Method and system for generating maintenance suggestions for vehicles

Abstract

The present disclosure provides a method for generating service suggestions for a vehicle. The method comprises the following steps: (S1) in response to a fault diagnosis request, collecting real-time operation data of a vehicle and sending the collected real-time operation data to a cloud server; (S2) the cloud server performs fault diagnosis on the vehicle based on the real-time operation data to generate a total fault diagnosis result and total maintenance suggestion information; and (S3) a personalized maintenance plan of the vehicle is generated, the personalized maintenance plan comprises the total fault diagnosis result and the total maintenance suggestion information, and the personalized maintenance plan is sent to a requester of the fault diagnosis request. According to the method, the model, the batch and the historical fault data of the vehicle are considered, the accuracy of fault diagnosis is improved, the maintenance time and cost can be effectively reduced, and the overall maintenance efficiency of the vehicle is improved.

Description

Technical Field

[0001] The present invention relates to vehicle fault diagnosis and more particularly to a method and system for generating maintenance suggestions for a vehicle. Background Art

[0002] Remote diagnosis of automobile faults is a technology that uses communication technology and big data analysis to monitor and diagnose faults in real time. On-board sensors can collect the operating data of the vehicle and the real-time status information of each subsystem / component. These data are sent to the cloud server in real time through a data transmission module, such as a wireless network or a cellular network. In the cloud server, the data is processed and analyzed using specialized analysis software to identify potential faults, and the vehicle fault diagnosis report is sent to, for example, a smartphone application or a vehicle dashboard, allowing users or maintenance personnel to view the vehicle status and receive fault alerts. The vehicle fault diagnosis report can include the fault type, fault location, severity, scope of impact, and one or more feasible maintenance suggestions. Afterwards, maintenance personnel can perform repairs based on the recommended maintenance suggestions in the diagnostic report.

[0003] The vehicle fault diagnosis in the prior art is only based on the fault code, without considering the fault cause, fault location, and repair plan related to the historical faults. However, for a specific vehicle, depending on the corresponding model and factory batch, the prone vehicle faults may be concentrated on several specific components or software. At the same time, according to the user's operating habits, personalized settings, etc., the prone vehicle faults often have certain rules, that is, the faults that have occurred have a greater probability of recurring. The above historical fault information is not taken into account in the fault diagnosis scheme of the prior art.

[0004] In addition, the maintenance suggestions in the vehicle fault diagnosis results of the prior art are only aimed at vehicle maintenance personnel, and do not take into account that some faults are caused by improper driver operation or vehicle software system defects. Therefore, no operation suggestions or software and hardware design improvement suggestions are provided to drivers and vehicle manufacturers. Summary of the invention

[0005] The first aspect of the present disclosure provides a method for generating maintenance suggestions for a vehicle. In some embodiments, the method for generating maintenance suggestions for a vehicle may include: (S1) in response to a fault diagnosis request, collecting real-time operating data of the vehicle and sending the collected real-time operating data to a cloud server; (S2) the cloud server performs fault diagnosis on the vehicle based on the real-time operating data to generate a total fault diagnosis result and total maintenance suggestion information; and (S3) generating a personalized maintenance plan for the vehicle, the personalized maintenance plan including the total fault diagnosis result and the total maintenance suggestion information, and sending the personalized maintenance plan to the requester of the fault diagnosis request. In some embodiments, step S2 may include: (S21) performing historical fault diagnosis on the vehicle based at least in part on real-time operating data to generate a first fault diagnosis result Res1 and a first maintenance suggestion information Sug1, and / or (S22) performing real-time fault diagnosis on the vehicle based at least in part on real-time operating data to generate a second fault diagnosis result Res2 and a second maintenance suggestion information Sug2; (S23) generating a total fault diagnosis result, the total fault diagnosis result including the first fault diagnosis result Res1 and / or the second fault diagnosis result Res2; and (S24) generating total maintenance suggestion information. The total maintenance suggestion information may include the first maintenance suggestion information Sug1 and / or the second maintenance suggestion information Sug2. The total maintenance suggestion information may be classified into operation guidance suggestions, hardware maintenance suggestions, and software / hardware improvement suggestions using a maintenance suggestion classifier.

[0006] In some embodiments, the vehicle's historical fault table Ta, production batch related fault table Tb, and vehicle model related fault table Tc may be stored and updated in the cloud server. Historical fault diagnosis may be performed in the order of historical fault information of the vehicle, historical fault information of the vehicle's production batch, and historical fault information of the vehicle's model.

