Shared vehicle suspected fault distinguishing method fusing multi-source data and related equipment

Abstract

The invention provides a shared vehicle suspected fault discrimination method fusing multi-source data and related equipment, and the method comprises the steps: obtaining the multi-source operation data of a to-be-discriminated vehicle in a preset data time period, and the multi-source operation data comprises order conversion data, riding order data, user feedback data, vehicle track data and battery data; calculating at least two abnormal indexes based on the multi-source operation data, wherein the abnormal indexes comprise a no-order abnormal rate, a short-time riding order rate, a fault feedback rate, an abnormal trajectory rate, a battery efficiency abnormal rate and a riding power-off rate; and according to the at least two obtained abnormal indexes, performing suspected fault scoring on the to-be-discriminated vehicle through a preset scoring model, comparing the suspected fault score with a preset fault scoring threshold, and discriminating a suspected fault vehicle. According to the distinguishing method, the multi-source operation data of the vehicle serves as the basis, fault distinguishing is carried out in cooperation with diversified index combinations, scene coverage is more comprehensive, and accuracy is higher.

Description

Technical Field

[0001] This invention relates to the field of shared vehicle technology, and in particular to a method and related equipment for identifying suspected faults in shared vehicles by integrating multi-source data. Background Technology

[0002] Shared electric bicycles are an important tool for short-distance travel, and timely fault diagnosis and handling are crucial to improving the operational efficiency of shared vehicles. Currently, the industry mainly uses two types of fault diagnosis methods: manual maintenance marking, where inspectors check and mark faulty vehicles on-site; and user-initiated feedback mechanisms, which collect user reports through channels such as app-based fault reporting and feedback. Manual maintenance marking suffers from slow response times, high labor costs, and limited coverage; while user-initiated feedback mechanisms have high false alarm rates, incomplete scenario coverage, and fragmented data. It is evident that existing methods still have certain shortcomings in terms of cost, scenario coverage, and diagnostic accuracy. There is an urgent need to develop a low-cost, all-scenario, and more reliable automated fault diagnosis solution.

[0003] Therefore, existing technologies still need to be improved and developed. Summary of the Invention

[0004] This invention provides a method and related equipment for identifying suspected faults in shared vehicles by integrating multi-source data. The main purpose of this invention is to solve the technical problems mentioned in the background art of the prior art.

[0005] The first aspect of this invention provides a method for identifying suspected faults in shared vehicles by fusing multi-source data, comprising: Obtain multi-source operational data of the vehicle to be identified within a preset data period. The multi-source operational data includes order conversion data, riding order data, user feedback data, vehicle trajectory data, and battery data. Based on the multi-source operational data, at least two abnormal indicators are calculated, including the no-order abnormality rate, short-time riding order rate, fault feedback rate, abnormal trajectory rate, battery efficiency abnormality rate, and riding power outage rate. Based on at least two of the calculated abnormal indicators, the vehicle to be identified is scored for suspected faults using a preset scoring model. The obtained suspected fault score is compared with a preset fault score threshold. When the suspected fault score is higher than the fault score threshold, the vehicle to be identified is identified as a suspected fault vehicle.

[0006] In an optional embodiment of the first aspect of the present invention, obtaining multi-source operational data of the vehicle to be identified within a preset data period includes: Obtain the maintenance records of the vehicle to be identified; The time point of the most recent maintenance of the vehicle to be identified is traced back in the maintenance record, and the time point is used as the starting time for acquiring multi-source operational data; Determine whether the vehicle to be judged meets the preset number of scans and number of ride orders within the preset data period from the start time to the current time; If the vehicle to be identified meets the preset number of scans and the number of ride orders, then the preset data period is determined to be valid, and multi-source operation data of the vehicle to be identified within the preset data period is obtained.

[0007] In an optional embodiment of the first aspect of the present invention, the calculation of at least two abnormal indicators based on the multi-source operational data includes: For the no-order anomaly rate, obtain the order conversion data of users who scanned the code but did not generate a ride order; In the order conversion data, records of loss due to user reasons and non-vehicle malfunction reasons are excluded. User reasons include user account arrears and user failure to complete real-name or facial authentication. Non-vehicle malfunction reasons include vehicle low battery and vehicle outside the operating area. The remaining order conversion data after excluding the lost records are counted, and the proportion of the remaining order conversion data to the total number of scans is calculated to obtain the no-order anomaly rate.

[0008] In an optional embodiment of the first aspect of the present invention, the calculation of at least two abnormal indicators based on the multi-source operational data includes: For the abnormal trajectory rate, based on the status field in the vehicle trajectory data, the trajectory points in the vehicle trajectory data are divided into stationary trajectory points and moving trajectory points. For the stationary trajectory point, if the spatial distance between the stationary trajectory point and the previous adjacent trajectory point is greater than a first preset threshold, then the stationary trajectory point is determined to be a stationary abnormal point. For the motion state trajectory point, the running speed is calculated based on the spatial distance and time difference between the motion state trajectory point and the previous adjacent trajectory point. If the running speed is greater than the second preset threshold, the motion state trajectory point is determined to be a motion anomaly point. The abnormal trajectory rate is obtained by statistically analyzing the proportion of stationary abnormal points and moving abnormal points to the total number of trajectory points.

