Vehicle maintenance method and device, computer equipment and storage medium

Abstract

The invention relates to the field of vehicle maintenance, and discloses a vehicle maintenance method and device, computer equipment and a storage medium, and the vehicle maintenance method comprises the steps: collecting vehicle maintenance comprehensive data, carrying out the preprocessing of the vehicle maintenance comprehensive data, obtaining a driving behavior feature set, and analyzing the driving behavior feature set through a supervised learning model; and identifying a driving style, inputting the driving style and the driving behavior characteristics in the driving behavior characteristic set into the time sequence prediction model, obtaining a health assessment index, and generating a personalized vehicle maintenance plan according to the health assessment index in combination with a preset safety priority rule and a cost benefit rule. According to the vehicle maintenance method, the accuracy and timeliness of maintenance decision can be effectively improved, resource waste is reduced, the service life of the vehicle is prolonged, and the service experience of a user is improved.

Description

Technical Field

[0001] This application relates to the technical field of vehicle maintenance, and particularly to a vehicle maintenance method, device, computer device, and storage medium. Background Art

[0002] With the continuous growth of the number of automobiles in use, regular maintenance has become a necessary measure to ensure the safety and stability of vehicles. Currently, automobile manufacturers and maintenance service providers generally provide maintenance suggestions based on preset time periods or mileage, such as "change the engine oil every 5,000 kilometers" and "conduct a routine inspection every six months". This maintenance mode based on fixed intervals has the advantages of strong universality and simple implementation, and is thus widely adopted.

[0003] However, this standardized maintenance strategy also exposes many limitations in practical applications. On the one hand, fixed intervals cannot effectively reflect the differences in driving behaviors of different users. For example, frequent hard braking, high-speed driving, or long-term idling will significantly affect the service life of components. On the other hand, environmental factors, such as temperature, humidity, road conditions, and meteorological changes, also have an important impact on vehicle performance and maintenance requirements. In addition, although many current systems have the ability to collect data, they lack in-depth analysis of the collected data, resulting in a lag in vehicle condition assessment, insufficient prediction accuracy, and real-time response. At the same time, the role of user feedback in traditional maintenance systems is limited, and there is a lack of an effective closed-loop mechanism for model optimization and plan update.

[0004] Therefore, how to dynamically perceive the vehicle condition, understand the user's driving pattern, and generate a scientific and reasonable personalized maintenance plan based on multi-source heterogeneous data in combination with artificial intelligence technology has become an important problem that needs to be solved urgently at present. Summary of the Invention

[0005] In view of this, the embodiments of this application provide a vehicle maintenance method, device, computer device, and storage medium, which can effectively solve the problems in the prior art that the maintenance strategy lacks personalization, cannot dynamically reflect the true state of the vehicle, and it is difficult to make accurate maintenance decisions based on multi-source data.

[0006] In a first aspect, the embodiments of this application provide a vehicle maintenance method, including:

[0007] Collect comprehensive vehicle maintenance data;

[0008] Preprocess the comprehensive vehicle maintenance data to obtain a set of driving behavior characteristics;

[0009] Use a supervised learning model to analyze the set of driving behavior characteristics to identify the driving style, and input the driving style and the set of driving behavior characteristics into a time series prediction model to obtain a health assessment index;

[0010] Generate a personalized vehicle maintenance plan based on the health assessment index and in combination with preset safety priority rules and cost-benefit rules.

[0011] In some embodiments, the collection of comprehensive vehicle maintenance data includes:

[0012] Collect vehicle operating status data through the on-vehicle OBD interface;

[0013] Collect environmental status data through external sensors and the GPS module;

[0014] Receive driving scenario data input by the user through the mobile application interface;

[0015] Obtain meteorological data by calling the weather API interface;

[0016] Align the vehicle operating status data, the environmental status data, the driving scenario data, and the meteorological data with a unified timestamp to form the comprehensive vehicle maintenance data.

[0017] In some embodiments, the preprocessing of the comprehensive vehicle maintenance data to obtain a driving behavior feature set includes:

[0018] Perform standardization processing on the vehicle operating status, the environmental status data, the driving scenario data, and the meteorological data respectively to obtain standard vehicle operating status data;

[0019] Extract key driving behavior features based on the standard vehicle operating status data to form the driving behavior feature set.

[0020] In some embodiments, the use of a supervised learning model to analyze the driving behavior feature set to identify the driving style, and input the driving style and the driving behavior feature set into a time series prediction model to obtain a health assessment index includes:

[0021] Use the supervised learning model to analyze the key driving behavior features in the driving behavior feature set to obtain the driving style classification result;

[0022] Input the key driving behavior features and the driving style classification result into the time series prediction model to evaluate the health status of each component of the vehicle and obtain the life prediction value of each component of the vehicle;

[0023] Calculate the health assessment index of each component of the vehicle according to the life prediction value.

[0024] In some embodiments, after generating the health assessment index of each component of the vehicle, it further includes:

[0025] Based on the historical vehicle operation status data and the real-time vehicle operation status data, correct the health assessment index to obtain the corrected health assessment index;

[0026] Divide each component of the vehicle into different maintenance levels according to the corrected health assessment index.