[0007] In some embodiments, when it is determined that a certain historical fault occurs again, the fault count corresponding to the historical fault is updated in the historical fault table Ta. When it is determined that a certain production batch related fault occurs again, the fault count corresponding to the production batch related fault is updated in the production batch related fault table Tb. When it is determined that a certain vehicle type related fault occurs again, the fault count corresponding to the vehicle type related fault is updated in the vehicle type related fault table Tc.

[0008] In some embodiments, historical fault diagnosis is performed for faults whose occurrence times or frequencies in the historical fault table Ta, the production batch related fault table Tb, and the vehicle model related fault table Tc exceed a threshold.

[0009] In some embodiments, historical fault diagnosis and / or real-time fault diagnosis can be further performed on the vehicle based on dynamic data. The dynamic data can include driving behavior data and / or environmental data.

[0010] In some embodiments, the historical fault table Ta may include the license plate number, production batch, model, fault information, the number of times the fault occurs, abnormal operation data, fault location, fault cause, and maintenance suggestion information corresponding to the fault information. The production batch related fault table Tb may include the production batch, fault information, the number of times the fault occurs, abnormal performance, fault cause, and maintenance suggestion information. The vehicle model related fault table Tc may include the model, fault information, the number of times the fault occurs, abnormal performance, fault cause, and maintenance suggestion information.

[0011] In some embodiments, the second fault diagnosis result Res2 may include abnormal operation data, fault location, possible fault cause, and the second maintenance suggestion information Sug2 may include fault symptom description, image of faulty component, maintenance method description, maintenance video, and any combination thereof.

[0012] In some embodiments, the total maintenance suggestion information can be sent together with the total fault diagnosis result to the requester of the fault diagnosis request. In some embodiments, the total fault diagnosis result and the operation guidance suggestion can be sent to the user of the vehicle, the total fault diagnosis result and the hardware maintenance suggestion can be sent to the maintenance personnel of the vehicle, and the total fault diagnosis result and the software / hardware improvement suggestion can be sent to the manufacturer of the vehicle.

[0013] In some embodiments, the method of the present disclosure may further include utilizing big data analysis to update the vehicle's production batch-related fault table Tb and vehicle model-related fault table Tc in the cloud server.

[0014] In some embodiments, the production batch related fault table Tb and the vehicle model related fault table Tc can be generated by counting the data in the historical fault table Ta.

[0015] In some embodiments, a machine learning algorithm may be used to adjust the threshold in real time based on changes in the vehicle's operating environment and / or the vehicle's actual operating conditions.

[0016] In some embodiments, if it is determined in the historical fault diagnosis that the same fault as the historical fault has occurred, the real-time fault diagnosis is not performed.

[0017] In some embodiments, whether to perform real-time fault diagnosis can be determined based on vehicle usage or user feedback.

[0018] In some embodiments, the method of the present disclosure may further include collecting user feedback and dynamically updating the maintenance suggestion classifier.

[0019] In some embodiments, the method of the present disclosure may further include removing duplicate or invalid data in the first fault diagnosis result Res1 and the second fault diagnosis result Res2 and / or the first maintenance suggestion information Sug1 and the second maintenance suggestion information Sug2 by data cleaning or feature selection.

[0020] In some embodiments, the method of the present disclosure may further include integrating the first fault diagnosis result Res1 and the second fault diagnosis result Res2 and / or the first maintenance suggestion information Sug1 and the second maintenance suggestion information Sug2 into a global view through multi-source data fusion.

[0021] In some embodiments, the method of the present disclosure may further include formulating a personalized maintenance plan based on the overall fault diagnosis results of the vehicle, the severity and type of the current fault, and dynamic data, wherein the dynamic data includes driving behavior data and / or environmental data.

[0022] In some embodiments, dynamic data is obtained from a sensor system of the vehicle and / or an external data source.

[0023] A second aspect of the present disclosure provides a system including one or more computer processors and a computer readable memory. The computer readable memory may include machine executable code. The machine executable code, when executed by the one or more computer processors, implements the method for generating maintenance recommendations for a vehicle of the present disclosure. BRIEF DESCRIPTION OF THE DRAWINGS

[0024] The novel features of the invention are set forth with particularity in the appended claims.A better understanding of the features and advantages of the invention will be obtained by referring to the following detailed description and accompanying drawings which set forth illustrative embodiments utilizing the principles of the invention.

