A Vehicle Odometer Tampering Detection Method Based on Multi-Source Odometer Reliability Scoring and Time-Sequence Path
By constructing a weighted temporal node network through multi-source mileage credibility scoring and temporal path optimization, the problems of single data and insufficient credibility in vehicle mileage detection are solved, achieving accurate odometer tamper detection and mileage prediction, which is applicable to multiple commercial scenarios.
Patent Information
- Application Number
- CN202611385132.X
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2026-09-08
- Publication Date
- 2026-11-17
AI Technical Summary
Existing vehicle mileage detection technologies suffer from problems such as limited data sources, lack of reliable quantification, and insufficient time-series processing capabilities, leading to invalid detection results, missed detections, false detections, and difficulties in mileage completion.
By constructing a multi-source mileage credibility scoring model, using a differentiated credibility scoring model to calculate the weight of each time-series mileage data, and combining it with time-series path optimization, a weighted time-series mileage node network is constructed to solve for the shortest path, fit the optimal real mileage time-series curve, identify odometer tampering behavior, and fill in missing mileage.
It achieves accurate identification of various odometer tampering behaviors, reduces false detection and false negative rates, completes historical mileage, and predicts current real mileage, and is applicable to scenarios such as used car transactions, car manufacturer maintenance, and insurance risk control.
Smart Images

Figure CN123116017A_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the fields of intelligent vehicle mileage verification, used car inspection, and vehicle risk control data analysis, specifically a method for detecting vehicle odometer tampering based on multi-source mileage credibility scoring and time-series paths. Background Technology
[0002] Vehicle mileage is a core parameter for assessing vehicle value, determining vehicle condition, verifying maintenance cycles, and controlling auto insurance risks. The accuracy of mileage data directly determines the fairness of used car transactions, the accuracy of vehicle maintenance, and the effectiveness of risk control assessments. Currently, mileage tampering (commonly known as odometer manipulation) is rampant. Some car owners lower the mileage by altering the odometer reading or disabling OBD data, thus artificially inflating the vehicle's value, disrupting the used car market order, and posing significant risks to vehicle maintenance and auto insurance underwriting.
[0003] Existing vehicle odometer tampering detection technologies have several shortcomings: First, they rely on a single data source, with most detection schemes depending solely on the displayed mileage or OBD data, failing to integrate multi-dimensional data such as maintenance records and third-party estimated mileage. If a single data source is tampered with, the detection results become completely invalid. Second, they lack a reliable data quantification mechanism, making it impossible to distinguish the authenticity and reliability differences of mileage data from different sources, and lacking a scientific basis for screening abnormal data. Third, their time-series data processing capabilities are insufficient. Judging odometer tampering solely through mileage comparison at a single time point cannot handle complex scenarios such as time-series mileage breakpoints, missing data, and periodic tampering, easily leading to missed or false detections. Fourth, they lack mileage completion and prediction capabilities. For scenarios where historical mileage data is missing or data for certain periods has been completely tampered with, they cannot reconstruct the vehicle's true mileage trajectory, making it difficult to accurately determine the extent of odometer tampering and the current true mileage.
[0004] In view of the shortcomings of the existing technologies, there is an urgent need for a scientific, accurate and multi-scenario-adaptable solution for vehicle odometer tampering detection and mileage prediction. Summary of the Invention
[0005] The purpose of this invention is to provide a vehicle odometer tampering detection method based on multi-source mileage credibility scoring and time-series path. By fusing multi-source data, quantifying credibility scoring, and optimizing the time-series shortest path, it accurately identifies various odometer tampering behaviors, restores the vehicle's true mileage time-series trajectory, and achieves missing mileage completion and accurate prediction of the current mileage. This solves the technical problems mentioned in the background art, such as the single type of vehicle odometer tampering detection data, lack of quantifiable credibility standards, low accuracy in identifying time-series anomalies, inability to complete missing mileage, and inaccurate prediction of the current mileage.