[0009] In an optional embodiment of the first aspect of the present invention, the calculation of at least two abnormal indicators based on the multi-source operational data includes: For the battery efficiency anomaly rate, the battery efficiency of all vehicles in the region in the previous battery swap cycle is statistically analyzed, and the battery efficiency percentile threshold is obtained from the statistical results according to a preset percentage. After this battery swap, the current battery percentage and the total mileage after the swap will be collected at fixed time intervals during the ride. When it is detected that the difference between the initial battery percentage after this battery swap and the current battery percentage is greater than or equal to 50%, the process of calculating the current battery efficiency based on the total riding mileage after the battery swap, the current battery percentage, and the initial battery percentage is initiated for this battery swap cycle. After each calculation, if the current battery efficiency is less than the battery efficiency quantile threshold, the battery efficiency is determined to be abnormal. The battery efficiency anomaly rate is obtained by calculating the ratio of the number of times the battery efficiency was determined to be abnormal to the total number of determinations during this battery swapping cycle.

[0010] In an optional embodiment of the first aspect of the present invention, the calculation of at least two abnormal indicators based on the multi-source operational data includes: For the aforementioned power outage rate during riding, battery power and battery voltage data of the vehicle are collected during riding. When the vehicle's battery level drops from a non-zero value to zero at an adjacent data collection time point, and the battery voltage at the current time point is less than or equal to the zero-battery voltage quantile threshold calculated based on historical vehicle data in the region, it is marked as an abnormal power outage of the vehicle. The number of abnormal power outages of vehicles within the battery swapping cycle is counted, and the proportion of the abnormal power outages to the total number of riding orders is calculated to obtain the riding power outage rate.

[0011] In an optional embodiment of the first aspect of the present invention, the step of scoring the suspected fault of the vehicle to be identified based on at least two calculated abnormal indicators using a preset scoring model includes: For each of the at least two abnormal indicators, obtain the preset weight coefficient of the abnormal indicator; A suspected fault score is obtained by weighting and summing at least two of the abnormal indicators according to the corresponding preset weight coefficients.

[0012] A second aspect of the present invention provides a device for identifying suspected faults in shared vehicles by fusing multi-source data, the device comprising: The data acquisition module is used to acquire multi-source operational data of the vehicle to be judged within a preset data period. The multi-source operational data includes order conversion data, riding order data, user feedback data, vehicle trajectory data, and battery data. An abnormal indicator calculation module is used to calculate at least two abnormal indicators based on the multi-source operation data. The abnormal indicators include no-order abnormality rate, short-time riding order rate, fault feedback rate, abnormal trajectory rate, battery efficiency abnormality rate, and riding power outage rate. The suspected fault scoring module is used to score the suspected fault of the vehicle to be identified based on at least two of the calculated abnormal indicators and through a preset scoring model. The suspected fault determination module is used to compare the obtained suspected fault score with a preset fault score threshold. When the suspected fault score is higher than the fault score threshold, the vehicle to be judged is judged as a suspected fault vehicle.

[0013] A third aspect of the present invention provides a shared vehicle suspected fault identification device that integrates multi-source data. The shared vehicle suspected fault identification device that integrates multi-source data includes: a memory and at least one processor. The memory stores instructions, and the memory and the at least one processor are interconnected by a line. The at least one processor invokes the instructions in the memory to cause the shared vehicle suspected fault detection device that integrates multi-source data to perform the shared vehicle suspected fault detection method that integrates multi-source data as described in any one of the first aspects of the present invention.

[0014] A fourth aspect of the present invention provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the method for identifying suspected faults in shared vehicles by fusing multi-source data as described in any one of the first aspects of the present invention.

[0015] Beneficial Effects: This invention provides a method and related equipment for identifying suspected faults in shared vehicles by integrating multi-source data. The method includes acquiring multi-source operational data of the vehicle to be identified within a preset data period. The multi-source operational data includes order conversion data, riding order data, user feedback data, vehicle trajectory data, and battery data. Based on the multi-source operational data, at least two abnormal indicators are calculated. These abnormal indicators include no-order abnormality rate, short-term riding order rate, fault feedback rate, abnormal trajectory rate, battery efficiency abnormality rate, and riding power outage rate. Based on the obtained at least two abnormal indicators, the vehicle to be identified is scored for suspected faults using a preset scoring model, and the scores are compared with a preset fault scoring threshold to identify the suspected faulty vehicle. This invention's identification method is based on multi-source operational data of the vehicle and combines diverse indicator combinations for fault identification, resulting in more comprehensive scenario coverage and higher accuracy. Attached Figure Description Figure 1 This is a schematic diagram illustrating an embodiment of the main steps of a method for identifying suspected faults in shared vehicles that integrates multi-source data according to the present invention; Figure 2 This is an exemplary cycling time distribution map of the present invention; Figure 3 This is an exemplary cycling mileage distribution map of the present invention; Figure 4 This is an exemplary static state-adjacent trajectory spatial distance distribution diagram of the present invention; Figure 5 This is an exemplary motion state-trajectory velocity distribution diagram of the present invention; Figure 6 This is an exemplary distribution map of expected full-charge riding range according to the present invention; Figure 7 This is an exemplary voltage distribution diagram at 0% charge according to the present invention; Figure 8 This is a schematic diagram of an embodiment of a shared vehicle suspected fault identification device that integrates multi-source data according to the present invention; Figure 9 This is a schematic diagram of an embodiment of a shared vehicle suspected fault identification device that integrates multi-source data according to the present invention. Detailed Implementation