[0025] Figure 1 The present invention is a flowchart of a method for generating maintenance suggestions for a vehicle according to an embodiment of the present application.

[0026] Figure 2 It is a flowchart of a method for generating overall fault diagnosis results and overall maintenance suggestion information according to an embodiment of the present application. DETAILED DESCRIPTION

[0027] In view of the problem that the vehicle fault diagnosis in the prior art only considers the fault information of components but does not consider the common problems related to the vehicle model and batch, resulting in poor pertinence of the maintenance suggestions, the present application provides a method for generating vehicle maintenance suggestions, which takes into account the vehicle model, batch and historical fault data, and not only improves the accuracy of fault diagnosis, but also effectively reduces maintenance time and cost, thereby improving the overall maintenance efficiency of the vehicle.

[0028] Figure 1 is a flow chart of a method for generating maintenance suggestions for a vehicle according to an embodiment of the present application. In some embodiments, Figure 1 The method shown may include steps S1-S3. In step S1, in response to a fault diagnosis request, real-time operating data of the vehicle is collected and the collected real-time operating data is sent to a cloud server. In step S2, the cloud server performs fault diagnosis on the vehicle based on the real-time operating data to generate a total fault diagnosis result and total maintenance suggestion information. In step S3, a personalized maintenance plan for the vehicle may be generated and sent to the requester of the fault diagnosis request. The personalized maintenance plan may include a total fault diagnosis result and total maintenance suggestion information.

[0029] The initiator / requester of the fault diagnosis request in step S1 may be the user of the vehicle, the maintenance personnel of the vehicle, the manufacturer of the vehicle, or a third-party monitoring service provider. The fault diagnosis request may be issued regularly, or triggered immediately when a specific abnormality is detected. The real-time operation data of the vehicle may include, but is not limited to, the mode, state, signal, and driving data of the vehicle. The mode of the vehicle may be the driving mode of the vehicle, such as economy, sports or off-road mode, or manual driving mode and automatic driving mode. The state of the vehicle may include starting, accelerating, decelerating, changing lanes, braking, and parking. The signal of the vehicle may include signals of various subsystems or components of the vehicle. The signal of the vehicle may be collected by on-board sensors, covering multiple subsystems such as the on-board power system, transmission system, braking system, and suspension system. Non-limiting examples of vehicle signals include battery temperature, battery charge, engine or motor temperature and speed, transmission temperature and state, brake system pressure and temperature, suspension system pressure, tire pressure, wheel speed, steering angle, etc. The driving data of the vehicle may include current speed, acceleration, total mileage, fuel or battery consumption rate, navigation position, driving route, tire pressure status, steering dynamics, and lateral acceleration, etc.

[0030] These data can be transmitted to the cloud platform in real time through the on-board diagnostic system (OBD) or other data interfaces. Examples of cloud platforms include but are not limited to Amazon's AWS, Microsoft's Azure, Google Cloud, IBM Cloud, OracleCloud, Alibaba Cloud, Tencent Cloud, and Huawei Cloud. These platforms can provide large-scale computing power, edge computing support, and AI and big data processing capabilities. Such platforms can not only store and analyze large amounts of vehicle data, but also generate personalized feedback such as maintenance suggestions and driving behavior adjustment suggestions in real time, and support various machine learning models for fault prediction and preventive maintenance. At the same time, the cloud platform can be integrated with the on-board diagnostic system through APIs to achieve automated fault diagnosis and push repair solutions.

[0031] In some embodiments, Figure 2 As shown, step S2 in which the cloud server performs fault diagnosis on the vehicle based on real-time operating data to generate overall fault diagnosis results and overall maintenance suggestion information may include steps S21-S24. Figure 2 It is a flowchart of a method for generating overall fault diagnosis results and overall maintenance suggestion information according to an embodiment of the present application.