[0006] To solve the above-mentioned technical problems, the present invention provides the following technical solution: The method for detecting vehicle odometer tampering based on multi-source mileage reliability scoring and time-series path analysis includes the following steps: For each target vehicle, time-series mileage data throughout its entire lifecycle is collected, along with its corresponding collection time, collection scenario, and data source identifier. A multi-source time-series mileage raw dataset is constructed, and preprocessing is performed on this raw dataset. Construct a differentiated credibility scoring model and use it to calculate the final credibility weight value for each time-series mileage data. Using the collection time of each mileage data as the time-series node, the mileage values from different sources and at different times as the node attributes, and the mileage increment, time interval, and confidence weight between adjacent time-series nodes as edge weight parameters, a weighted time-series mileage node network is constructed. With the optimization objectives of maximizing the credibility weight and minimizing the time series error, the shortest path between adjacent nodes in the weighted time series mileage node network is solved, and the optimal real mileage time series curve for the entire vehicle life cycle is obtained by fitting. By comparing the optimal real mileage time-series curve of the vehicle's entire life cycle with the vehicle's entire life cycle mileage time-series curve, the time point of vehicle mileage tampering and the mileage tampering value are determined. Finally, by combining all the data, it is determined whether the vehicle has mileage tampering behavior. At the same time, the tampering time period, tampering range, and tampering confidence level can be output. Based on the optimal real mileage time-series curve of the vehicle's entire life cycle, the missing mileage nodes in history are filled in by the time-series fitting algorithm. At the same time, combined with the vehicle's usage time, average annual mileage, and operating conditions, the vehicle's real mileage at the current moment is predicted.
[0007] The operating condition characteristics include maintenance records or accident data, as well as customer complaint data.
[0008] Furthermore, the full lifecycle time-series mileage data includes maintenance mileage, OBD mileage, estimated mileage, and displayed mileage.
[0009] This invention constructs a four-dimensional multi-source mileage data system, integrating maintenance mileage, OBD mileage, estimated mileage, and odometer-displayed mileage, covering official records, vehicle-mounted underlying data, third-party estimated data, and terminal display data, thus completely solving the problem of distortion from a single data source.
[0010] Furthermore, the confidence weight value calculation step includes: The steps for calculating the credibility weight value include: Determine the basic weights for maintenance mileage, OBD mileage, estimated mileage, and displayed mileage; Calculate the historical error rate of maintenance mileage, OBD mileage, estimated mileage, and displayed mileage; Calculate the corresponding correction coefficients based on the historical error rates of maintenance mileage, OBD mileage, estimated mileage, and odometer-displayed mileage, respectively. The time interval between the collection time and the current reference time for maintenance mileage, OBD mileage, estimated mileage, and displayed mileage is obtained respectively, and a time decay coefficient is generated. Based on the data source identifiers of maintenance mileage, OBD mileage, estimated mileage, and displayed mileage, a source credibility coefficient is matched respectively; The basic weights of maintenance mileage, OBD mileage, estimated mileage, and displayed mileage are respectively multiplied by the time decay coefficient, the credibility coefficient, and the correction coefficient to correct the basic weights. The corrected base weights of maintenance mileage, OBD mileage, estimated mileage, and displayed mileage are normalized to obtain the final credibility weights corresponding to maintenance mileage, OBD mileage, estimated mileage, and displayed mileage.
[0011] This process involves: first, constructing a mapping table containing various data source codes and corresponding source credibility coefficients; extracting data source identifiers for maintenance mileage, OBD mileage, estimated mileage, and displayed mileage; matching the extracted data source identifiers with the mapping table to obtain the source credibility coefficient for each type of mileage data; the higher the source authority and the greater the difficulty of tampering, the higher the obtained source credibility coefficient; and then using the source credibility coefficient to correct the static basic weights.
[0012] Furthermore, the time decay coefficient : ; The correction coefficient : ; In the formula, This represents the time interval between the acquisition time of the full lifecycle time-series mileage data and the current reference time, where N represents the minimum time interval and M represents the maximum time interval. Indicates the total effective period in years. Indicates the error rate. This represents the maximum baseline threshold for the error rate. Indicates the tolerance error threshold. Indicates the lower limit of error. This indicates the upper limit of the error.
[0013] Furthermore, the edge weight parameters : ; In the formula, the time interval normalization coefficient Mileage increment compliance coefficient Node credibility weight .
[0014] The maximum time span is obtained by taking the total time from when the vehicle leaves the factory to the current time. The time interval between each mileage data collection time and the current time is calculated. The time interval normalization coefficient is obtained by calculating the ratio of each time interval to the maximum time span.
[0015] Among them, the mileage increment compliance coefficient is used to determine whether the mileage growth between adjacent time nodes conforms to the normal vehicle driving pattern, and to prevent abnormal situations such as mileage rollback, zero growth, and sudden drop.