[0016] The terms "first," "second," "third," "fourth," etc. (if present) in the specification, claims, and accompanying drawings of this invention are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments described herein can be implemented in orders other than those illustrated or described herein. Furthermore, the terms "comprising" or "having" and any variations thereof are intended to cover a non-exclusive inclusion; for example, a process, method, system, product, or apparatus that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or apparatus.

[0017] For ease of understanding, the specific process of the embodiments of the present invention is described below. Please refer to [link / reference]. Figure 1 The first aspect of this invention provides a method for identifying suspected faults in shared vehicles by fusing multi-source data, comprising: S100: Obtain multi-source operational data of the vehicles to be identified within a preset data period. This multi-source operational data includes order conversion data, riding order data, user feedback data, vehicle trajectory data, and battery data. Before step S100, vehicles to be identified need to be filtered out in the shared vehicle operation system. Specifically, this can be done by filtering out vehicles that are not yet deployed, have been marked as faulty, or are in special abnormal states such as being in storage, impounded, lost, or offline, based on the vehicle's real-time status table. The remaining vehicles are the vehicles to be identified. The core of this invention's technical solution lies in using multi-source data such as "riding data (mileage, duration) + order conversion data (records of loss at each stage such as scanning, verification, unlocking, and riding) + trajectory data (threshold-based judgment of static / dynamic states) + user feedback data (feedback, fault reporting, fee appeals) + battery data (battery efficiency + power outage inspection) + vehicle maintenance data (storage, maintenance, code replacement)" to construct a full-process fault screening system, overcoming the limitations of traditional fault identification methods that rely too heavily on manual labor or a single data source. Construct a fault diagnosis model that covers the entire scenario of "with orders + without orders".

[0018] To avoid interference from invalid data and maintain sensitive fault tracking, a reasonable and effective data time period range needs to be selected. Based on vehicle maintenance records (including vehicle entry, repair, or change of vehicle number), for each vehicle, data is retrieved from the current time backward to the most recent maintenance time, until the user's scan count reaches 10 times (consecutive scans from the same user within 1 hour will not be counted repeatedly), or the number of ride orders reaches 5 times. If the data period exceeds 30 days and the scan and ride order conditions are still not met, the vehicle is considered to have insufficient data and cannot be used to identify suspected faults.

[0019] Specifically, in an optional embodiment of step S100 of the present invention, obtaining multi-source operational data of the vehicle to be judged within a preset data period includes: obtaining the maintenance record of the vehicle to be judged; tracing back the time node of the most recent maintenance of the vehicle to be judged from the maintenance record, and using the time node as the starting time for obtaining multi-source operational data; determining whether the vehicle to be judged meets the preset number of scans and riding orders within the preset data period from the starting time to the current time; if the vehicle to be judged meets the preset number of scans and riding orders, then the preset data period is determined to be valid, and multi-source operational data of the vehicle to be judged within the preset data period is obtained. The present invention automatically pulls data from "the most recent maintenance to the current time" based on the vehicle maintenance record, for example, using "10 scans + 5 rides" or "30 days" as dual termination conditions to ensure the validity of the data sample size; at the same time, it improves the purity of abnormal data without orders by automatically eliminating non-fault factors (filtering out interference such as low battery, out-of-area, and user authentication failure).

[0020] S200. Based on the multi-source operational data, at least two abnormal indicators are calculated. The abnormal indicators include no-order abnormality rate, short-term riding order rate, fault feedback rate, abnormal trajectory rate, battery efficiency abnormality rate, and riding power outage rate. Specifically, another core aspect of the technical solution of this invention is to dynamically calculate six indicators—no-order abnormality rate, short-term riding order rate, fault feedback rate, abnormal trajectory rate, battery efficiency abnormality rate, and riding power outage rate—and combine them with custom weights and thresholds to achieve a comprehensive score for suspected faults, balancing the accuracy of judgment with the number of faults covered, and adapting to the operational needs of different cities.