[0032] In step S21, the vehicle can be diagnosed with historical faults according to the real-time operation data of the vehicle, and the first fault diagnosis result Res1 and the first maintenance suggestion information Sug1 can be generated. In the cloud server, the vehicle's historical fault table Ta, the production batch-related fault table Tb, and the vehicle model-related fault table Tc can be stored and updated. The historical fault diagnosis in step S21 can be performed in the order of the vehicle's historical fault information, the vehicle's production batch's historical fault information, and the vehicle model's historical fault information. For example, if the vehicle has ever had a "low battery voltage" fault and has a corresponding maintenance record, it is possible to prioritize checking whether there is sensor data indicating the "low battery voltage" fault in the vehicle's real-time operation data. For another example, if other vehicles in the production batch to which the vehicle belongs have ever had a "engine temperature sensor abnormality" fault and have a corresponding maintenance record, it is possible to continue to check whether there is sensor data indicating the "engine temperature sensor abnormality" fault in the vehicle's real-time operation data. Similarly, if other vehicles of the same model as the current vehicle have experienced the "brake fluid volume abnormality" fault and have corresponding maintenance records, it is possible to continue to check whether there is sensor data indicating the "brake fluid volume abnormality" fault in the vehicle's real-time operation data. In this way, if multiple vehicles of a certain model or a certain production batch have experienced a certain fault multiple times in the historical records, the method disclosed in the present invention can focus on checking the fault of the vehicle to determine whether there is a similar problem.

[0033] In order to avoid the influence of occasional faults on historical fault diagnosis, in the process of historical fault diagnosis, only the faults whose number of occurrences or frequencies in the historical fault table Ta, the production batch related fault table Tb and the vehicle model related fault table Tc exceed a certain threshold can be diagnosed. In order to improve the accuracy of diagnosis, the threshold can be dynamically adjusted according to different vehicles, use environments and operating conditions by introducing an adaptive threshold. For example, the threshold can be automatically adjusted according to changes in the vehicle operating environment (such as temperature, humidity, altitude, etc.) to adapt to the possibility of faults under different working conditions. For example, in a high temperature environment, the threshold of abnormal battery voltage may need to be more sensitive, while the threshold of engine start problems in a cold environment may need to be adjusted. For example, the threshold can be adjusted in real time according to the actual operation of the vehicle (such as long-term low-speed driving, frequent braking, etc.). The method disclosed in the present invention can continuously optimize the threshold setting based on a machine learning algorithm. By analyzing the vehicle operation data under different environments, the machine learning model can gradually improve the prediction ability of different faults and optimize the threshold adjustment strategy. For example, by comparing the occurrence of faults in different vehicles under the same conditions, the method disclosed in the present invention can learn the optimal threshold setting to reduce misjudgment and missed diagnosis.

[0034] Some faults may have little impact on vehicle safety or performance, while others may lead to serious consequences. In some instances, different weights can be set for different fault types to optimize diagnostic efficiency. Faults that are low in frequency but potentially serious in impact can be diagnosed and handled first, even if they do not reach the threshold.

[0035] When it is determined that a certain historical fault has occurred again, the fault count corresponding to the historical fault can be updated in the historical fault table Ta, for example, the fault count + 1. Similarly, when it is determined that a certain production batch related fault has occurred again, the fault count corresponding to the production batch related fault is updated in the production batch related fault table Tb. When it is determined that a certain vehicle type related fault has occurred again, the fault count corresponding to the vehicle type related fault is updated in the vehicle type related fault table Tc.

[0036] In some embodiments, the historical fault diagnosis of the vehicle based at least in part on the real-time operation data of the vehicle in step S21 may include retrieving historical fault data identical or similar to the real-time operation data of the vehicle in the historical fault table Ta, the production batch related fault table Tb and / or the vehicle model related fault table Tc. In some instances, the search may be implemented based on case-based reasoning (CBR). In some instances, the method disclosed herein may extract the features of the real-time operation data of the vehicle and compare the similarity with the features of the data in the historical fault table Ta, the production batch related fault table Tb and / or the vehicle model related fault table Tc, thereby searching for historical fault data similar to the real-time operation data of the vehicle in the historical fault table Ta, the production batch related fault table Tb and / or the vehicle model related fault table Tc. The similarity comparison between data may be implemented by feature matching algorithm, weighted vector algorithm, and multidimensional data space mapping method. In the feature matching algorithm, the matching weight may be dynamically adjusted to adapt to different fault types and models. In the weighted vector algorithm, the weight may be adjusted according to the importance of each fault data to improve the accuracy of the retrieval.

