text through the vehicle system. The output carrier includes the engineer's end, the car owner's end, and the car manufacturer's server end.

[0086] Here, when performing fault diagnosis using a fault classification library and machine learning model, the matching results of fault features, the fault types in OTA diagnostic data, and the fault handling strategies output in subsequent diagnostic results are collected. These data are then integrated into a single text file to obtain the diagnostic report text.

[0087] The matching results of fault features, i.e., the fault features associated with the fault, such as ECU model, power battery voltage, and charging status. If the fault type in the OTA diagnostic data is confirmed through machine learning model diagnosis, it needs to be labeled with attributes specific to new energy vehicles, such as incompatibility with the 800V high-voltage platform BMS version. The fault handling strategy, root cause analysis of the OTA upgrade fault diagnosed in this case, and targeted solutions are also included.

[0088] Finally, the diagnostic report text is pushed out through various output platforms.

[0089] Here, a multi-terminal output platform is designed to realize fault diagnosis push notifications, including the engineer's end, the car owner's end, and the car manufacturer's server end. Specifically, the push to the engineer's end is displayed on the web interface of the automotive R&D personnel, where a diagnostic chain is configured to display information, including screenshots of CAN bus fault frames, ECU interaction timing diagrams, and power battery voltage fluctuation curves. The push to the car owner's end is displayed on the control screen interface of the in-vehicle system application, providing the car owner with fault prompts and operation guidance. The push to the car manufacturer's server utilizes the automatic work order generation function to generate corresponding fault work orders and send them to the car manufacturer's server for unified processing.

[0090] In batch OTA upgrade scenarios, a car manufacturer-specific report generation module is generated during the push process, including the distribution of faulty vehicle models, the scope of impact, and the solution push plan, which is then uniformly fed back to the car manufacturer's server.

[0091] In this embodiment of the invention, the fault diagnosis result output module is designed to present the diagnostic results and push them to multiple output terminals, thereby providing a basis for verifying the effectiveness of OTA upgrade fault solutions. Car owners can directly and independently resolve most common faults without contacting the car manufacturer, significantly improving the success rate of OTA upgrades and the convenience of fault resolution, thus enhancing the car owner experience.

[0092] To ensure that the diagnostic predictions of the machine learning model can cover as many OTA upgrade faults of new energy vehicle models as possible, this invention proposes a feedback mechanism to achieve a self-optimizing closed loop for the machine learning model, which is described in detail below.

[0093] See also Figure 1 In one possible implementation, the fault intelligent diagnosis module is also used to obtain feedback results of the fault handling strategy for handling OTA upgrade faults.

[0094] The fault handling strategy is communicated to various output devices via push notifications, thus being implemented in real-world scenarios. At this point, the effectiveness of the solutions implemented in the fault handling strategy, i.e., the resolution status of OTA upgrade faults, is obtained. Whether the fault is successfully resolved or not, a validity label is generated.

[0095] The feedback results are stored as new fault data in the historical database of the TSP platform, resulting in an updated historical database.

[0096] Here, a complete data chain is constructed based on the diagnostic type, fault characteristics (derived characteristics), fault handling strategy, and validity label of the OTA upgrade fault. This complete data chain is then stored as new fault data in the historical database of the TSP platform, enabling iterative updates to the historical database and thus facilitating the backflow of new fault data.

[0097] When the incremental value of newly added fault data in the updated historical database exceeds a preset incremental threshold, the machine learning model is updated based on the updated historical database.

[0098] This paper proposes an incremental training mechanism for the machine learning model. By monitoring the incremental value of newly added fault data in the historical database of the TSP platform in real time, when this incremental value exceeds a preset incremental threshold (e.g., the number of newly added fault data exceeds five), the machine learning model is automatically updated. The machine learning model is then retrained using the updated historical database, achieving continuous optimization of the model and ensuring a continuous improvement in its diagnostic accuracy and ability to adapt to new faults.

[0099] This invention employs an incremental training mechanism in the intelligent fault diagnosis module to update the model and synchronously update feature labels. This eliminates the need to reconstruct the core engine, significantly shortening the adaptation cycle for new vehicle models and enhancing the system's scalability. Furthermore, these incremental update and model iteration mechanisms create a closed-loop knowledge accumulation system, enabling the knowledge base of faults and countermeasures specific to new energy vehicles to continuously iterate throughout the diagnostic process. This significantly improves the accuracy of solution recommendations when similar faults recur, reducing repetitive work.