[0021] In an optional embodiment of step S200 of the present invention, the calculation of at least two abnormal indicators based on the multi-source operational data includes: For the no-order anomaly rate, order conversion data of users who scanned the code but did not generate a ride order is obtained; in the order conversion data, drop-off records caused by user reasons and non-vehicle malfunction reasons are excluded. User reasons include user account arrears and user failure to complete real-name or facial authentication. Non-vehicle malfunction reasons include vehicle low battery and vehicle outside the operating area; the remaining order conversion data after excluding the drop-off records are counted, and the proportion of the remaining order conversion data to the total number of scans is calculated to obtain the no-order anomaly rate.

[0022] Specifically, in this invention, order conversion data can be categorized into stages such as scanning, vehicle verification, unlocking, facial recognition, and riding according to the business process. Records where users scanned the code but ultimately failed to convert it into a ride order are filtered out, and the stage at which the user ultimately dropped off is examined. After filtering out reasons such as low battery, exceeding the designated area, vehicle being occupied, user arrears, and user failure to pass facial recognition or real-name authentication, user drop-off due to other reasons is marked as a no-order anomaly for that vehicle. Furthermore, a maximum of one no-order anomaly is recorded per hour for the same user and the same vehicle.

[0023] Regarding the short-term cycling order rate, this invention will statistically analyze the cycling duration and mileage of cycling orders in the region over the past 30 days, and the results are as follows: Figure 2 and Figure 3 As shown, based on comprehensive data statistics and business experience, riding orders with a riding time of less than 2 minutes and a distance of less than 100 meters are classified as short-duration riding orders. The short-duration riding order rate is then obtained by dividing the number of short-duration riding orders by the total number of riding orders. Regarding the fault feedback rate, this invention calculates the proportion of user fault feedback received within 30 days to the total number of riding orders.

[0024] For the abnormal trajectory rate, based on the status field in the vehicle trajectory data, the trajectory points in the vehicle trajectory data are divided into stationary trajectory points and moving trajectory points. For a stationary trajectory point, if the spatial distance between the stationary trajectory point and its previous adjacent trajectory point is greater than a first preset threshold, the stationary trajectory point is determined to be a stationary abnormal point. For a moving trajectory point, the running speed is calculated based on the spatial distance and time difference between the moving trajectory point and its previous adjacent trajectory point. If the running speed is greater than a second preset threshold, the moving trajectory point is determined to be a moving abnormal point. The abnormal trajectory rate is obtained by statistically analyzing the proportion of stationary abnormal points and moving abnormal points to the total number of trajectory points.

[0025] Specifically, in this invention, the logic for identifying abnormal vehicle trajectories is as follows: based on the vehicle status data field in the trajectory data, it can be divided into two categories: moving state and stationary state, and analyzed separately. For example... Figure 4 As shown, the spatial distance between adjacent trajectory points in a stationary state. Highly concentrated around 0, with a 99.7% quantile. This is the threshold for abnormal trajectories. If adjacent trajectories are spatially offset... ≥ If the trajectory is abnormal while stationary, it is recorded as an anomaly. In the stationary state, the 99.7% quantile of the spatial distance between adjacent trajectory points is used as a threshold to detect positioning drift; in the moving state, the velocities of adjacent trajectory points are calculated, and the 99.7% quantile of the velocity is used as a threshold to identify velocity jump anomalies. The dual-state model addresses the technical blind spot of traditional trajectory analysis that ignores "stationary positioning faults."

[0026] For motion trajectory data, calculate the spatial distance between adjacent trajectory points. With time difference And calculate the vehicle's speed between these two points. :

[0027] Vehicle speed distribution under motion state as follows Figure 5 As shown, a 99.7% quantile was set. This is the threshold for abnormal trajectories. If the speed of adjacent trajectories... ≥ If it is, it will be recorded as an abnormal trajectory during motion.

[0028] In summary, the offset of adjacent trajectory points in a static state ≥ or vehicle speed in motion ≥ If so, it will be judged as having an abnormal trajectory point.

[0029] For the battery efficiency anomaly rate, the battery efficiency of all vehicles in the region during the previous battery swap cycle is first statistically analyzed, and a battery efficiency quantile threshold is obtained from the statistical results according to a preset percentage. After the current battery swap, the current battery percentage and the total mileage after the swap are collected at fixed time intervals during the ride. When it is detected that the difference between the initial battery percentage after the current battery swap and the current battery percentage is greater than or equal to 50%, the process of calculating the current battery efficiency based on the total mileage after the swap, the current battery percentage, and the initial battery percentage for the current battery swap cycle is initiated. After each calculation, if the current battery efficiency is less than the battery efficiency quantile threshold, the battery efficiency is determined to be abnormal. The ratio of the number of battery efficiency anomaly determinations to the total number of determinations in the current battery swap cycle is calculated to obtain the battery efficiency anomaly rate.