[0027] In some embodiments, according to the health assessment index, combined with the preset safety priority rule and cost-benefit rule, generate a personalized vehicle maintenance plan, including:

[0028] Obtain each component of the vehicle whose health assessment index is lower than the threshold corresponding to the maintenance level, and determine the corresponding candidate maintenance tasks;

[0029] For each candidate maintenance task, evaluate the degree of influence of each component of the vehicle on vehicle safety, and generate the corresponding safety risk weight;

[0030] Combined with the cost-benefit rule, calculate the risk reduction benefit value for each candidate maintenance task;

[0031] Based on the safety risk weight and the risk reduction benefit value, calculate the utility score for each candidate maintenance task;

[0032] Rank all the candidate maintenance tasks according to the utility score to generate the personalized vehicle maintenance plan.

[0033] In some embodiments, after generating the personalized vehicle maintenance plan, it further includes:

[0034] Based on the historical maintenance records and the preset maintenance cycle, regularly check the status of the maintenance tasks, and mark the maintenance tasks that meet the deadline judgment conditions as the upcoming expired maintenance tasks;

[0035] According to the preset push strategy, send the upcoming expired maintenance tasks including the maintenance content and the deadline to the user terminal to remind the user;

[0036] And, receive the user's maintenance consultation based on the built-in voice module, use natural language processing technology to identify the problem type, and match the solution based on the knowledge base to generate a personalized reply.

[0037] In a second aspect, an embodiment of the present application provides a vehicle maintenance device, including:

[0038] A data acquisition module for acquiring comprehensive vehicle maintenance data;

[0039] A data processing module for preprocessing the comprehensive vehicle maintenance data to obtain a driving behavior feature set;

[0040] An index acquisition module, configured to analyze the driving behavior feature set by using a supervised learning model, obtain a driving style classification result, combine the driving style classification result with a time series prediction model, evaluate the health status of each component of the vehicle, and obtain a health evaluation index;

[0041] A maintenance plan generation module, configured to generate a personalized vehicle maintenance plan according to the health evaluation index, in combination with preset safety priority rules and cost-benefit rules.

[0042] In a third aspect, an embodiment of the present application provides a computer device, which includes a processor and a memory. The memory stores a computer program, and the processor is configured to execute the computer program to implement the vehicle maintenance method in the first aspect above.

[0043] In a fourth aspect, an embodiment of the present application provides a computer-readable storage medium, and when the computer program is executed on a processor, the vehicle maintenance method in the first aspect above is implemented.

[0044] The embodiments of the present application have the following beneficial effects:

[0045] The vehicle maintenance method of the present application collects comprehensive maintenance data such as vehicle operation status, environmental data, and user driving behavior, extracts a driving behavior feature set through preprocessing, analyzes the driving behavior feature set by using a supervised learning model, identifies the driving style, inputs the driving style and driving behavior features into a time series prediction model, generates a health evaluation index for each key component of the vehicle, and dynamically reflects its performance degradation trend and potential failure risk. On this basis, in combination with preset safety priority rules and cost-benefit rules, a personalized maintenance plan that highly matches the actual state of the current vehicle is output. The vehicle maintenance method of the present application not only avoids waste of resources caused by over-maintenance, but also effectively reduces safety hazards caused by maintenance delays, helps to extend the service life of the vehicle and improve the operation and maintenance efficiency. In addition, through intelligent processing of the content input by the user based on a natural language processing model, a qualitative improvement has been achieved in terms of prediction accuracy and response speed, enabling users to obtain more forward-looking and personalized maintenance suggestions, further enhancing the service experience and customer satisfaction. Description of the Drawings

[0046] In order to more clearly illustrate the technical solutions of the embodiments of the present application, the following will briefly introduce the drawings required in the embodiments. It should be understood that the following drawings only show some embodiments of the present application, and therefore should not be regarded as limiting the scope. For those of ordinary skill in the art, other relevant drawings can be obtained based on these drawings without creative efforts.

[0047] Figure 1Shows a flowchart of a vehicle maintenance method according to an embodiment of the present application;

[0048] Figure 2 Shows a schematic diagram of generating a personalized vehicle maintenance plan according to the utility score in a vehicle maintenance method according to an embodiment of the present application;

[0049] Figure 3 Shows a schematic structural diagram of a vehicle maintenance device according to an embodiment of the present application. Detailed implementation manners

[0050] Next, the technical solutions in the embodiments of the present application will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present application. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all the embodiments.

[0051] Generally, the components of the embodiments of the present application described and shown in the drawings here can be arranged and designed in various different configurations. Therefore, the following detailed description of the embodiments of the present application provided in the drawings is not intended to limit the scope of the present application claimed, but merely represents selected embodiments of the present application. All other embodiments obtained by those skilled in the art based on the embodiments of the present application without creative efforts belong to the scope of protection of the present application.