[0100] In summary, this invention presents a system processing flow that integrates multi-dimensional diagnostic data acquisition, fault classification library construction, feature matching, and model fault identification for dual diagnosis, ultimately outputting diagnostic results. This process eliminates the need for cross-platform fault analysis and significantly reduces OTA fault location time for various new energy vehicle models. Furthermore, the fault classification library is further subdivided into second fault types based on vehicle model under the first fault type corresponding to the data dimensions. This covers a wide range of unique and complex faults specific to various new energy vehicle models, enhancing OTA fault identification for new energy vehicle models and improving adaptability in new energy vehicle scenarios. Moreover, deploying the machine learning model on the TSP platform effectively utilizes the platform's cross-vehicle data aggregation capabilities, accelerating model computation and enabling unified diagnostic prediction based on multi-dimensional OTA diagnostic data. This improves diagnostic accuracy, supports large-scale parallel diagnostic needs for new energy vehicles, and meets the rapid response requirements of automakers' batch operations.

[0101] The following section details the intelligent OTA upgrade fault diagnosis method based on the TSP platform provided by this invention. This method can be applied to the intelligent OTA upgrade fault diagnosis system based on the TSP platform proposed in the above embodiments.

[0102] Figure 2 This is a flowchart illustrating the intelligent fault diagnosis method for OTA upgrades based on the TSP platform provided by the present invention, as shown below. Figure 2 As shown, the intelligent diagnosis method for OTA upgrade faults based on the TSP platform can be implemented through the following steps 201 to 204, which are explained in detail below.

[0103] Step 201: When the vehicle undergoes an OTA upgrade and triggers fault diagnosis, collect OTA diagnostic data according to the preset data dimensions.

[0104] The data dimensions include: cloud download dimension, vehicle dimension, historical fault dimension, vehicle operating status dimension, vehicle operation dimension, and vehicle model dimension.

[0105] Step 202: Determine the fault type of OTA upgrade failure, and construct fault characteristics and fault handling strategies for each fault type to obtain a fault classification library.

[0106] The fault types include a first fault type that corresponds one-to-one with the data dimensions and a second fault type that is classified under the vehicle model.

[0107] Step 203: Match the fault classification library with the OTA diagnostic data for fault features. If the feature matching is successful, determine the fault type of the OTA diagnostic data based on the matching result. If the feature matching fails, call the machine learning model deployed in the TSP platform to identify faults in the OTA diagnostic data and obtain the fault type of the OTA diagnostic data.

[0108] Step 204: Generate the corresponding fault handling strategy based on the fault type in the OTA diagnostic data.

[0109] The OTA upgrade fault intelligent diagnosis method based on the TSP platform provided in the above embodiments can realize the technical solutions described in the embodiments of each module of the OTA upgrade fault intelligent diagnosis system based on the TSP platform. Specifically, step 201 can realize the technical solution described in the embodiment of the vehicle multi-source data synchronous acquisition module; step 202 can realize the technical solution described in the embodiment of the fault classification library construction module; step 203 can realize the technical solution described in the embodiment of the fault intelligent diagnosis module; and step 204 can realize the technical solution described in the embodiment of the fault diagnosis result output module. The specific implementation principles of each step can be found in the corresponding content in the embodiments of each module of the OTA upgrade fault intelligent diagnosis system based on the TSP platform, and their technical effects can also be referred to accordingly, which will not be repeated here.

[0110] like Figure 3 As shown, the present invention also provides an electronic device 300. The electronic device 300 includes a processor 301, a memory 302, and a display 303. Figure 3 Only some components of the electronic device 300 are shown, but it should be understood that it is not required to implement all of the components shown, and more or fewer components may be implemented instead.

[0111] In some embodiments, memory 302 may be an internal storage unit of electronic device 300, such as a hard disk or memory of electronic device 300. In other embodiments, memory 302 may also be an external storage device of electronic device 300, such as a plug-in hard disk, smart media card (SMC), secure digital (SD) card, flash card, etc. equipped on electronic device 300.

[0112] Furthermore, the memory 302 may include both internal storage units of the electronic device 300 and external storage devices. The memory 302 is used to store application software and various types of data installed on the electronic device 300.

[0113] In some embodiments, processor 301 may be a central processing unit (CPU), microprocessor, or other data processing chip, used to run program code stored in memory 302 or process data, such as the intelligent diagnosis method for OTA upgrade faults based on the TSP platform in this invention.

[0114] In some embodiments, display 303 may be an LED display, a liquid crystal display, a touch-sensitive liquid crystal display, or an OLED (Organic Light-Emitting Diode) touchscreen. Display 303 is used to display information from electronic device 300 and to display a visual user interface. Components 301-303 of electronic device 300 communicate with each other via a system bus.