[0030] In this invention, to identify abnormal batteries, it is necessary to first statistically analyze the battery status of all vehicles within the area. A battery swap cycle is defined as the time from one battery swap to the next. If the initial battery percentage of a certain swap cycle is... The percentage of battery charge at the end of the battery swapping cycle is The riding distance during the battery swapping cycle is Then the expected full-charge riding range for this battery swapping cycle is... (i.e., the expected full-charge battery efficiency) is:

[0031] The expected range of riding on a full charge in a certain area is as follows: Figure 6 As shown, it typically exhibits a good approximate normal distribution. Calculate the 3% quantile for the fully charged riding mileage in this region. (i.e., battery efficiency quantile threshold), if a vehicle's current battery percentage is... The percentage of battery charge after the last battery swap was: The cycling distance is If and satisfy:

[0032] If the battery efficiency of the vehicle is abnormally low, this invention proposes a formula for calculating battery efficiency within a battery swapping cycle. Based on the 3% percentile of the full-charge riding mileage in a region, it quantifies the conditions for determining abnormally low battery efficiency, making the judgment more accurate.

[0033] For the aforementioned power outage rate, battery power and voltage data of the vehicle are collected during the riding process. When the vehicle power drops from a non-zero value to zero at an adjacent data collection time point, and the battery voltage at the current time point is less than or equal to the zero power voltage quantile threshold calculated based on the historical data of vehicles in the region, it is marked as an abnormal power outage. The number of abnormal power outages of the vehicle within the battery swapping cycle is counted, and the proportion of the number of abnormal power outages to the total number of riding orders is calculated to obtain the riding power outage rate.

[0034] This invention primarily uses battery power data to detect abnormal power outages during vehicle operation. Since malfunctions in the battery sensor, data acquisition module, data communication, or server parsing module can all lead to abnormal individual battery power data, a simple jump to 0% battery level cannot be used to determine a power outage. A comprehensive assessment must be made in conjunction with battery voltage. The battery voltage distribution in a certain area at 0% battery power is shown below. Figure 7 As shown, the 3% quantile of the voltage is calculated. If, during the riding process, the battery level of a vehicle was >0% at the previous data point, and the current battery level has dropped to 0% and the voltage is ≤0%, then... If the voltage drops to 0% and the voltage is less than or equal to the 3% percentile of the area's 0% voltage level, then it is marked as an abnormal power outage. This invention combines the dual conditions of "0% power level change + ≤ 0% voltage level in the area" to achieve accurate identification of abnormal power outages and avoid misjudgments caused by a single abnormal power level data.

[0035] In this invention, the suspected fault identification system updates the data weekly and recalculates. , , , Data metrics, such as regional data quantiles (e.g., stationary trajectory offset thresholds), are updated weekly. Motion speed threshold Fully charged riding range threshold Power outage voltage threshold This enables adaptability to different operating areas and time periods, improving cross-scenario judgment stability. Every hour, for all vehicles awaiting suspected fault detection, the following data results are updated: No-Order Anomaly Rate = Percentage of No-Order Anomalies to Total Scanned Codes; Short-Time Ride Order Rate = Percentage of Short-Time Ride Orders to Total Ride Orders; Fault Feedback Rate = Percentage of Fault Feedback to Total Ride Orders; Abnormal Trajectory Rate = Percentage of Abnormal Trajectory Points to Total Trajectory Points; Battery Efficiency Anomaly Rate = Percentage of Battery Efficiency within the Current Battery Swapping Cycle. Initially, the percentage of times the battery efficiency assessment result is abnormal; the riding power outage rate is the percentage of power outages within the current battery swapping cycle out of the total number of riding orders.

[0036] S300. Based on at least two of the calculated abnormal indicators, a suspected fault score is assigned to the vehicle to be identified using a preset scoring model. In this invention, the weights of various faults (to highlight the importance of a certain fault) and fault scoring thresholds can be preset according to the requirements of the regional operations managers (a higher threshold can improve the accuracy of suspected fault identification, but reduce the number of identifications). The weighted average of the quantiles of each vehicle's indicators within the region is recorded as the suspected fault score for that vehicle. Vehicles with suspected fault scores higher than the fault scoring threshold are considered suspected faulty vehicles.

[0037] In an optional embodiment of step S300 of the present invention, the step of scoring the suspected fault of the vehicle to be judged based on the calculated at least two abnormal indicators through a preset scoring model includes: for each of the at least two abnormal indicators, obtaining a preset weight coefficient for the abnormal indicator; and performing a weighted summation of the at least two abnormal indicators corresponding to the preset weight coefficients to obtain a suspected fault score.

[0038] S400. The obtained suspected fault score is compared with a preset fault score threshold. When the suspected fault score is higher than the fault score threshold, the vehicle to be judged is judged as a suspected fault vehicle.