[0052] In the following, the terms "including", "having" and their cognates that can be used in various embodiments of the present application are only intended to represent specific features, numbers, steps, operations, elements, components or combinations of the foregoing items, and should not be construed as first excluding the existence of one or more other features, numbers, steps, operations, elements, components or combinations of the foregoing items or increasing the possibility of one or more features, numbers, steps, operations, elements, components or combinations of the foregoing items. In addition, the terms "first", "second", "third", etc. are only used for distinguishing descriptions and cannot be understood as indicating or implying relative importance.

[0053] Unless otherwise defined, all terms (including technical terms and scientific terms) used here have the same meaning as commonly understood by those of ordinary skill in the art to which the various embodiments of the present application belong. The terms (such as those defined in a general-use dictionary) will be interpreted as having the same meaning as the contextual meaning in the relevant technical field and will not be interpreted as having an idealized meaning or being overly formal, unless clearly defined in the various embodiments of the present application.

[0054] Next, some embodiments of the present application will be described in detail in conjunction with the accompanying drawings. Without conflict, the following embodiments and the features in the embodiments can be combined with each other.

[0055] Considering the problems in the prior art that the maintenance strategy lacks personalization, cannot dynamically reflect the true state of the vehicle, and it is difficult to make accurate maintenance decisions based on multi-source data, therefore, a vehicle maintenance method is proposed. By collecting multi-source data such as the vehicle running state, environmental information, and driving behavior, a driving behavior feature set is obtained after processing. A supervised learning model is used to identify the driving style. Combining the driving style and driving behavior features input into the time series prediction model, a health index is obtained to dynamically evaluate the health state of each component of the vehicle. Combining the safety priority and cost-benefit rules, a personalized maintenance plan is generated to achieve intelligent and refined vehicle maintenance management.

[0056] Figure 1 A flowchart of the vehicle maintenance method according to an embodiment of the present application is shown. Exemplarily, the method includes the following steps:

[0057] Step S100, collect comprehensive vehicle maintenance data.

[0058] Exemplarily, the comprehensive vehicle maintenance data is multi-source data collected from multiple dimensions such as vehicle operation, environmental perception, user input, and external information services, which can comprehensively reflect the current running state, usage environment, driving behavior, and meteorological conditions of the vehicle, providing data support for subsequent health assessment and maintenance decision-making.

[0059] In an alternative embodiment, collecting comprehensive vehicle maintenance data includes:

[0060] First, collect vehicle running state data through the on-vehicle OBD interface. The vehicle running state data includes but is not limited to real-time running parameters such as engine speed, accelerator pedal depth, braking frequency, braking pressure, and vehicle speed. Among them, the sampling frequency is set to not less than 10Hz to ensure the timeliness and continuity of the collected data and capture short-term driving behavior characteristics such as sudden acceleration and sudden braking. It should be noted that the collected data is uploaded in the way of "mainly real-time collection and supplemented by event triggering": under normal circumstances, the raw data collected at high frequency is compressed and packaged in a 1-minute cycle and uploaded to the cloud to balance real-time performance and network load; under abnormal working conditions, such as detecting a sudden drop in braking pressure or the engine speed exceeding the set threshold, a high-priority data upload channel is automatically triggered to ensure the immediate reporting and processing of key safety data.

[0061] Secondly, environmental status data is collected through temperature and humidity sensors and GPS modules deployed outside the vehicle. The environmental status data includes the temperature, humidity, and corresponding geographical location information at the vehicle's location. To ensure the consistency and alignability of multi-source sensor data, the data collected by each sensor is attached with an accurate timestamp and synchronized using a unified time reference for subsequent time-series analysis and model input integration. In practical applications, sensor calibration strategies and fusion algorithms (e.g., Kalman filtering) can be used to preprocess the collected data, thereby improving the accuracy and stability of temperature and humidity readings and geolocation data, and ensuring the reliability and analysis effectiveness of environmental status data under complex road conditions and climate conditions.

[0062] Thirdly, driving scenario data input by the user is received through the mobile application interface. The driving scenario data includes subjective descriptions by the user of historical maintenance and repair records, long-distance driving conditions, and special road conditions, as well as information such as personalized maintenance preferences. "Special road conditions" refer to environmental or road conditions different from normal driving conditions, such as extreme weather (heavy snow, heavy rain), complex roads (muddy, potholed), and situations such as long-distance driving and mountain driving. These factors may impose additional loads on key vehicle components and affect the component wear rate and health assessment results. The data input by the user is formatted and then structurally integrated into the vehicle data system as an important reference input for subsequent modeling and maintenance plan generation to improve the personalization and accuracy of maintenance strategies.

[0063] Subsequently, the weather API interface is called to obtain meteorological data, which includes the air quality index, rainfall, temperature fluctuations, and meteorological warning information, etc. For example, by spatially matching the meteorological data with the GPS location, accurate environmental meteorological characteristics can be obtained.

[0064] Finally, after the above data is collected, the vehicle operation status data, environmental status data, driving scenario data, and meteorological data are aligned with unified timestamps, thus forming vehicle maintenance comprehensive data with consistent structure and unified time series.