[0115] In some embodiments of the present invention, when the processor 301 executes the computer program in the memory 302, the following steps can be implemented: when a vehicle OTA upgrade is performed and fault diagnosis is triggered, OTA diagnostic data is collected according to preset data dimensions, including: cloud download dimension, vehicle dimension, historical fault dimension, vehicle operating status dimension, vehicle operation dimension, and vehicle model dimension; the fault type of the OTA upgrade fault is determined, and fault features and fault handling strategies are constructed for each fault type to obtain a fault classification library, wherein the fault type includes a first fault type corresponding one-to-one with the data dimensions and a second fault type divided under the vehicle model; the fault classification library is matched with the OTA diagnostic data for fault features. If the feature matching is successful, the fault type of the OTA diagnostic data is determined according to the fault feature matching result. If the feature matching fails, the machine learning model deployed in the TSP platform is called to perform fault identification on the OTA diagnostic data to obtain the fault type of the OTA diagnostic data; a corresponding fault handling strategy is generated according to the fault type of the OTA diagnostic data.

[0116] It should be understood that when the processor 301 executes the computer program in the memory 302, in addition to the functions described above, it can also perform other functions, as can be found in the description of the corresponding method embodiments above.

[0117] Furthermore, the embodiments of the present invention do not specifically limit the type of electronic device 300 mentioned. Electronic device 300 can be a mobile phone, tablet computer, personal digital assistant (PDA), wearable device, laptop computer, or other portable electronic device. Exemplary embodiments of portable electronic devices include, but are not limited to, portable electronic devices running iOS, Android, Microsoft, or other operating systems. The aforementioned portable electronic device can also be other portable electronic devices, such as a laptop computer with a touch-sensitive surface (e.g., a touch panel). It should also be understood that in some other embodiments of the present invention, electronic device 300 may not be a portable electronic device, but rather a desktop computer with a touch-sensitive surface (e.g., a touch panel).

[0118] In another aspect, the present invention also provides a non-transitory computer-readable storage medium storing a computer program thereon. When executed by a processor, the computer program performs the above-described intelligent diagnostic method for OTA upgrade faults based on a TSP platform. This method includes: when a vehicle undergoes an OTA upgrade and triggers fault diagnosis, collecting OTA diagnostic data according to preset data dimensions, including: cloud download dimension, vehicle dimension, historical fault dimension, vehicle operating status dimension, vehicle operation dimension, and vehicle model dimension; determining the fault type of the OTA upgrade fault, and constructing fault features and fault handling strategies for each fault type to obtain a fault classification library, wherein the fault type includes a first fault type corresponding one-to-one with the data dimensions and a second fault type classified under the vehicle model; matching the fault classification library with the OTA diagnostic data for fault features; if the feature matching is successful, determining the fault type of the OTA diagnostic data based on the fault feature matching result; if the feature matching fails, calling a machine learning model deployed in the TSP platform to perform fault identification on the OTA diagnostic data to obtain the fault type of the OTA diagnostic data; and generating a corresponding fault handling strategy based on the fault type of the OTA diagnostic data.

[0119] Those skilled in the art will understand that all or part of the processes of the methods described in the above embodiments can be implemented by a computer program instructing related hardware, and the program can be stored in a computer-readable storage medium. The computer-readable storage medium may be a disk, optical disk, read-only memory, or random access memory, etc.

[0120] The above provides a detailed description of the intelligent diagnostic system and method for OTA upgrade faults based on the TSP platform provided by this invention. Specific examples have been used to illustrate the principles and implementation methods of this invention. The descriptions of the above embodiments are only for the purpose of helping to understand the method and core ideas of this invention. At the same time, for those skilled in the art, there will be changes in the specific implementation methods and application scope based on the ideas of this invention. Therefore, the content of this specification should not be construed as a limitation of this invention.

Claims

1. An intelligent fault diagnosis system for OTA upgrades based on the TSP platform, characterized in that, include: The vehicle-mounted multi-source data synchronous acquisition module is used to collect OTA diagnostic data according to preset data dimensions when the vehicle is upgraded over-the-air (OTA) and fault diagnosis is triggered. The data dimensions include: cloud download dimension, vehicle dimension, historical fault dimension, vehicle operating status dimension, vehicle operation dimension, and vehicle model dimension. The fault classification library construction module is used to determine the fault type of OTA upgrade failure, and to construct fault features and fault handling strategies for each fault type to obtain the fault classification library. The fault type includes a first fault type that corresponds one-to-one with the data dimension and a second fault type that is divided according to the first fault type under the vehicle model. The fault intelligent diagnosis module is used to match the fault classification library with the OTA diagnostic data for fault features. If the feature matching is successful, the fault type of the OTA diagnostic data is determined based on the fault feature matching result. If the feature matching fails, the machine learning model deployed in the TSP platform is called to identify the fault in the OTA diagnostic data and obtain the fault type of the OTA diagnostic data. The fault diagnosis result output module is used to generate corresponding fault handling strategies based on the fault type of the OTA diagnostic data.