[0016] The system retrieves the collection time and corresponding mileage values of two adjacent sets of mileage records chronologically, calculates the time interval between the two sets of records and the actual mileage increment, and combines this with a preset threshold for the average daily normal mileage of the vehicle to calculate the reasonable mileage increment range within that time interval. Based on the matching relationship between the actual mileage increment and the reasonable mileage increment range, the mileage increment compliance coefficient is calculated. Specifically: If the actual mileage increment is between the lower limit and the upper limit of the reasonable mileage increment, the mileage increment compliance coefficient is set to 1; if the actual mileage increment is greater than 0 and less than the lower limit of the reasonable mileage increment, the ratio of the actual mileage increment to the lower limit of the reasonable mileage increment is used as the mileage increment compliance coefficient; if the actual mileage increment is greater than the upper limit of the reasonable mileage increment, the ratio of the upper limit of the reasonable mileage increment to the actual mileage increment is used as the mileage increment compliance coefficient; if the actual mileage increment is less than or equal to 0, a preset minimum compliance coefficient is assigned.
[0017] The greater the deviation of the actual mileage increment from the reasonable range, the lower the mileage increment compliance coefficient will be; if the mileage increment is less than or equal to zero, a fixed minimum compliance coefficient will be set; the mileage increment compliance coefficient will be included in the weighted joint correction calculation.
[0018] Furthermore, an optimized Bellman-Ford algorithm is used to solve for the shortest path between adjacent nodes in the weighted temporal mileage node network. The algorithm is optimized to take advantage of the characteristics of the temporal network—no negative weights and unidirectional temporal progression—to avoid temporal loop problems. It traverses all combined paths of multi-source mileage nodes and selects the path with the minimum overall weight as the optimal true mileage path.
[0019] This invention introduces a time-weighted node network and a shortest path algorithm to transform the mileage screening problem into a time-series path optimization problem. By filtering the global path optimally, it eliminates locally abnormal and tampered data and restores the complete, continuous, and true time-series trajectory of vehicle mileage. Compared with the traditional single-point comparison method, the detection accuracy is greatly improved.
[0020] Furthermore, the criteria for determining odometer tampering include: if the difference between the optimal actual mileage and the displayed mileage exceeds a preset threshold and lasts for more than T days (30 days is acceptable), it is determined as deterministic odometer tampering; if the mileage rollback occurs in the time-series mileage without maintenance or malfunction, it is determined as malicious odometer tampering; if the phased mileage increment is significantly lower than the average increment of vehicles under the same operating conditions, it is determined as covert odometer tampering. The phased mileage increment is obtained by dividing the vehicle's usage cycle into multiple continuous time-series stages and using big data to calculate the phased mileage increment by the time span of each stage and the difference between the first and last mileages.
[0021] Furthermore, a time-series weighted fitting algorithm is employed, based on the optimal mileage time-series curve, integrating data on vehicle age, driving conditions, and maintenance records to dynamically adjust the prediction model parameters and output the current true mileage prediction value and prediction error range. For example, the average annual driving rate is dynamically reduced according to the increase in vehicle age to match the actual vehicle usage pattern of reduced driving frequency; the mileage growth slope is dynamically adjusted according to different driving conditions such as long-term commuting, long-distance driving, and long-term inactivity; and multiple high-reliability maintenance mileage records are used as time-series anchors to continuously correct the baseline slope, reasonable upper and lower thresholds of fluctuation, and attenuation parameters of the fitting model, thereby continuously optimizing the fit of the optimal mileage time-series curve.
[0022] The system includes a technical solution: a vehicle odometer tampering detection system based on multi-source mileage credibility scoring and time-series path analysis, comprising: a data acquisition module, a credibility scoring module, a network construction module, a path solving module, an odometer tampering detection module, and a mileage prediction module. The data acquisition module is used to collect multi-source time-series data of vehicle maintenance mileage, OBD mileage, estimated mileage, and odometer-displayed mileage; The credibility scoring module is used to assign basic scores and dynamically correct them for multi-source mileage data, and generate credibility weights for each mileage node. The network construction module is used to build a weighted temporal mileage node network and define node and edge weight parameters; The path solving module is used to solve for the shortest time path and fit the optimal real mileage time-series curve. The odometer tampering detection module is used to compare and analyze mileage deviations, identify various abnormal odometer tampering behaviors, and output detection results. The mileage prediction module is used to fill in missing mileage nodes and predict the vehicle's current actual mileage.
[0023] The invention includes a technical solution, an electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, when the processor executes the computer program, it implements the vehicle odometer tampering detection method based on multi-source mileage reliability scoring and time-series path as described in any of the above technical solutions.