[0065] For example, assume that vehicle operation status data is collected at a frequency of 10 Hz, generating 10 data points per second, with timestamps such as 10:15:00.123, 10:15:00.223, etc.; while environmental status data is collected once per minute by a temperature and humidity sensor, with the recorded time being 10:15:00; at the same time, the time when the user submits driving scenario data through a mobile application may be 10:15:05, and the recorded time of the meteorological data returned by the weather API is also 10:15:00. By aligning the timestamps, these data are mapped to the same time window. For example, all data is summarized within the time period from 10:15:00 to 10:15:59, ensuring that data from each data source is synchronously processed under the same time sequence, enabling multi-source data to have a consistent time sequence basis in subsequent feature fusion and model analysis, thereby improving the accuracy of data integration and the reliability of model prediction.

[0066] In the case of unstable or interrupted network connection, the comprehensive vehicle maintenance data is temporarily stored in the local buffer to ensure data integrity and continuity; after the network is restored, the data is uploaded to the cloud server through an encryption protocol for subsequent health status assessment and personalized maintenance plan generation.

[0067] Step S200: Preprocess the comprehensive vehicle maintenance data to obtain a set of driving behavior features.

[0068] Exemplarily, use data preprocessing techniques to standardize multi-source data such as vehicle operation status, environmental status data, driving scenario data, and meteorological data to obtain standard vehicle operation status data. Then, based on these standard data, use machine learning algorithms to extract key driving behavior features to form a set of driving behavior features of the user. This process can effectively integrate different types of vehicle data, improve the structured degree and consistency of the data, and provide a solid foundation for subsequent driving style recognition and vehicle health assessment.

[0069] In an optional embodiment, preprocessing the comprehensive vehicle maintenance data to obtain a set of driving behavior features includes:

[0070] Step S201: Standardize the vehicle operation status, environmental status data, driving scenario data, and meteorological data respectively to obtain standard vehicle operation status data.

[0071] For vehicle operation status data, according to the valid value ranges of vehicle models and preset configuration parameters, identify and eliminate outliers. For example, an engine speed exceeding 7000 RPM is marked as an "overspeed event", a brake pressure below 5 Bar triggers an emergency alarm, and a vehicle speed exceeding 220 km / h is treated as abnormal data. For missing items, use linear interpolation of adjacent time points or a method based on historical means to fill them to ensure data integrity. Subsequently, use the min-max normalization method to normalize all numerical parameters to a unified standard interval of [0,1] to improve data comparability and calculation efficiency.

[0072] For environmental status data: temperature, humidity, air quality index, etc., perform the same standardization process to ensure data consistency and comparability. For example, build an association mapping between environmental variables and vehicle component status based on a large amount of historical data to obtain the impact of ambient temperature on vehicle battery performance, the corrosion risk of ambient humidity on the vehicle's electrical system, etc. And calculate the performance loss coefficient under the temperature-performance curve, the corrosion risk score under the humidity level, etc.

[0073] For driving scenario data: special road condition descriptions, long-distance driving records, etc., perform encoding and standardization to convert them into numerical codes for easy model processing.

[0074] Meteorological data: real-time weather data (air quality index, rainfall), etc., also undergo corresponding standardization. For example, normalize rainfall to a fixed range of [0,1].

[0075] Step S202, based on the standard vehicle operation status data, extract key driving behavior characteristics to form a driving behavior characteristic set.

[0076] After completing data standardization, based on the standard vehicle operation status data, use machine learning algorithms to extract key driving behavior characteristics to form a driving behavior characteristic set. Specifically, by statistically analyzing behavior parameters such as average acceleration, average deceleration rate, number of emergency brakes, and continuous idling time within a unit time or unit mileage. For example, the number of emergency brakes is cumulatively counted based on the vehicle deceleration exceeding -3 m / s 2 and the continuous idling time is calculated by detecting the time period when the engine speed is below the idling threshold and the vehicle is stationary. These characteristics comprehensively reflect the driver's operation style and driving habits, providing a basis for subsequent driving style modeling.

[0077] Step S300, use a supervised learning model to analyze the driving behavior characteristic set, identify the driving style, and input the driving style and the driving behavior characteristic set into a time series prediction model to obtain a health assessment index.

[0078] Analyze the driving behavior feature set through a supervised learning model to obtain the driving style. Then, input the driving style and the driving behavior feature set into a time series prediction model to evaluate the health status of each vehicle component, thereby obtaining a health assessment index. This process can accurately identify the driving styles of different users and, based on this, conduct personalized evaluations of the health status of vehicle components, providing a scientific basis for subsequent personalized maintenance plans.

[0079] In an alternative embodiment, a supervised learning model is used to analyze the driving behavior feature set, identify the driving style, and input the driving style and the driving behavior feature set into a time series prediction model to obtain a health assessment index, including:

[0080] Step S301: Use a supervised learning model to analyze the key driving behavior features in the driving behavior feature set to obtain a driving style classification result.