2. The OTA upgrade fault intelligent diagnosis system based on the TSP platform according to claim 1, characterized in that, The vehicle-mounted multi-source data synchronous acquisition module is also used to trigger fault diagnosis when the vehicle OTA upgrade meets any of the following fault conditions; The status code for a single vehicle's OTA upgrade indicates that the OTA upgrade has failed. The TSP platform detected an error code from the vehicle's embedded data points during an OTA upgrade of a single vehicle. When performing OTA upgrades on the same vehicle model in batches, the percentage of vehicles with failed OTA upgrades or error codes detected by the TSP platform at the vehicle mounting points reaches a preset threshold.

3. The OTA upgrade fault intelligent diagnosis system based on the TSP platform according to claim 1, characterized in that, The vehicle-mounted multi-source data synchronous acquisition module is also used to establish a communication link from the vehicle to the cloud after triggering fault diagnosis. The communication link is used to transmit the collected OTA raw data. The original OTA data is timestamped, and data features associated with OTA upgrade failures are extracted from the original OTA data. Each data feature is then defined as a data dimension. The raw OTA data is divided according to the data dimensions to obtain OTA diagnostic data.

4. The OTA upgrade fault intelligent diagnosis system based on the TSP platform according to claim 1, characterized in that, The intelligent fault diagnosis module is also used to perform fault feature matching in the following manner; Based on the data dimensions of the OTA diagnostic data, preset rules are retrieved from the fault classification library; The corresponding fault characteristics are determined from the preset rules retrieved; The fault type corresponding to the fault feature is determined as the fault type of the OTA diagnostic data; If no corresponding preset rule is found in the fault classification library based on the data dimensions of the OTA diagnostic data, then feature matching is determined to have failed.

5. The OTA upgrade fault intelligent diagnosis system based on the TSP platform according to claim 1, characterized in that, The intelligent fault diagnosis module is also used to train a machine learning model in the following manner: Fault data samples of various vehicle models during OTA upgrades were obtained from the historical database of the TSP platform. The faulty data sample is cleaned to obtain a clean sample; Extract derived features of the corresponding vehicle models from the clean samples, and calculate the correlation coefficient between the derived features and the fault type of the OTA upgrade failure; Remove derived features with correlation coefficients less than a threshold, and train a machine learning model based on the remaining derived features.

6. The OTA upgrade fault intelligent diagnosis system based on the TSP platform according to claim 1, characterized in that, The fault intelligent diagnosis module is also used to obtain feedback results of the fault handling strategy in handling OTA upgrade faults; The feedback results are stored as new fault data in the historical database of the TSP platform to obtain the updated historical database. When the incremental value of newly added fault data in the updated historical database is greater than a preset incremental threshold, the machine learning model is updated based on the updated historical database.

7. The OTA upgrade fault intelligent diagnosis system based on the TSP platform according to claim 1, characterized in that, The fault diagnosis result output module is also used to construct a diagnostic report text based on the matching results of the fault features, the fault type of the OTA diagnostic data, and the fault handling strategy. The diagnostic report text is pushed through multiple output carriers, including engineer's terminal, car owner's terminal, and car manufacturer's server terminal.

8. A method for intelligent diagnosis of OTA upgrade faults based on the TSP platform, characterized in that, include: When a vehicle undergoes an OTA upgrade and triggers fault diagnosis, OTA diagnostic data is collected according to preset data dimensions, including: cloud download dimension, vehicle dimension, historical fault dimension, vehicle operating status dimension, vehicle operation dimension, and vehicle model dimension. The fault types of OTA upgrade failures are determined, and fault characteristics and fault handling strategies are constructed for each fault type to obtain a fault classification library. The fault types include a first fault type that corresponds one-to-one with the data dimensions and a second fault type that is divided under the vehicle model. The fault classification library is matched with the OTA diagnostic data for fault features. If the feature matching is successful, the fault type of the OTA diagnostic data is determined based on the fault feature matching result. If the feature matching fails, the machine learning model deployed in the TSP platform is called to identify faults in the OTA diagnostic data to obtain the fault type of the OTA diagnostic data. A corresponding fault handling strategy is generated based on the fault type in the OTA diagnostic data.

9. An electronic device, characterized in that, Including memory and processor, among which, The memory is used to store programs; The processor, coupled to the memory, is used to execute the program stored in the memory to implement the steps of the OTA upgrade fault intelligent diagnosis method based on the TSP platform as described in claim 8.

10. A non-transitory computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it implements the steps of the OTA upgrade fault intelligent diagnosis method based on the TSP platform as described in claim 8.

Images