[0037] The advantage of the present disclosure of saving and updating the historical fault table Ta, the production batch related fault table Tb and / or the vehicle model related fault table Tc in the cloud server is that it can continuously optimize its own diagnostic capabilities by accumulating new cases, realize self-learning of fault detection, and improve accuracy. In some implementations, when performing historical fault diagnosis of S21, in addition to historical faults, production batch related faults and / or vehicle model related faults, additional data sources can also be introduced, such as driving behavior (for example, the driver's acceleration, braking habits), environmental conditions (such as climate, road conditions, temperature, etc.), etc. These additional data can help to more comprehensively analyze the possible causes of the current fault, optimize the fault diagnosis process, and improve the adaptability of the method disclosed in this disclosure in a variety of scenarios. In one example, the vehicle had difficulty starting in cold weather many times. In this case, when performing historical fault diagnosis and / or real-time fault diagnosis, the method disclosed in the present invention can not only retrieve the historical faults of the vehicle, production batch-related faults and / or vehicle model-related faults, but also consider dynamic data such as changes in ambient temperature (for example, the temperature is below zero for a long time) and / or the driver's driving behavior (for example, frequent short-distance driving or frequent sudden acceleration, which may cause excessive battery consumption or increased engine load), and combine these dynamic data into fault diagnosis to identify possible battery or fuel system faults. Dynamic data can be obtained from the vehicle's sensor system or from external data sources (for example, public weather forecast systems, traffic dispatch systems, and navigation systems). By combining these dynamic data, the system can more accurately determine the root cause of the fault and make reasonable maintenance suggestions or preventive measures. In another example, the vehicle frequently brakes in a high temperature environment, which may cause abnormal wear of the brake pads. In this scenario, when performing historical fault diagnosis and / or real-time fault diagnosis, the method disclosed in the present invention can not only retrieve the historical faults of the vehicle, production batch-related faults and / or vehicle model-related faults, but also collect data on ambient temperature (e.g., above 40°C for a long time) and driving behavior (e.g., the driver frequently brakes suddenly). Combined with these dynamic data, the method disclosed in the present invention can determine that high temperature and sudden braking have caused overheating or increased wear of the braking system. This analysis can help identify potential faults in the braking system and suggest that the driver adjust his driving habits or check the braking system.

[0038] In some embodiments, the historical fault table Ta may include the license plate number, production batch, model, fault information, the number of fault occurrences, abnormal operation data, fault location, fault cause, and maintenance suggestion information corresponding to the fault information. The production batch related fault table Tb may include the production batch, fault information, the number of fault occurrences, abnormal performance, fault cause, and maintenance suggestion information. The vehicle model related fault table Tc may include the model, fault information, the number of fault occurrences, abnormal performance, fault cause, and maintenance suggestion information. In some instances, the cloud server may only store the historical fault table Ta in advance, and when a fault occurs, generate the production batch related fault table Tb and vehicle model related fault table Tc of the vehicle by counting the data in the historical fault table Ta. When generating the production batch fault table Tb and vehicle model fault table Tc, not only can they be generated based on the historical data of the vehicle, but also relevant data of other dimensions, such as environmental factors, vehicle usage, etc., can be introduced to further enrich the accuracy of the fault cause analysis. This makes the diagnosis process more intelligent and can adapt to different environments and vehicle usage scenarios. The data in the historical fault table Ta, the production batch related fault table Tb, and the vehicle model related fault table Tc can provide support for the AI ​​algorithm. For example, by analyzing the historical fault information in the historical fault table Ta, potential fault modes can be automatically identified and personalized maintenance suggestions can be generated. For example, by analyzing the data in the production batch-related fault table Tb, it can be identified that vehicles in a certain production batch are prone to a certain fault under specific circumstances, so other vehicles in the same batch can be warned to take maintenance measures in advance to reduce the probability of failure. For example, for certain production batches and specific models, by comparing the historical fault data of multiple vehicles, the fault performance of different vehicles in similar environments can be analyzed to generate a more general troubleshooting and maintenance strategy. This function can provide strong support for large-scale fleet management or product improvements of manufacturers.

[0039] In some embodiments, the first fault diagnosis result Res1 may describe the abnormal operation data, fault location and fault cause in detail. For example, the first fault diagnosis result Res1 of a vehicle may show that the abnormal operation data is "low battery charge signal", the fault location is "battery pack", and the fault cause may be attributed to "insufficient charging", "too long use time", "reduced battery capacity" or "low rated capacity". This structured result helps to quickly locate the problem and formulate corresponding maintenance plans. In some instances, the format of the first fault diagnosis result Res1 may adopt a standardized template. For example, the first fault diagnosis result Res1 may include sensor readings, timestamps, and occurrence frequencies. In some embodiments, the first maintenance suggestion information Sug1 may include the location of the faulty component, maintenance measures, etc.