[0039] In general, under the traditional manual inspection and user feedback model, the fault discovery cycle is usually 1 to 3 days. This invention, however, shortens the suspected fault identification cycle to within 24 hours by analyzing data in real time, enabling rapid fault location and interception. It integrates multi-dimensional information such as riding order data, trajectory data, battery data, and user feedback data. By constructing a static + dynamic dual-threshold model for trajectory and a battery efficiency-voltage joint verification mechanism, it achieves fault identification functions for modules such as the central control unit, communication, positioning, and battery. It comprehensively considers various fault types, quantifies the identification results, and improves the accuracy of identification while reducing the false positive rate. Addressing the issue that existing technologies cannot cover "no-order scenarios such as vehicle unlocking failure and central control unit malfunction," it integrates lost data from the entire process of scanning, verification, unlocking, and riding, combined with a deduplication logic of "user ID + vehicle ID + 1-hour time window," to achieve accurate screening of no-order faults. By combining dynamic calculation of battery efficiency and abnormal power outage detection within the battery swap cycle, the system can identify inefficient batteries and vehicles with abnormal power outages, reducing the number of invalid battery swaps and lowering maintenance costs. At the same time, it can prevent users from experiencing riding interruptions due to battery problems, thus improving the user experience.

[0040] See Figure 8 The second aspect of the present invention provides a device for identifying suspected faults in shared vehicles by integrating multi-source data, the device comprising: The data acquisition module 10 is used to acquire multi-source operational data of the vehicle to be judged within a preset data period. The multi-source operational data includes order conversion data, riding order data, user feedback data, vehicle trajectory data, and battery data. The abnormal indicator calculation module 20 is used to calculate at least two abnormal indicators based on the multi-source operation data. The abnormal indicators include no-order abnormality rate, short-time riding order rate, fault feedback rate, abnormal trajectory rate, battery efficiency abnormality rate, and riding power failure rate. The suspected fault scoring module 30 is used to score the suspected fault of the vehicle to be judged based on at least two of the calculated abnormal indicators and through a preset scoring model. The suspected fault determination module 40 is used to compare the obtained suspected fault score with a preset fault score threshold. When the suspected fault score is higher than the fault score threshold, the vehicle to be judged is judged as a suspected fault vehicle.

[0041] In an optional embodiment of the second aspect of the present invention, the data acquisition module includes: The maintenance record acquisition unit is used to acquire the maintenance records of the vehicle to be identified. The start time determination unit is used to trace back the time node of the most recent maintenance of the vehicle to be judged in the maintenance record, and use the time node as the start time for acquiring multi-source operation data; The effective judgment unit is used to determine whether the vehicle to be judged meets the preset number of scans and number of ride orders within the preset data period from the start time to the current time. The multi-source data acquisition unit is used to determine that the preset data period is valid if the vehicle to be identified meets the preset number of scans and the number of riding orders, and to acquire the multi-source operation data of the vehicle to be identified within the preset data period.

[0042] In an optional embodiment of the second aspect of the present invention, the anomaly index calculation module includes: The no-order anomaly rate processing unit is used to obtain order conversion data for users who scanned the code but did not generate a ride order, in order conversion data; excluding drop-off records caused by user reasons and non-vehicle malfunctions, where user reasons include user account arrears and user failure to complete real-name or facial authentication, and non-vehicle malfunctions include vehicle low battery and vehicle being outside the operating area; counting the remaining order conversion data after excluding the drop-off records, and calculating the proportion of the remaining order conversion data to the total number of scans, to obtain the no-order anomaly rate.

[0043] In an optional embodiment of the second aspect of the present invention, the anomaly index calculation module includes: An abnormal trajectory rate processing unit is used to, based on the status field in the vehicle trajectory data, divide the trajectory points in the vehicle trajectory data into stationary trajectory points and moving trajectory points; for the stationary trajectory points, if the spatial distance between the stationary trajectory point and its previous adjacent trajectory point is greater than a first preset threshold, then the stationary trajectory point is determined to be a stationary abnormal point; for the moving trajectory points, the running speed is calculated based on the spatial distance and time difference between the moving trajectory point and its previous adjacent trajectory point, and if the running speed is greater than a second preset threshold, then the moving trajectory point is determined to be a moving abnormal point; the abnormal trajectory rate is obtained by calculating the ratio of the number of stationary abnormal points and the number of moving abnormal points to the total number of trajectory points.

[0044] In an optional embodiment of the second aspect of the present invention, the anomaly index calculation module includes: The battery efficiency anomaly rate processing unit is used to first statistically analyze the battery efficiency of all vehicles in the region during the previous battery swap cycle, and obtain the battery efficiency percentile threshold from the statistical results according to a preset percentage. After the current battery swap, the current battery percentage and the total riding distance after the swap are collected at fixed time intervals during the ride. When it is detected that the difference between the initial battery percentage after the current battery swap and the current battery percentage is greater than or equal to 50%, the process of calculating the current battery efficiency based on the total riding distance after the swap, the current battery percentage, and the initial battery percentage is initiated for the current battery swap cycle. After each calculation, if the current battery efficiency is less than the battery efficiency percentile threshold, the battery efficiency is determined to be abnormal. The ratio of the number of battery efficiency anomaly determinations to the total number of determinations in the current battery swap cycle is calculated to obtain the battery efficiency anomaly rate.