[0081] Exemplarily, a pre-trained supervised learning model is used to analyze the key driving behavior features in the driving behavior feature set. The supervised learning model can adopt algorithms such as XGBoost and LightGBM, and use samples containing known driving style labels for fitting during the training stage to accurately identify the style type of newly input data during the prediction stage. According to the input driving behavior features, such as the frequency of hard braking, average acceleration, and idle time, the model outputs the driving style, which is usually divided into aggressive (rapid acceleration and hard braking), mild (smooth acceleration and deceleration), economical (minimizing fuel consumption), etc. And it can also output the vehicle status, which covers various degrees from good to in need of repair, such as the degree of newness, wear level, and fault history.

[0082] Step S302: Input the driving behavior features and the driving style classification result into a time series prediction model to evaluate the health status of each vehicle component and obtain the life prediction values of each vehicle component.

[0083] Subsequently, the extracted driving behavior features and the driving style classification result obtained by the supervised learning model are jointly input into a time series prediction model. Among them, the time series model for prediction can adopt deep learning structures such as LSTM (Long Short-Term Memory Network) and GRU (Gated Recurrent Unit) to mine the trend of input features changing over time and predict the remaining life or failure probability of each vehicle component (such as the braking system, battery system, cooling system, etc.). This time series prediction model models and dynamically analyzes the health status of each vehicle component based on historical and real-time data during the training process, thereby predicting the remaining service life or failure probability of each component. These life prediction values provide a reliable quantitative basis for subsequent maintenance decisions and the formulation of maintenance plans.

[0084] Step S303: Calculate the health assessment index of each vehicle component according to the life prediction value.

[0085] After obtaining the life prediction result, calculate the health assessment index of each component based on the remaining life or potential failure probability of the component obtained from the prediction model. One is the remaining useful life (RUL) of the component, which is used to evaluate the time or mileage that the component can continue to work under the current state; the other is the probability of failure of the component, that is, the possibility of predicting its abnormal function or performance degradation in a future period.

[0086] The health index (HI) is a standardized index, usually set in the range of 0 to 100, and is used to quantify the current health status of each vehicle component. Different calculation strategies are selected according to different model output types:

[0087] When the model output is the remaining life, the health assessment index can be obtained by normalizing the ratio of the current remaining life to the reference life at the time of factory shipment or at the beginning of maintenance. The longer the remaining life, the better the component state, and the higher the corresponding HI value;

[0088] When the model output is the failure probability, the health assessment index can be calculated by the reverse mapping method, that is, HI = (1 - failure probability) × 100, which represents the health and safety level under the current state. The higher the failure probability, the lower the HI value.

[0089] In an alternative embodiment, after generating the health assessment index of each vehicle component, it further includes:

[0090] Based on the historical vehicle operation state data and the real-time vehicle operation state data, correct the health assessment index to obtain the corrected health assessment index.

[0091] Exemplarily, the health assessment index reflects the current performance status and potential failure risks of key components of the vehicle. Although this initial value has a certain predictive reference significance, due to the large individual differences in the wear condition, service life, and historical fault repair records of the vehicle during use, there may be deviations when directly using this value. Therefore, historical vehicle operation state data and real-time collected vehicle operation state data are introduced to dynamically correct the health assessment index. The historical vehicle operation state data includes the cumulative driving duration, wear degree, service life, past maintenance records, and repair fault frequency of the component since it left the factory; the real-time vehicle operation state data includes the working temperature, load condition, and abnormal events (such as abnormal vibration, overheating, or current fluctuation) within the latest cycle. By constructing a fusion model or rule engine, these historical vehicle operation state data and real-time vehicle operation state data are weighted and input to adjust the deviation of the health assessment index to improve its consistency with the actual component state. This processing method can keep the model prediction result synchronized with the actual use situation of the vehicle.

[0092] Each component of each vehicle is divided into different maintenance levels according to the corrected health assessment index.

[0093] Specifically, the corrected health assessment value is compared with the preset maintenance level threshold interval, and the corresponding maintenance urgency is judged according to the interval where the index falls. The following maintenance levels can be set:

[0094] When the health assessment value is lower than the first threshold (for example, HI value < 30), it is judged as the emergency maintenance level, indicating that there is a significant failure risk for this component and it should be repaired or replaced immediately;

[0095] When the health assessment value is in the second interval (for example, 30 ≤ HI < 70), it is judged as the recommended maintenance level, and it is recommended that the user arrange maintenance operations in the near future;

[0096] When the health assessment value is higher than the third threshold (for example, HI ≥ 70), it is judged as the optional maintenance level, that is, the component is operating well and no forced maintenance is required for the time being.

[0097] Step S400, according to the health assessment index of each component of the vehicle, combined with the preset safety priority rules and cost-benefit rules, generate a personalized vehicle maintenance plan.