[0040] In step S22, the vehicle can be subjected to real-time fault diagnosis according to the real-time operation data of the vehicle to generate a second fault diagnosis result Res2 and a second maintenance suggestion information Sug2. The second fault diagnosis result Res2 may include abnormal operation data, fault location, possible causes of the fault. The real-time fault diagnosis method disclosed in the present invention may be a fault diagnosis method based on a knowledge graph or a fault diagnosis method based on a fault diagnosis model, etc. According to the information in the second fault diagnosis result Res2, the second maintenance suggestion information Sug2 may be generated. The second maintenance suggestion information Sug2 may include a description of fault symptoms, an image of a faulty component, a description of a maintenance method, a maintenance video, and any combination thereof. A detailed description of the fault symptoms can help users or maintenance personnel accurately determine the degree of match between the fault and the observed performance. This description may not only include the text of common symptoms, but also add specific fault occurrence scenarios (such as performance during acceleration and braking) to enhance the practicality of the information. The image of the faulty component not only helps to quickly locate the faulty component, but also can be compared with the structural diagram of the vehicle to provide a more intuitive visual guide. The maintenance method description can guide the user or maintenance personnel to perform maintenance operations in text form. The maintenance video allows the user or maintenance personnel to repair the faulty component according to the teaching video.

[0041] In some embodiments, the real-time fault diagnosis of step S22 is not only based on the real-time data of the vehicle, but also can integrate data from other sources, such as road environment, climate conditions or driving behavior data, so as to further optimize the accuracy of fault diagnosis. In some instances, high-frequency faults or faults with serious impacts can be identified and handled with priority through comprehensive analysis of real-time data and historical data. Different types of faults can be weighted and scored to automatically adjust the diagnostic priority. This process can be implemented using a machine learning algorithm.

[0042] In some embodiments, if it is determined in the historical fault diagnosis of step S21 that a fault identical to a historical fault has occurred, the real-time fault diagnosis step of step S22 may also be skipped. This decision can improve the efficiency of fault detection and avoid repeated diagnostic work, thereby saving time and resources. This is particularly important for vehicle models that frequently have the same faults, and can effectively shorten the repair time. In some cases, it may also be decided whether to skip the real-time fault diagnosis based on the usage of the vehicle or user feedback. For example, if the user reports recent changes in driving behavior or environmental factors, real-time diagnosis may still be chosen to capture any potential new problems. When it is determined in the historical fault diagnosis of step S21 that no fault identical to a historical fault has occurred, the real-time fault diagnosis step of step S22 is also continued.

[0043] In step S23, a total fault diagnosis result may be generated. Depending on whether the step S22 of real-time fault diagnosis is performed, the total fault diagnosis result may include the first fault diagnosis result Res1 and / or the second fault diagnosis result Res2. In step S24, total maintenance suggestion information may be generated. Depending on whether the step S22 of real-time fault diagnosis is performed, the total maintenance suggestion information may include the first maintenance suggestion information Sug1 and / or the second maintenance suggestion information Sug2. In some instances, the method disclosed herein may remove duplicate or invalid data in the first fault diagnosis result Res1 and the second fault diagnosis result Res2 and the first maintenance suggestion information Sug1 and the second maintenance suggestion information Sug2 through techniques such as data cleaning and feature selection, thereby reducing redundancy and improving diagnostic efficiency. For example, a deduplication algorithm or similarity detection may be used to ensure that only unique fault diagnosis information and maintenance suggestion information are retained. In some instances, the method disclosed herein may integrate multiple fault diagnosis results or multiple maintenance suggestion information into a global view through multi-source data fusion. The fusion method may use weighted average, Bayesian reasoning, etc. to ensure the effective combination of different data sources, thereby generating more accurate fault diagnosis results and maintenance suggestion information. In some instances, the disclosed method can predict or discover potential faults by analyzing hidden patterns or associations in existing data through data mining and machine learning algorithms. This can be achieved through methods such as anomaly detection and cluster analysis, so that when generating fault diagnosis results, possible fault risks can be identified in advance to enhance preventive maintenance.