[0024] Compared with existing technologies, the beneficial effects achieved by this invention are as follows: This invention utilizes the basic score assignment and dynamic correction of multi-source mileage data to generate credibility weights for each mileage node. A weighted time-series mileage node network is constructed based on these credibility weights. The shortest time-series path is solved within this network to fit the optimal real mileage time-series curve. Mileage deviations are compared and analyzed against the fitted optimal real mileage time-series curve to identify various odometer tampering behaviors and output detection results. These results are then used to complete missing mileage nodes and predict the vehicle's current actual mileage. This invention quantifies data quality through credibility scoring and combines it with global optimization using the shortest time-series path to avoid interference from single-point data anomalies. It can accurately identify various tampering behaviors that are difficult to detect using traditional methods, such as mileage rollback, phased downsizing, and hidden odometer tampering, significantly reducing false positive and false negative rates. Furthermore, it not only enables accurate determination of odometer tampering behavior and quantification of the tampering period and magnitude, but also completes missing historical mileage and predicts the current actual mileage, providing a one-stop solution for the entire process of vehicle mileage traceability, verification, and prediction. This invention is applicable to multiple scenarios such as used car transaction inspection, car manufacturer maintenance supervision, insurance company risk control, and vehicle asset management, and has extremely high commercial value and industry promotion value. Attached Figure Description
[0025] The accompanying drawings are provided to further illustrate the invention and form part of the specification. They are used in conjunction with embodiments of the invention to explain the invention and do not constitute a limitation thereof. In the drawings: Figure 1 This is a flowchart of the steps of the vehicle odometer tampering detection method based on multi-source mileage reliability scoring and time-series path according to the present invention. Detailed Implementation
[0026] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.
[0027] Example 1: Routine Vehicle Odometer Tampering Detection and Mileage Prediction S1. Data Collection: Select a family passenger vehicle that has been used for 5 years and collect its full life cycle mileage data, including: 12 official maintenance mileage records from 4S stores, 36 real-time mileage data from the vehicle's OBD backend, 8 estimated mileage records from third-party platforms, and 40 real-time mileage records displayed on the dashboard. Simultaneously record the collection time, collection channel, and data status of each data record to construct the original time-series mileage dataset.
[0028] S2. Credibility Score: The default base weights for maintenance mileage, OBD mileage, estimated mileage, and displayed mileage are 0.75, 0.60, 0.20, and 0.90, respectively.
[0029] This embodiment addresses four mainstream mileage data sources for vehicle mileage verification scenarios. Based on each data source's data generation mechanism, storage security level, tampering difficulty, traceability capabilities, and long-term statistical error level, a pre-defined credibility level is established. A unique, fixed data source code is assigned to each data source type, along with a dedicated source credibility coefficient, forming a static mapping relationship that the system can directly access. For example, estimated mileage is the result of multi-source fusion and fitting by an algorithm, with no room for human intervention or tampering, resulting in the highest credibility; maintenance mileage is official backend archived data, traceable, and with an extremely low error rate; OBD mileage is permanently stored in the vehicle's electronic control unit and can only be tampered with using specialized equipment; and displayed mileage is modifiable data from the local instrument panel, with the lowest tampering threshold. Therefore, the mapping relationship can be preset as follows: source code S01 corresponds to estimated mileage, source confidence coefficient 0.92; source code S02 corresponds to maintenance mileage, source confidence coefficient 0.78; source code S03 corresponds to OBD mileage, source confidence coefficient 0.60; source code S04 corresponds to displayed mileage, source confidence coefficient 0.20, thus completing the pre-fixed storage of the mapping table.
[0030] Then, the data source identifiers attached to each of the four types of mileage data are extracted, and the extracted data source identifiers are searched and matched with a preset mapping table to retrieve the corresponding confidence coefficients.
[0031] Based on the historical average error rates of various data sources using a large sample of vehicles, the historical error rates for maintenance mileage (0.02), OBD mileage (0.08), estimated mileage (0.01), and instrument mileage (0.30) were used to calculate correction coefficients. : ; In the formula, Indicates the error rate. This represents the maximum baseline threshold for the error rate. Indicates the tolerance error threshold. Indicates the lower limit of error. This indicates the lower limit of the error. Among them... It can be 1%. It can take the value 3%. It can take the value 5%.