[0098] Exemplarily, the safety priority rules are formulated based on the risks that component failures may pose to the safety of the entire vehicle. For example, when the health assessment index of the braking system is lower than a set threshold (e.g., HI < 30), a higher safety weight (e.g., 0.95) is assigned to this component, indicating that maintenance should be arranged as soon as possible even if the maintenance cost is high; for components with less impact on safety, a lower weight is given. At the same time, the cost-benefit rules are determined based on the estimated maintenance cost of each maintenance task and the benefit of reducing risks after maintenance. For example, although the cost of changing the brake fluid is high, if it can significantly reduce the risk of braking failure, its benefit value is high, thereby increasing the cost-benefit ratio of this task.

[0099] By comprehensively calculating the safety weight and the risk reduction benefit value of each candidate maintenance task, a utility score is generated, and all tasks are prioritized according to the score, finally forming a personalized maintenance plan that not only ensures vehicle safety but also has economic benefits.

[0100] In an alternative embodiment, in step S400, according to the health assessment index, combining the preset safety priority rules and cost-benefit rules, a personalized vehicle maintenance plan is generated, including:

[0101] Step S401, obtain each component of the vehicle whose health assessment index is lower than the threshold corresponding to the maintenance level, and determine the corresponding set of candidate maintenance tasks.

[0102] Specifically, based on the above maintenance level division results, components with a health assessment index lower than the recommended maintenance level (e.g., HI < 70) are screened out, and a set of candidate maintenance tasks is constructed accordingly.

[0103] Step S402, for each candidate maintenance task, evaluate the degree of impact of each vehicle component on vehicle safety, and generate the corresponding safety risk weight.

[0104] Then, a safety assessment is carried out for each candidate maintenance task. According to the preset safety priority rules, evaluate the potential impact of its failure on the safety of the entire vehicle operation, and assign a safety risk weight to it. For example, components directly affecting driving safety such as the braking system and tire system have a relatively high safety weight set (e.g., 0.95), while components with less impact on safety such as the ignition system are set to a relatively lower weight (e.g., 0.7).

[0105] Step S403, in combination with the cost-benefit rules, calculate the risk reduction benefit value for each candidate maintenance task.

[0106] On this basis, combined with the cost-benefit rule, a risk mitigation benefit analysis is carried out for each maintenance task. This analysis comprehensively considers factors such as the extended service life after maintenance, the degree of potential risk reduction, and maintenance costs, and calculates the risk reduction benefit value of each task. For example, replacing the brake fluid may bring a benefit value of 0.9, and replacing the spark plugs brings a benefit value of 0.8.

[0107] Step S404: Calculate the utility score of each candidate maintenance task based on the safety risk weight and the risk reduction benefit value.

[0108] Exemplarily, calculate the utility score of each maintenance task according to the following formula:

[0109] Utility score = safety risk weight × risk reduction benefit value ÷ maintenance cost;

[0110] As Figure 2 shown, two examples are used to illustrate:

[0111] Task A: Brake fluid replacement, safety weight 0.95, benefit 0.9, cost 70 yuan, and the utility score is (0.95×0.9) / 70≈0.012;

[0112] Task B: Spark plug replacement, safety weight 0.7, benefit 0.8, cost 70 yuan, and the utility score is (0.7×0.8) / 70≈0.008;

[0113] Prioritize Task A with a higher score and include it in the vehicle maintenance plan.

[0114] Step S405: Sort all candidate maintenance tasks according to the utility score to generate a personalized vehicle maintenance plan.

[0115] Finally, sort all candidate maintenance tasks according to the utility score to generate a personalized vehicle maintenance plan. For example, the content of this maintenance plan may include, but is not limited to: a list of maintenance items, recommended execution times, estimated workloads, and priority levels, etc.

[0116] In an alternative embodiment, after generating the personalized vehicle maintenance plan in step S400, the method further includes:

[0117] Regularly check the status of maintenance tasks based on historical maintenance records and preset maintenance cycles, and mark the maintenance tasks as upcoming due when the maintenance task status meets the deadline judgment conditions.

[0118] According to the preset push strategy, the maintenance tasks that are about to expire, including maintenance content and deadlines, are sent to the user terminal to remind the user. In addition, the AI assistant receives the user's maintenance consultation, uses natural language processing technology to identify the type of problem, and matches the solution based on the knowledge base to generate personalized replies.

[0119] First, based on the most recent maintenance time of each component in the historical maintenance records and the corresponding preset maintenance cycle (which can be based on time or mileage), a periodic status tracking table for maintenance tasks is established. For example: the oil change cycle is 6 months or 5,000 kilometers, the air conditioning filter replacement cycle is 12 months, etc. Run the scheduled task module regularly (for example, daily or every time the vehicle starts), call the current time and mileage data, and compare it with the last completion time and cycle conditions of each maintenance task.

[0120] If a maintenance task meets the deadline judgment condition (for example, it is close to 80% of the maintenance cycle, or there are less than 7 days left before the maintenance deadline), the task is marked as "expiring soon" and its status field in the maintenance schedule is updated. For example, when a tire maintenance task is set to "6 months / 10,000 kilometers", it is currently the 5th month and has traveled 9,200 kilometers, so it is judged to meet the expiration warning condition.