[0044] In order to facilitate different fault repairers to quickly find the required maintenance solutions, the method for generating vehicle maintenance suggestions disclosed in the present invention can provide different maintenance suggestions to the vehicle driver, maintenance personnel and manufacturer's development and design personnel according to the different fault repairers (such as the vehicle driver, maintenance personnel and manufacturer's development and design personnel). This personalization ensures that each user can quickly find the relevant solution, thereby improving the efficiency of fault handling. In some embodiments, the total maintenance suggestion information can be classified into operation guidance suggestions, hardware maintenance suggestions and software / hardware improvement suggestions using a maintenance suggestion classifier. The maintenance suggestion classifier can include decision trees, K-nearest neighbor algorithms, support vector machines, logistic regression and naive Bayes classifiers, etc. User feedback can be continuously collected, and the maintenance suggestion classifier can be dynamically updated based on the actual results during the maintenance process.

[0045] In step S3, a personalized maintenance plan for the vehicle can be generated and sent to the requester of the fault diagnosis request. The requester of the fault diagnosis request can be the user of the vehicle, the maintenance personnel of the vehicle, the manufacturer of the vehicle, or a third-party monitoring service provider. The personalized maintenance plan may include the overall fault diagnosis results and the overall maintenance suggestion information. In some instances, in order to ensure the pertinence and effectiveness of the personalized maintenance plan, the method disclosed in the present invention can formulate a personalized maintenance plan based on the historical maintenance records of the vehicle, the severity and type of the current fault, and dynamic data (e.g., driving habits, environmental factors, road conditions). The dynamic data can be obtained from the vehicle's sensor system or from an external data source (e.g., a public weather forecast system, a traffic dispatch system, and a navigation system).

[0046] The personalized maintenance plan can list the execution steps of each suggestion in detail to ensure that the user can clearly understand the required operations. The personalized maintenance plan can be sent in a variety of ways, such as email, mobile application notification or SMS, to ensure that the user can receive the information in real time. The requester can be allowed to provide feedback on the personalized maintenance plan so that future maintenance suggestions are more accurate. In some instances, the overall fault diagnosis results and the vehicle's operating guidance suggestions can be sent to the user of the vehicle. The vehicle's operating guidance suggestions may include suggestions for the driver, such as adjusting driving habits, regularly checking specific parts, etc. The overall fault diagnosis results and hardware maintenance suggestions can be sent to the vehicle's maintenance personnel. Hardware maintenance suggestions may include detailed hardware maintenance steps, such as replacing parts, calibrating sensors, etc. The overall fault diagnosis results and software / hardware improvement suggestions can be sent to the software developers of the vehicle manufacturer. Software / hardware improvement suggestions may include, for example, updating software versions, fixing specific code problems, etc.

[0047] In one example, when generating a personalized maintenance plan for an electric vehicle, if it is found that (1) the vehicle has multiple battery overheating cases in its historical maintenance records, (2) the current "battery temperature abnormality" fault is assessed to be of medium severity, (3) the driver frequently performs fast charging, and (4) the area where the vehicle is located is currently experiencing high temperature weather, then the personalized maintenance plan generated by the method of the present disclosure includes not only maintenance suggestions "check the coolant level in the cooling system" and "detect each cell of the battery pack to find damaged or failed cells" sent to the vehicle's maintenance personnel, but also vehicle operation guidance sent to the vehicle's driver. The vehicle operation guidance may include suggestions that the owner reduce the frequency of fast charging, avoid charging in high temperature environments, and regularly check the battery cooling system.

[0048] In one example, when generating a personalized maintenance plan for a vehicle, if it is found that (1) the vehicle's historical maintenance record has multiple brake pad replacements, (2) the current "insufficient brake pad thickness" fault is assessed to be of urgent severity, (3) the driver usually drives the vehicle in urban congestion, resulting in frequent emergency braking, and (4) the road surface in the area where the vehicle is currently located is icy, then the personalized maintenance plan generated by the method of the present disclosure includes not only maintenance suggestions "replace brake pads" and "check brake fluid" sent to the vehicle's maintenance personnel, but also vehicle operation instructions sent to the vehicle's driver and software / hardware improvement suggestions sent to the vehicle manufacturer. The vehicle operation instructions may include suggesting that the owner reduce the habit of frequent emergency braking in the next few months. Software / hardware improvement suggestions may include suggesting that the manufacturer upgrade the software to optimize the electronic control strategy and early warning function of the braking system.

