A method and system for vehicle trajectory verification based on the fusion data of BeiDou positioning and ETC.

Abstract

This invention discloses a vehicle trajectory auditing method and system based on the fusion of BeiDou positioning and ETC data, belonging to the field of intelligent transportation technology. When there are missing ETC gantry transaction records for a target vehicle's travel segment, this method extracts the BeiDou positioning trajectory point sequence between the starting and ending gantry, calculates the trajectory interval speed deviation factor, and eliminates outlier trajectory points based on the path completion characteristic value, obtaining an optimized trajectory point sequence. Candidate driving paths are generated based on projection distance constraints and direction consistency constraints. A comprehensive consistency cost is calculated based on time deviation, path length, and speed deviation, and the candidate path with the lowest cost is selected as the completion path. Its accumulated mileage replaces the default path to complete the toll audit. This invention can improve the accuracy of vehicle trajectory recovery, toll accuracy, and reliability of toll auditing when ETC gantry transaction records are missing.

Description

Technical Field

[0001] This invention relates to the field of intelligent transportation technology, specifically to a vehicle trajectory auditing method and system based on the fusion of BeiDou positioning and ETC data. Background Technology

[0002] With the continuous development of highway network toll collection systems, ETC gantry transaction data has become an important basis for vehicle toll calculation and auditing. However, in actual operation, due to factors such as communication abnormalities, equipment failures, network transmission delays, or data loss, some vehicles may have missing gantry transaction records, resulting in the incomplete acquisition of the vehicle's actual passage trajectory between adjacent gantries.

[0003] To address the aforementioned issues, existing technologies typically determine the corresponding travel path based on gantry information with transaction records before and after the vehicle, according to the highway network topology, and then perform tolling or auditing accordingly. However, when there is a discrepancy between the vehicle's actual route and the route estimated based on the road network, the calculated mileage can easily differ from the vehicle's actual mileage, thus affecting the accuracy of the tolling results. Furthermore, in complex scenarios involving detours or abnormal traffic, relying solely on gantry transaction records for route estimation is insufficient to accurately reflect the vehicle's actual travel status, hindering subsequent toll collection and auditing efforts. Summary of the Invention

[0004] This invention provides a vehicle trajectory auditing method and system based on the fusion of BeiDou positioning and ETC data, in order to solve the technical problems of insufficient accuracy in vehicle trajectory recovery, low billing accuracy, and insufficient reliability of toll auditing when ETC gantry transaction records are missing in the prior art.

[0005] The technical solution of the present invention to solve the above-mentioned technical problems is as follows:

[0006] In a first aspect, the present invention provides a vehicle trajectory auditing method based on the fusion of BeiDou positioning and ETC data, comprising:

[0007] When there are missing ETC gantry transaction records on the road segment through which the target vehicle passes, the gantry with a transaction record at the last upstream position of the missing road segment is taken as the starting gantry and the gantry with a transaction record at the first downstream position is taken as the ending gantry. The Beidou positioning trajectory point sequence of the target vehicle between the starting gantry and the ending gantry is extracted.

[0008] The BeiDou positioning trajectory point sequence is projected onto the highway network, and the cumulative deviation of the average speed between adjacent points relative to the historical speed limit is calculated as the speed deviation factor for each trajectory interval.

[0009] Based on the positional relationship of each trajectory point relative to the starting gantry, the path completion characterization value is calculated, and combined with the road network projection distance, outlier trajectory points are removed to obtain an optimized trajectory point sequence;

[0010] Starting from the starting gantry and ending from the ending gantry, each trajectory point in the optimized trajectory point sequence is matched to an arc segment in the road network that satisfies the projection distance constraint and the direction consistency constraint, and connected in the matching order to generate at least one candidate driving path.

[0011] Based on the time deviation, arc length, and velocity deviation of the corresponding trajectory intervals in each candidate path, the comprehensive consistency cost of each path is calculated, and the candidate path with the lowest cost is selected as the completion path.

[0012] The cumulative mileage of the completed path is output as the actual mileage of the target vehicle in the missing road segment, which is used to replace the default path to complete the billing audit.