[0121] When a maintenance task is detected to be in the "expiring soon" state, the task is pushed to the user terminal according to the preset push strategy. The push strategy may include: reminder time (7 days / 3 days / 1 day in advance), priority level (whether it is a critical safety component), push method (APP pop-up window, SMS, email), etc.

[0122] The push content includes: maintenance item name; recommended maintenance time or mileage node; deadline; explanatory advice (for example, "To ensure driving safety, it is recommended to arrange this maintenance as soon as possible").

[0123] Users can click the notification on the terminal to jump to the maintenance plan details page, view the plan items and make an appointment to execute them. The model records user clicks and feedback information as input for subsequent AI interaction optimization.

[0124] After the maintenance plan is pushed or during daily use, users can use the terminal's AI assistant module to provide maintenance consultation in voice or text form. For example, ask questions such as "Why have the brakes become soft recently?" "Should I change the tires now?" etc.

[0125] Based on the natural language processing model (NLP), the user input content is semantically parsed to extract keywords and question intent. After identification, the question is mapped to a preset question type label (for example, "brake problem", "maintenance timing", "vehicle noise", etc.), and the professional maintenance knowledge base is called for question matching and answer generation.

[0126] Among them, the knowledge base includes various types of information such as standard maintenance suggestions, component failure mechanisms, and fault diagnosis paths. Based on the identified tags and context parameters, personalized and actionable solution suggestions are generated. For example, for the problem of "brake not sensitive", the suggestions may be "check the brake fluid level", "check whether the thickness of the brake pads is less than 2 mm", etc. The model will also record the user feedback results, such as whether it is solved and whether the maintenance appointment button is clicked, for optimizing the next round of model answers and strategy scheduling.

[0127] Figure 3 A schematic structural diagram of a vehicle maintenance device according to an embodiment of the present application is shown. Exemplarily, the vehicle maintenance device 100 includes:

[0128] A data acquisition module 110, configured to acquire comprehensive vehicle maintenance data;

[0129] A data processing module 120, configured to preprocess the comprehensive vehicle maintenance data to obtain a driving behavior feature set;

[0130] An index acquisition module 130, configured to analyze the key driving behavior features in the driving behavior feature set by using the supervised learning model to obtain the driving style classification result;

[0131] A maintenance plan generation module 140, configured to generate a personalized vehicle maintenance plan according to the health assessment index, in combination with preset safety priority rules and cost-benefit rules.

[0132] It can be understood that the device in this embodiment corresponds to the method in the above embodiment, and the optional items in the above embodiment are equally applicable to this embodiment, so they will not be repeated here.

[0133] The present application also provides a computer device. Exemplarily, the computer device includes a processor and a memory. Among them, the memory stores a computer program, and the processor runs the computer program to enable the computer device to execute the functions of the above method or each module of the above device.

[0134] Among them, the processor can be an integrated circuit chip with signal processing capabilities. The processor can be a general-purpose processor, including at least one of a central processing unit (CPU), a graphics processing unit (GPU), a network processor (NP), a digital signal processor (DSP), an application-specific integrated circuit (ASIC), a field-programmable gate array (FPGA), or other programmable logic devices, discrete gate or transistor logic devices, and discrete hardware components. The general-purpose processor can be a microprocessor or the processor can also be any conventional processor, etc., and can implement or execute the various methods, steps, and logic block diagrams disclosed in the embodiments of the present application.

[0135] The memory can be, but is not limited to, a random access memory (RAM), a read-only memory (ROM), a programmable read-only memory (PROM), an erasable programmable read-only memory (EPROM), an electrically erasable programmable read-only memory (EEPROM), etc. Among them, the memory is used to store a computer program, and after receiving an execution instruction, the processor can execute the computer program accordingly.

[0136] The present application also provides a computer-readable storage medium for storing the computer program used in the above computer device. For example, the computer-readable storage medium can include, but is not limited to: various media such as a USB flash drive, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a magnetic disk, or an optical disc that can store program codes.

[0137] In several embodiments provided by this application, it should be understood that the disclosed devices and methods can also be implemented in other ways. The device embodiments described above are merely illustrative. For example, the flowcharts and structural diagrams in the drawings show the possible architectures, functions, and operations of devices, methods, and computer program products according to multiple embodiments of this application. In this regard, each block in the flowchart or block diagram can represent a module, a program segment, or a part of code, and the module, program segment, or part of code contains one or more executable instructions for implementing the specified logical function. It should also be noted that in an alternative implementation, the functions marked in the blocks can occur in a different order than that marked in the drawings. For example, two consecutive blocks can actually be executed substantially in parallel, and they can sometimes be executed in the reverse order, depending on the functions involved. It should also be noted that each block in the structural diagram and / or flowchart, as well as the combination of blocks in the structural diagram and / or flowchart, can be implemented by a dedicated hardware-based system that performs the specified functions or actions, or can be implemented by a combination of dedicated hardware and computer instructions.

[0138] In addition, each functional module or unit in various embodiments of this application can be integrated together to form an independent part, or each module can exist separately, or two or more modules can be integrated to form an independent part.