[0049] The present disclosure also provides a system including one or more computer processors and a computer readable memory. The computer readable memory may include machine executable code, which, when executed by the one or more computer processors, implements the method for generating maintenance suggestions for a vehicle of the present disclosure.

[0050] Although preferred embodiments of the present invention have been shown and described herein, it will be apparent to those skilled in the art that these embodiments are provided by way of example only. Other variations, changes and substitutions may be conceived by those skilled in the art without departing from the present invention.

Claims

1. A method for generating maintenance recommendations for a vehicle, the method comprising: (S1) in response to a fault diagnosis request, collecting real-time operating data of the vehicle and sending the collected real-time operating data to a cloud server; (S2) the cloud server performs fault diagnosis on the vehicle based on the real-time operation data to generate a total fault diagnosis result and total maintenance suggestion information, and the step S2 includes: (S21) performing historical fault diagnosis on the vehicle based at least in part on the real-time operation data to generate a first fault diagnosis result Res1 and first maintenance suggestion information Sug1, and / or (S22) performing real-time fault diagnosis on the vehicle based at least in part on the real-time operation data, To generate a second fault diagnosis result Res2 and a second maintenance suggestion information Sug2; (S23) generating the total fault diagnosis result, wherein the total fault diagnosis result includes the first fault diagnosis result Res1 and / or the second fault diagnosis result Res2; and (S24) generating the total maintenance suggestion information, wherein the total maintenance suggestion information includes the first maintenance suggestion information Sug1 and / or the second maintenance suggestion information Sug2, wherein the total maintenance suggestion information is classified into operation guidance suggestions, hardware maintenance suggestions, and software / hardware improvement suggestions by using a maintenance suggestion classifier; and (S3) The cloud server generates a personalized maintenance plan for the vehicle, wherein the personalized maintenance plan includes the overall fault diagnosis result and the overall maintenance suggestion information, and sends the personalized maintenance plan to the requester of the fault diagnosis request.

2. The method as claimed in claim 1, wherein the historical fault table Ta, the production batch related fault table Tb and the model related fault table Tc of the vehicle are stored and updated in the cloud server, and In step S21, the historical fault diagnosis is performed in the order of the historical fault information of the vehicle, the historical fault information of the production batch of the vehicle, and the historical fault information of the model of the vehicle.

3. The method as claimed in claim 2, wherein in the step S21, the historical fault diagnosis is performed for the faults whose number of occurrences or frequencies in the historical fault table Ta, the production batch related fault table Tb and the vehicle model related fault table Tc exceed a threshold.

4. The method as claimed in claim 2, wherein in the steps S21 and / or S22, the historical fault diagnosis and / or the real-time fault diagnosis are further performed on the vehicle based on dynamic data, and the dynamic data includes driving behavior data and / or environmental data.

5. The method as claimed in claim 1, wherein in step S3, the overall fault diagnosis result and the operation guidance suggestion are sent to the user of the vehicle, the overall fault diagnosis result and the hardware maintenance suggestion are sent to the maintenance personnel of the vehicle, and the overall fault diagnosis result and the software / hardware improvement suggestion are sent to the manufacturer of the vehicle.

6. The method as claimed in claim 2, wherein the production batch related fault table Tb and the vehicle type related fault table Tc are generated by statistically analyzing the data in the historical fault table Ta.

7. The method as claimed in claim 3, wherein a machine learning algorithm is used to adjust the threshold in real time according to changes in the operating environment of the vehicle and / or the actual operating conditions of the vehicle.

8. The method of claim 1, wherein if it is determined in the historical fault diagnosis of step S21 that a fault identical to a historical fault has occurred, the real-time fault diagnosis of step S22 is not performed.

9. The method as claimed in claim 1, wherein step S3 comprises formulating a personalized maintenance plan based on the overall fault diagnosis result of the vehicle, the severity and type of the current fault, and dynamic data, wherein the dynamic data comprises driving behavior data and / or environmental data.

10. A system comprising one or more computer processors and a computer readable memory, the computer readable memory comprising machine executable code, which when executed by the one or more computer processors implements the method for generating maintenance suggestions for a vehicle of any one of claims 1-9.

Images