[0013] Secondly, the present invention provides a vehicle trajectory auditing system based on the fusion of BeiDou positioning and ETC data, comprising:

[0014] The trajectory data acquisition module is used to extract the BeiDou positioning trajectory point sequence of the target vehicle between the starting gantry and the ending gantry when there is a missing ETC gantry transaction record in the road segment through which the target vehicle passes.

[0015] The speed deviation analysis module is used to project the Beidou positioning trajectory point sequence onto the highway network and calculate the cumulative deviation of the average speed between adjacent points relative to the historical speed limit, which serves as the speed deviation factor for each trajectory interval.

[0016] The trajectory optimization module is used to calculate the path completion characterization value based on the positional relationship of each trajectory point relative to the starting gantry, and combine it with the road network projection distance to remove outlier trajectory points and obtain an optimized trajectory point sequence.

[0017] The candidate path generation module is used to match each trajectory point in the optimized trajectory point sequence to an arc segment in the road network that satisfies the projection distance constraint and the direction consistency constraint, with the starting gantry as the starting point and the ending gantry as the ending point, and connect them in the matching order to generate at least one candidate driving path.

[0018] The path completion module is used to calculate the comprehensive consistency cost of each path based on the time deviation, arc length, and speed deviation of the corresponding trajectory interval in each candidate path, and select the candidate path with the lowest cost as the completion path.

[0019] The billing audit module is used to output the cumulative mileage of the completed path as the actual mileage of the target vehicle in the missing road segment, and is used to replace the default path to complete the billing audit.

[0020] One or more technical solutions provided in this invention have at least the following technical effects or advantages:

[0021] In the event of missing ETC gantry transaction records, this invention combines BeiDou positioning trajectory data and highway network information to analyze and recover the actual driving trajectory of a vehicle. By calculating speed deviation factors, path completion characterization values, and comprehensive consistency costs, a complete path that matches the actual driving situation of the vehicle is selected, and the actual driving mileage is determined based on the complete path to complete the tolling audit.

[0022] Compared to billing methods that rely solely on default routes, this invention can more accurately restore the actual travel path of vehicles, improve the accuracy of actual mileage calculation, enhance the reliability of toll collection and trajectory auditing, and facilitate the identification of abnormal travel behavior, thus providing reliable technical support for the refined management of highway network toll collection. Attached Figure Description

[0023] Figure 1 This is a schematic diagram of a vehicle trajectory auditing method based on the fusion of BeiDou positioning and ETC data provided in an embodiment of the present invention;

[0024] Figure 2 This is a flowchart of trajectory recovery and path auditing provided in an embodiment of the present invention;

[0025] Figure 3 This is a schematic diagram of a vehicle trajectory auditing system module based on the fusion of BeiDou positioning and ETC data provided in an embodiment of the present invention;

[0026] In the attached diagram, the components represented by each number are as follows:

[0027] The module includes: trajectory data acquisition module 11, speed deviation analysis module 12, trajectory optimization module 13, candidate path generation module 14, path completion module 15, and billing audit module 16. Detailed Implementation

[0028] This invention provides a vehicle trajectory auditing method and system based on the fusion of BeiDou positioning and ETC data, which specifically solves the technical problems in the prior art such as insufficient accuracy of vehicle trajectory recovery, low billing accuracy, and insufficient reliability of toll auditing when ETC gantry transaction records are missing.

[0029] It should be noted that the terms "comprising" and "having" are intended to cover non-exclusive inclusion. For example, a process, method, system, product, or server that includes a series of steps or units is not necessarily limited to those steps or units that are explicitly listed, but may include other steps or modules that are not explicitly listed or that are inherent to these processes, methods, products, or devices.

[0030] Example 1, as Figure 1 and Figure 2 As shown, this invention provides a vehicle trajectory auditing method based on the fusion of BeiDou positioning and ETC data, the method comprising:

[0031] S100: When there is a missing ETC gantry transaction record on the road segment through which the target vehicle passes, the gantry with a transaction record at the last position upstream of the missing road segment is taken as the starting gantry and the gantry with a transaction record at the first position downstream is taken as the ending gantry. The Beidou positioning trajectory point sequence of the target vehicle between the starting gantry and the ending gantry is extracted.