[0045] In an optional embodiment of the second aspect of the present invention, the anomaly index calculation module includes: The cycling power outage rate processing unit is used to collect battery power data and battery voltage data of the vehicle during the cycling process for the cycling power outage rate; when the vehicle power drops from a non-zero value to zero at an adjacent data collection time point, and the battery voltage at the current time point is less than or equal to the zero power voltage quantile threshold calculated based on the historical data of vehicles in the region, it is marked as an abnormal power outage of the vehicle; the number of abnormal power outages of the vehicle within the battery swapping cycle is counted, and the proportion of the number of abnormal power outages to the total number of cycling orders is calculated to obtain the cycling power outage rate.

[0046] In an optional embodiment of the second aspect of the present invention, the suspected fault scoring module includes: The weight coefficient acquisition unit is used to acquire a preset weight coefficient for each of the at least two abnormal indicators. The indicator weighting calculation unit is used to perform a weighted summation of at least two of the abnormal indicators with corresponding preset weight coefficients to obtain a suspected fault score.

[0047] Figure 9 This is a schematic diagram of a shared vehicle suspected fault detection device that integrates multi-source data according to an embodiment of the present invention. This device can vary significantly due to differences in configuration or performance, and may include one or more processors 50 (central processing units, CPUs) (e.g., one or more processors) and a memory 60, and one or more storage media 70 (e.g., one or more mass storage devices) for storing applications or data. The memory and storage media can be temporary or persistent storage. The program stored in the storage media may include one or more modules (not shown in the diagram), each module including a series of instruction operations on the shared vehicle suspected fault detection device integrating multi-source data. Furthermore, the processor may be configured to communicate with the storage media to execute the series of instruction operations in the storage media on the shared vehicle suspected fault detection device integrating multi-source data.

[0048] The shared vehicle suspected fault identification device of the present invention, which integrates multi-source data, may further include one or more power supplies 80, one or more wired or wireless network interfaces 90, one or more input / output interfaces 100, and / or one or more operating systems, such as Windows Server, Mac OS X, Unix, Linux, FreeBSD, etc. Those skilled in the art will understand that... Figure 9 The illustrated structure of the shared vehicle suspected fault detection device that integrates multi-source data does not constitute a limitation on the shared vehicle suspected fault detection device that integrates multi-source data. It may include more or fewer components than illustrated, or combine certain components, or have different component arrangements.

[0049] The present invention also provides a computer-readable storage medium, which can be a non-volatile computer-readable storage medium or a volatile computer-readable storage medium, wherein the computer-readable storage medium stores instructions that, when the instructions are executed on a computer, cause the computer to perform the steps of the method for identifying suspected faults in shared vehicles by fusing multi-source data.

[0050] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the specific working process of the system or system / unit described above can be referred to the corresponding process in the foregoing method embodiments, and will not be repeated here.

[0051] If the integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of the present invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0052] The above-described embodiments are only used to illustrate the technical solutions of the present invention, and are not intended to limit it. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for identifying suspected faults in shared vehicles by integrating multi-source data, characterized in that, include: Obtain multi-source operational data of the vehicle to be identified within a preset data period. The multi-source operational data includes order conversion data, riding order data, user feedback data, vehicle trajectory data, and battery data. Based on the multi-source operational data, at least two abnormal indicators are calculated, including the no-order abnormality rate, short-time riding order rate, fault feedback rate, abnormal trajectory rate, battery efficiency abnormality rate, and riding power outage rate. Based on at least two of the calculated abnormal indicators, the vehicle to be identified is scored for suspected faults using a preset scoring model. The obtained suspected fault score is compared with a preset fault score threshold. When the suspected fault score is higher than the fault score threshold, the vehicle to be identified is identified as a suspected fault vehicle.

2. The method for identifying suspected faults in shared vehicles by integrating multi-source data according to claim 1, characterized in that, The acquisition of multi-source operational data of the vehicle to be identified within a preset data period includes: Obtain the maintenance records of the vehicle to be identified; The time point of the most recent maintenance of the vehicle to be identified is traced back in the maintenance record, and the time point is used as the starting time for acquiring multi-source operational data; Determine whether the vehicle to be judged meets the preset number of scans and number of ride orders within the preset data period from the start time to the current time; If the vehicle to be identified meets the preset number of scans and the number of ride orders, then the preset data period is determined to be valid, and multi-source operation data of the vehicle to be identified within the preset data period is obtained.

3. The method for identifying suspected faults in shared vehicles by integrating multi-source data according to claim 1, characterized in that, The calculation of at least two abnormal indicators based on the multi-source operational data includes: For the no-order anomaly rate, obtain the order conversion data of users who scanned the code but did not generate a ride order; In the order conversion data, records of loss due to user reasons and non-vehicle malfunction reasons are excluded. User reasons include user account arrears and user failure to complete real-name or facial authentication. Non-vehicle malfunction reasons include vehicle low battery and vehicle outside the operating area. The remaining order conversion data after excluding the lost records are counted, and the proportion of the remaining order conversion data to the total number of scans is calculated to obtain the no-order anomaly rate.