[0032] The time interval between the collection time and the current reference time for maintenance mileage, OBD mileage, estimated mileage, and displayed mileage is obtained respectively, and a time decay coefficient is generated; the time decay coefficient : ; In the formula, This represents the time interval between the acquisition time of the full lifecycle time-series mileage data and the current reference time, where N represents the minimum time interval and M represents the maximum time interval. It indicates the total effective period in years.
[0033] The basic weights of maintenance mileage, OBD mileage, estimated mileage, and displayed mileage are respectively multiplied by the time decay coefficient, the credibility coefficient, and the correction coefficient to correct the basic weights. The corrected base weights of maintenance mileage, OBD mileage, estimated mileage, and displayed mileage are normalized to obtain the final credibility weights corresponding to maintenance mileage, OBD mileage, estimated mileage, and displayed mileage.
[0034] S3. Network Construction: Using the minimum daily collection time as the time-series node, mileage data from different sources are used as node attributes. The time interval and mileage increment of adjacent time-series nodes are calculated. The edge weights are calculated in combination with the credibility weights to construct a complete time-series weighted node network.
[0035] S4. Path Determination: The optimized Bellman-Ford algorithm is used to traverse all time-series paths, select the shortest path with the smallest comprehensive weight, fit the optimal true mileage time-series curve, and remove 3 nodes with abnormally low displayed mileage and 2 nodes with estimated mileage errors.
[0036] S5. Odometer Tampering Detection: By comparing the optimal actual mileage curve with the displayed mileage, it was found that the mileage dropped significantly in the third year of vehicle use, with the actual mileage exceeding the displayed mileage by 42,000 kilometers. This indicates that there was a definite malicious odometer tampering behavior, and the time period and extent of the tampering were accurately identified.
[0037] S6. Mileage Prediction: Based on the optimal time-series curve, the five missing mileage nodes during the vehicle's use period are completed. Combined with the vehicle's average annual mileage and household operating conditions, the current vehicle's actual mileage is predicted to be 126,000 kilometers, with a prediction error range of ±3,000 kilometers.
[0038] Example 2: Table Adjustment Detection in Data Missing Scenarios For older vehicles with missing OBD data or incomplete maintenance records, this invention employs a detection method: by retaining valid maintenance mileage, estimated mileage, and historical normal mileage data, a reliability score and time-series network are constructed. The shortest path algorithm is used to avoid missing data nodes, and a continuous mileage trajectory is fitted based on valid nodes. This successfully identifies the vehicle's periodic hidden odometer tampering behavior and accurately completes the mileage data for missing periods, achieving effective prediction of the current mileage. In actual testing, even with 30% data loss, the detection accuracy of this method still reaches over 95%.
[0039] It should be noted that, in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article, or apparatus.
[0040] Finally, it should be noted that the above descriptions are merely preferred embodiments of the present invention and are not intended to limit the present invention. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art can still modify the technical solutions described in the foregoing embodiments or make equivalent substitutions for some of the technical features. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.
Claims
1. A method for detecting vehicle odometer tampering based on multi-source mileage reliability scoring and time-series path analysis, characterized in that, The steps include: For each target vehicle, time-series mileage data throughout its entire lifecycle is collected, along with its corresponding collection time, collection scenario, and data source identifier. A multi-source time-series mileage raw dataset is constructed, and preprocessing is performed on this raw dataset. Construct a differentiated credibility scoring model and use it to calculate the credibility weight of each time-series mileage data point; Using the collection time of each mileage data as the time-series node, the mileage values from different sources and at different times as the node attributes, and the mileage increment, time interval, and confidence weight between adjacent time-series nodes as edge weight parameters, a weighted time-series mileage node network is constructed. With the optimization objectives of maximizing the credibility weight and minimizing the time series error, the shortest path between adjacent nodes in the weighted time series mileage node network is solved, and the optimal real mileage time series curve for the entire vehicle life cycle is obtained by fitting. By comparing the optimal real mileage time-series curve of the vehicle's entire life cycle with the mileage time-series curve of the vehicle's entire life cycle, the time point of mileage tampering and the value of mileage tampering are determined, and finally, all data are combined to determine whether the vehicle has engaged in mileage tampering. Based on the optimal real mileage time-series curve of the vehicle's entire life cycle, the missing mileage nodes in history are filled in by the time-series fitting algorithm. At the same time, combined with the vehicle's usage time, average annual mileage, and operating conditions, the vehicle's real mileage at the current moment is predicted.