[0032] In the field of vehicle trajectory verification technology, existing technologies typically rely on the default path between the starting and ending gantries to complete the trajectory when ETC gantry transaction records are missing. This method struggles to accurately reflect the vehicle's actual travel route, potentially impacting subsequent billing accuracy and the reliability of toll collection verification. Furthermore, directly utilizing all BeiDou positioning data for analysis can introduce irrelevant trajectory information, increasing computational complexity and reducing trajectory recovery accuracy. Therefore, accurately extracting the BeiDou positioning trajectory data corresponding to the missing road segment when ETC gantry transaction records are missing becomes the primary technical challenge to be addressed in this step.

[0033] Step S100 in the method provided in this embodiment of the invention includes:

[0034] A time window is defined based on the transaction time of the starting gantry and the transaction time of the ending gantry, and the original BeiDou positioning data of the target vehicle within the time window is extracted.

[0035] Redundant positioning points with inverted or duplicate timestamps in the original BeiDou positioning data are removed, and the times of the remaining positioning points are uniformly transformed to the time reference of the gantry device.

[0036] The average velocity between adjacent points and the straight-line distance between adjacent points are calculated sequentially for the time-aligned positioning point sequence, and each positioning point is projected onto the highway network to obtain the projected distance of each point;

[0037] Location points that meet any of the following conditions will be removed: the average speed between two adjacent points is greater than twice the historical speed limit, or the projected distance of the location point is greater than one-third of the straight-line distance between two adjacent points.

[0038] The remaining positioning points are arranged in chronological order to form the BeiDou positioning trajectory point sequence.

[0039] In this embodiment, the time window refers to the time interval for intercepting the BeiDou positioning data of the target vehicle, which is determined based on the start gantry transaction time and the end gantry transaction time. It is used to limit the trajectory extraction range and avoid introducing irrelevant travel data.

[0040] The gantry device time reference refers to the standard time reference uniformly adopted by the ETC gantry system, which is used to uniformly calibrate the timestamps uploaded by Beidou positioning terminals to eliminate time deviations between different devices and ensure that the positioning data and gantry transaction data have a consistent time reference.

[0041] Projection distance refers to the shortest distance after the BeiDou positioning point is vertically mapped onto the highway network. It is used to measure the degree of deviation of the positioning point from the actual road position. The smaller the projection distance, the more the positioning result matches the actual driving position of the vehicle.

[0042] In this step, a time window is first defined based on the transaction time of the starting gantry and the transaction time of the ending gantry. Then, raw BeiDou positioning data for the target vehicle within the corresponding time period is extracted from the BeiDou positioning database according to this time window. By using the gantry transaction time as the trajectory extraction boundary, it is ensured that the extracted data completely covers the missing transaction segments, while avoiding the introduction of vehicle positioning information from other road segments, thus improving the targeting of subsequent trajectory recovery.

[0043] Subsequently, redundant positioning points with inverted or duplicate timestamps in the original BeiDou positioning data are removed, and the times of the remaining positioning points are uniformly transformed to the gantry device time reference to obtain a time-aligned positioning point sequence. After time alignment, the average velocity between adjacent points and the straight-line distance between adjacent points are calculated sequentially for the time-aligned positioning point sequence, and each positioning point is projected onto the highway network to obtain the corresponding projected distance.

[0044] The average velocity between two adjacent points is calculated based on the straight-line distance between the adjacent positioning points and the corresponding time difference. The calculation method is as follows:

[0045] Average speed = Straight-line distance between two adjacent points ÷ Time difference between two adjacent points

[0046] The straight-line distance between two adjacent points is calculated based on the spatial coordinates of the two points. The calculation method is as follows:

[0047] Straight-line distance = √((difference in x-coordinates)² + (difference in y-coordinates)²).

[0048] For example, if the time difference between two adjacent positioning points is 2s, and the straight-line distance between the two points is 60m calculated by coordinates, then the corresponding average speed can be obtained as 60÷2=30m / s according to the above calculation method; if the time difference remains unchanged but the straight-line distance increases to 80m, then the corresponding average speed is 80÷2=40m / s.