4. The method for identifying suspected faults in shared vehicles by integrating multi-source data according to claim 1, characterized in that, The calculation of at least two abnormal indicators based on the multi-source operational data includes: For the abnormal trajectory rate, based on the status field in the vehicle trajectory data, the trajectory points in the vehicle trajectory data are divided into stationary trajectory points and moving trajectory points. For the stationary trajectory point, if the spatial distance between the stationary trajectory point and the previous adjacent trajectory point is greater than a first preset threshold, then the stationary trajectory point is determined to be a stationary abnormal point. For the motion state trajectory point, the running speed is calculated based on the spatial distance and time difference between the motion state trajectory point and the previous adjacent trajectory point. If the running speed is greater than the second preset threshold, the motion state trajectory point is determined to be a motion anomaly point. The abnormal trajectory rate is obtained by statistically analyzing the proportion of stationary abnormal points and moving abnormal points to the total number of trajectory points.

5. The method for identifying suspected faults in shared vehicles by integrating multi-source data according to claim 1, characterized in that, The calculation of at least two abnormal indicators based on the multi-source operational data includes: For the battery efficiency anomaly rate, the battery efficiency of all vehicles in the region in the previous battery swap cycle is statistically analyzed, and the battery efficiency percentile threshold is obtained from the statistical results according to a preset percentage. After this battery swap, the current battery percentage and the total mileage after the swap will be collected at fixed time intervals during the ride. When it is detected that the difference between the initial battery percentage after this battery swap and the current battery percentage is greater than or equal to 50%, the process of calculating the current battery efficiency based on the total riding mileage after the battery swap, the current battery percentage, and the initial battery percentage is initiated for this battery swap cycle. After each calculation, if the current battery efficiency is less than the battery efficiency quantile threshold, the battery efficiency is determined to be abnormal. The battery efficiency anomaly rate is obtained by calculating the ratio of the number of times the battery efficiency was determined to be abnormal to the total number of determinations during this battery swapping cycle.

6. The method for identifying suspected faults in shared vehicles by integrating multi-source data according to claim 1, characterized in that, The calculation of at least two abnormal indicators based on the multi-source operational data includes: For the aforementioned power outage rate during riding, battery power and battery voltage data of the vehicle are collected during riding. When the vehicle's battery level drops from a non-zero value to zero at an adjacent data collection time point, and the battery voltage at the current time point is less than or equal to the zero-battery voltage quantile threshold calculated based on historical vehicle data in the region, it is marked as an abnormal power outage of the vehicle. The number of abnormal power outages of vehicles within the battery swapping cycle is counted, and the proportion of the abnormal power outages to the total number of riding orders is calculated to obtain the riding power outage rate.

7. The method for identifying suspected faults in shared vehicles by integrating multi-source data according to claim 1, characterized in that, The step of scoring the suspected fault of the vehicle to be identified based on at least two of the calculated abnormal indicators using a preset scoring model includes: For each of the at least two abnormal indicators, obtain the preset weight coefficient of the abnormal indicator; A suspected fault score is obtained by weighting and summing at least two of the abnormal indicators according to the corresponding preset weight coefficients.

8. A device for identifying suspected faults in shared vehicles by integrating multi-source data, characterized in that, The shared vehicle suspected fault detection device that integrates multi-source data includes: The data acquisition module is used to acquire multi-source operational data of the vehicle to be judged within a preset data period. The multi-source operational data includes order conversion data, riding order data, user feedback data, vehicle trajectory data, and battery data. An abnormal indicator calculation module is used to calculate at least two abnormal indicators based on the multi-source operation data. The abnormal indicators include no-order abnormality rate, short-time riding order rate, fault feedback rate, abnormal trajectory rate, battery efficiency abnormality rate, and riding power outage rate. The suspected fault scoring module is used to score the suspected fault of the vehicle to be identified based on at least two of the calculated abnormal indicators and through a preset scoring model. The suspected fault determination module is used to compare the obtained suspected fault score with a preset fault score threshold. When the suspected fault score is higher than the fault score threshold, the vehicle to be judged is judged as a suspected fault vehicle.

9. A device for identifying suspected faults in shared vehicles by integrating multi-source data, characterized in that, The shared vehicle suspected fault detection device that integrates multi-source data includes: a memory and at least one processor, wherein the memory stores instructions, and the memory and the at least one processor are interconnected via a line; The at least one processor invokes the instructions in the memory to cause the shared vehicle suspected fault detection device that integrates multi-source data to perform the shared vehicle suspected fault detection method that integrates multi-source data as described in any one of claims 1-7.

10. A computer-readable storage medium storing a computer program thereon, characterized in that, When the computer program is executed by the processor, it implements the method for identifying suspected faults in shared vehicles that integrates multi-source data as described in any one of claims 1-7.

Images