2. The vehicle odometer tampering detection method based on multi-source mileage reliability scoring and time-series path as described in claim 1, characterized in that, The full lifecycle time-series mileage data includes maintenance mileage, OBD mileage, estimated mileage, and mileage displayed on the instrument panel.
3. The vehicle odometer tampering detection method based on multi-source mileage reliability scoring and time-series path as described in claim 2, characterized in that, The steps for calculating the credibility weight value include: Determine the basic weights for maintenance mileage, OBD mileage, estimated mileage, and displayed mileage; Calculate the historical error rate of maintenance mileage, OBD mileage, estimated mileage, and displayed mileage; Calculate the corresponding correction coefficients based on the historical error rates of maintenance mileage, OBD mileage, estimated mileage, and odometer-displayed mileage, respectively. The time interval between the collection time and the current reference time for maintenance mileage, OBD mileage, estimated mileage, and displayed mileage is obtained respectively, and a time decay coefficient is generated. Based on the data source identifiers of maintenance mileage, OBD mileage, estimated mileage, and displayed mileage, a source credibility coefficient is matched accordingly; The basic weights of maintenance mileage, OBD mileage, estimated mileage, and displayed mileage are respectively multiplied by the time decay coefficient, the credibility coefficient, and the correction coefficient to correct the basic weights. The corrected base weights of maintenance mileage, OBD mileage, estimated mileage, and displayed mileage are normalized to obtain the final credibility weights corresponding to maintenance mileage, OBD mileage, estimated mileage, and displayed mileage.
4. The vehicle odometer tampering detection method based on multi-source mileage reliability scoring and time-series path as described in claim 3, characterized in that, The time decay coefficient : ; The correction coefficient : ; In the formula, This represents the time interval between the acquisition time of the full lifecycle time-series mileage data and the current reference time, where N represents the minimum time interval and M represents the maximum time interval. Indicates the total effective period in years. Indicates the error rate. This represents the maximum baseline threshold for the error rate. Indicates the tolerance error threshold. Indicates the lower limit of error. This indicates the upper limit of the error.
5. The vehicle odometer tampering detection method based on multi-source mileage reliability scoring and time-series path as described in claim 1, characterized in that: The edge weight parameters : ; In the formula, Represents the normalization coefficient of the time interval. Indicates the compliance coefficient for mileage increments. Represents the node's credibility weight .
6. The vehicle odometer tampering detection method based on multi-source mileage reliability scoring and time-series path as described in claim 1, characterized in that: An optimized Bellman-Ford algorithm is used to solve for the shortest path between adjacent nodes in the weighted temporal mileage node network.
7. The vehicle odometer tampering detection method based on multi-source mileage reliability scoring and time-series path as described in claim 1, characterized in that: The criteria for determining odometer tampering include: if the optimal real mileage of the vehicle throughout its entire life cycle exceeds a preset threshold and the duration is greater than T days, it is determined to be deterministic odometer tampering; if the time-series mileage shows a mileage rollback that is not due to maintenance or fault, it is determined to be malicious odometer tampering; if the mileage increment in a certain period is lower than the average increment of vehicles under the same operating conditions, it is determined to be implicit odometer tampering.
8. The vehicle odometer tampering detection method based on multi-source mileage reliability scoring and time-series path as described in claim 1, characterized in that: The time-series weighted fitting algorithm is adopted, based on the optimal mileage time-series curve, and integrates data on vehicle age, driving conditions, and maintenance records to dynamically correct the prediction model parameters and output the current true mileage prediction value and prediction error range.
9. A vehicle odometer tampering detection system based on multi-source mileage reliability scoring and time-series path analysis, characterized in that, include: Data acquisition module, credibility scoring module, network construction module, path solving module, odometer tamper detection module, mileage prediction module; The data acquisition module is used to collect multi-source time-series data of vehicle maintenance mileage, OBD mileage, estimated mileage, and odometer-displayed mileage; The credibility scoring module is used to correct the basic weights of multi-source mileage data and generate credibility weights for each mileage node. The network construction module is used to build a weighted temporal mileage node network and define node and edge weight parameters; The path solving module is used to solve for the shortest time path and fit the optimal real mileage time-series curve. The odometer tampering detection module is used to analyze the deviation between the fitted optimal real mileage and the collected mileage, identify abnormal odometer tampering behavior, and output the detection results. The mileage prediction module is used to fill in missing mileage nodes and predict the vehicle's current actual mileage.
10. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the computer program, it implements the method of any one of claims 1.