[0049] Furthermore, to ensure the quality of trajectory data, anomaly filtering is performed on the positioning points. When the average speed between two adjacent points is greater than twice the historical speed limit, the corresponding positioning point is identified as an anomaly and removed. This is because the common speed limit on highways is generally 120 km / h, and twice the historical speed limit corresponds to approximately 240 km / h. However, the maximum design speed of ordinary passenger cars and trucks is usually difficult to reach this speed, and due to road conditions and traffic flow limitations, it is almost impossible for vehicles to maintain an average speed above 240 km / h continuously during actual driving. Therefore, when the calculated average speed of adjacent positioning points exceeds twice the historical speed limit, it can be basically determined that the anomaly is caused by incorrect BeiDou positioning timestamps, positioning drift, or position jumps, rather than the actual movement of the vehicle. Using twice the historical speed limit as the judgment threshold can effectively filter out false trajectory points and avoid misjudging normal high-speed driving data.

[0050] On the other hand, when the projected distance of a positioning point is greater than one-third of the straight-line distance between two adjacent points, the positioning point is also judged as an abnormal positioning point and removed. Normally, when a vehicle is traveling normally along a highway lane, the BeiDou positioning error is only a few meters, while the actual displacement of the vehicle within a continuous sampling period usually reaches tens of meters. For example, when a vehicle is traveling at 100 km / h and the sampling interval is 1 second, the displacement between adjacent positioning points is approximately 27.8 meters. Therefore, under normal circumstances, the ratio of projected distance to displacement distance is generally much less than 0.1. This embodiment sets the judgment threshold to one-third, or approximately 0.33, which is significantly higher than the normal positioning error range. This condition is only met when the positioning point deviates significantly from the road centerline, experiences significant drift, or a positioning jump. This effectively identifies outlier positioning points while avoiding the accidental deletion of normal trajectory data, improving the trajectory data retention rate and trajectory recovery accuracy.

[0051] Finally, the remaining positioning points after the above anomaly screening are rearranged in chronological order to form the BeiDou positioning trajectory point sequence.

[0052] For example, suppose the starting gantry transaction time for the target vehicle is 10:00:00 and the ending gantry transaction time is 10:12:00. First, using 10:00:00 to 10:12:00 as a time window, all raw positioning data uploaded within this time period is extracted from the BeiDou positioning database. Assuming five consecutive positioning points P1, P2, P3, P4, and P5 are extracted, and the time difference between positioning point P2 and positioning point P3 is 1 second, with a straight-line distance of 278m, the average speed between the two adjacent points is calculated to be 278 ÷ 1 = 278 m / s = 1000.8 km / h. If the historical speed limit for the current highway segment is 120 km / h, then twice the historical speed limit corresponds to 240 km / h. Since 1000.8 km / h is significantly greater than 240 km / h, it can be determined that the positioning data has an abnormal timestamp or positioning jump. Therefore, positioning point P3 is removed as an abnormal positioning point.

[0053] Furthermore, assuming the straight-line distance between positioning point P4 and positioning point P5 is 30m, and the projected distance of positioning point P5 onto the highway network is 12m, then the ratio of the projected distance to the straight-line distance is 12 / 30 = 0.40, which is greater than one-third of the preset threshold (approximately 0.33). This indicates that the positioning point has significantly deviated from the actual road position and can be identified as an outlier caused by positioning drift. Therefore, positioning point P5 is removed. If the straight-line distance of another positioning point is 30m, and the projected distance is only 5m, then the corresponding ratio is approximately 0.17, which is less than one-third. This indicates that the positioning error is within the normal range and should be retained.

[0054] It should be noted that the above values ​​are for illustrative purposes only and do not constitute a limitation on the present invention.

[0055] Step S100 in the method provided in this embodiment of the invention further includes:

[0056] Based on the highway network between the starting gantry and the ending gantry, all arc segments are extracted to form a candidate arc segment set.

[0057] For each positioning point, the straight-line distance between the positioning point and the adjacent positioning point is taken as the search radius. In the candidate arc segment set, arc segments whose vertical distance to the positioning point does not exceed the search radius are selected to form an initial set of arc segments.

[0058] In the initial screening arc segment set, calculate the direction angle between the direction of each arc segment and the direction of the line connecting two adjacent points of the positioning point, remove arc segments with a direction angle greater than 90 degrees, and take the value with the smallest vertical distance among the remaining arc segments as the projected distance of the positioning point.

[0059] If the directional angle of all arc segments in the initial screening arc segment set is greater than 90 degrees, then the minimum vertical distance in the initial screening arc segment set is taken as the projected distance of the positioning point.