[0139] If the above functions are implemented in the form of software functional modules and sold or used as an independent product, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or a part of this 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 enable a computer device (which can be a smart phone, a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the methods described in various embodiments of this application.

[0140] As described above, the above are only the specific implementation manners of this application, but the protection scope of this application is not limited thereto. Any person skilled in the art within the technical scope disclosed by this application can easily think of changes or substitutions, which should all be covered by the protection scope of this application.

Claims

1. A vehicle maintenance method, characterized in that, The method includes: Collect comprehensive vehicle maintenance data; Preprocess the comprehensive vehicle maintenance data to obtain a set of driving behavior characteristics; Use a supervised learning model to analyze the set of driving behavior characteristics to identify the driving style, and input the driving style and the set of driving behavior characteristics into a time series prediction model to obtain a health assessment index; Generate a personalized vehicle maintenance plan according to the health assessment index, in combination with preset safety priority rules and cost-benefit rules.

2. The vehicle maintenance method according to claim 1, characterized in that, The collection of comprehensive vehicle maintenance data includes: Collect vehicle operation status data through an in-vehicle OBD interface; Collect environmental status data through external sensors and a GPS module; Receive driving scenario data input by the user through a mobile application interface; Obtain meteorological data by calling a weather API interface; Align the timestamps of the vehicle operation status data, the environmental status data, the driving scenario data, and the meteorological data to form the comprehensive vehicle maintenance data.

3. The vehicle maintenance method according to claim 2, wherein The preprocessing of the comprehensive vehicle maintenance data to obtain a set of driving behavior characteristics includes: Perform standardization processing on the vehicle operation status, the environmental status data, the driving scenario data, and the meteorological data respectively to obtain standard vehicle operation status data; Extract key driving behavior characteristics based on the standard vehicle operation status data to form the set of driving behavior characteristics.

4. The vehicle maintenance method according to claim 1, characterized in that, The use of a supervised learning model to analyze the set of driving behavior characteristics, identify the driving style, and input the driving style and the set of driving behavior characteristics into a time series prediction model to obtain a health assessment index includes: Use the supervised learning model to analyze the key driving behavior characteristics in the set of driving behavior characteristics to obtain the driving style classification result; Input the key driving behavior characteristics and the driving style classification result into the time series prediction model to evaluate the health status of each component of the vehicle and obtain the life prediction value of each component of the vehicle; Calculate the health assessment index of each component of the vehicle according to the life prediction value.

5. The vehicle maintenance method according to claim 1, characterized in that, After generating the health assessment index of each component of the vehicle, it further includes: Based on historical vehicle operation status data and real-time vehicle operation status data, correct the health assessment index to obtain a corrected health assessment index; Divide each component of the vehicle into different maintenance levels according to the corrected health assessment index.

6. The vehicle maintenance method according to claim 1, wherein The generation of a personalized vehicle maintenance plan according to the health assessment index, in combination with preset safety priority rules and cost-benefit rules, includes: Obtain each component of the vehicle whose health assessment index is lower than the threshold corresponding to the maintenance level, and determine the corresponding candidate maintenance tasks; For each candidate maintenance task, evaluate the degree of impact of each component of the vehicle on vehicle safety and generate a corresponding safety risk weight; In combination with the cost-benefit rule, calculate the risk reduction benefit value for each candidate maintenance task; Calculate the utility score of each candidate maintenance task based on the safety risk weight and the risk reduction benefit value. Prioritize all the candidate maintenance tasks according to the utility scores to generate the personalized vehicle maintenance plan.

7. The vehicle maintenance method according to claim 1, characterized in that, After generating the personalized vehicle maintenance plan, it further includes: Regularly check the status of maintenance tasks based on historical maintenance records and preset maintenance cycles, and mark the maintenance tasks that are about to expire when the status of the maintenance tasks meets the deadline judgment conditions; According to the preset push strategy, send the maintenance tasks that are about to expire, including maintenance content and deadline, to the user terminal to remind the user; And, receive the user's maintenance consultation based on the built-in voice module, use natural language processing technology to identify the problem type, and match the solution based on the knowledge base to generate a personalized reply.

8. A vehicle maintenance device, characterized in that, The device includes: A data acquisition module for acquiring comprehensive vehicle maintenance data; A data processing module for preprocessing the comprehensive vehicle maintenance data to obtain a set of driving behavior characteristics; An index acquisition module for using a supervised learning model to analyze the set of driving behavior characteristics to identify the driving style, and inputting the driving style and the set of driving behavior characteristics into a time series prediction model to obtain a health assessment index; A maintenance plan generation module for generating a personalized vehicle maintenance plan according to the health assessment index, in combination with preset safety priority rules and cost-benefit rules.

9. A computer device, characterized in that, The computer device includes a processor and a memory, the memory stores a computer program, and the processor is used to execute the computer program to implement the vehicle maintenance method according to any one of claims 1-7.

10. A computer-readable storage medium, characterized in that, It stores a computer program, and when the computer program is executed on the processor, it implements the vehicle maintenance method according to any one of claims 1-7.

Images