[0060] In this embodiment, the highway network refers to the highway topology structure constructed using nodes and arc segments. Nodes represent locations such as toll stations, ETC gantries, interchanges, or road connection points, while arc segments represent road sections that connect adjacent nodes and have defined driving directions and lengths.

[0061] The vertical distance is the shortest vertical distance from a specified point to a certain highway arc segment, used to measure the spatial proximity between the location point and the road arc segment;

[0062] The directional angle refers to the angle formed between the driving direction of the road arc segment and the direction of the line connecting the current positioning point and its adjacent positioning points. It is used to evaluate the consistency between the vehicle's movement direction and the road direction.

[0063] In this step, all road arcs within the highway network between the starting gantry and the ending gantry are first extracted to form a candidate arc set. Since the actual driving path of a vehicle during the period when ETC transaction records are missing must be between the starting gantry and the ending gantry, limiting the search range to the corresponding road network area can effectively narrow the calculation range, reduce the complexity of subsequent arc matching, and reduce the interference of irrelevant roads on the positioning results.

[0064] Subsequently, for each BeiDou positioning point, the straight-line distance between the positioning point and its adjacent positioning points is used as the search radius. Road arc segments whose vertical distance from the positioning point to the arc segment does not exceed the search radius are selected from the candidate arc segment set to form a preliminary set of arc segments.

[0065] The straight-line distance between adjacent positioning points can approximately reflect the actual movement scale of the vehicle within the current sampling period. Therefore, this distance is used to dynamically construct the search radius, which can automatically adjust the search range according to the vehicle's operating status. When the vehicle is traveling at high speed, the distance between adjacent positioning points is larger, and the search range expands accordingly; when the vehicle is traveling at low speed or in congested traffic, the distance between adjacent positioning points is smaller, and the search range shrinks accordingly, thus balancing matching efficiency and matching accuracy.

[0066] Furthermore, after obtaining the initial set of arc segments, the directional angle between the direction of each candidate arc segment and the direction of the line connecting the current positioning point and its adjacent positioning points is calculated.

[0067] The direction of the line connecting the positioning points is determined by the direction vector formed by the current positioning point and the next positioning point. The direction of the candidate arc segment is determined by the direction vector formed by the starting point and ending point of the corresponding road arc segment. The angle between the directions is calculated based on the angle between the two direction vectors. The calculation method is as follows:

[0068] The angle between directions is equal to arccos[(direction vector one · direction vector two) ÷ (the magnitude of direction vector one × the magnitude of direction vector two)].

[0069] The above calculations yield the angle between the vehicle's actual direction of motion and the candidate road's direction of travel. When the angle is no greater than 90 degrees, it indicates that the vehicle's direction of motion is essentially consistent with the road's direction of travel, and the candidate arc segment meets the direction consistency requirement. When the angle is greater than 90 degrees, it indicates that the vehicle's direction of motion is opposite to the road direction or deviates significantly from it, and the candidate arc segment does not conform to normal vehicle travel patterns, therefore it is discarded. By simultaneously considering both spatial distance constraints and direction consistency constraints, we can avoid incorrectly matching vehicles to adjacent reverse lanes, oncoming ramps, or parallel roads based solely on the nearest distance principle, thus improving the reliability of the location point road matching results.

[0070] For example, if the direction of the line connecting the current positioning point and the next positioning point is 20° east of north, and the road direction of candidate arc segment A is 30° east of north, then the angle between the two directions is 10°, which satisfies the direction consistency constraint and can be used as a valid candidate arc segment; if the road direction of candidate arc segment B is 40° west of south, then the calculated direction angle is about 160°, which is greater than 90°, indicating that if the vehicle is matched to this arc segment, it needs to travel in the opposite direction. Therefore, candidate arc segment B is removed and does not participate in the subsequent projection distance calculation.

[0071] Finally, if the included angle of all candidate arc segments in the initial screening arc segment set is greater than 90 degrees, it indicates that the current positioning point may be affected by factors such as BeiDou positioning errors, special road geometry, or complex topology of interchanges, resulting in the inability to find a candidate road that meets the direction consistency requirement. In this case, to avoid interrupting the subsequent trajectory analysis process due to the positioning point's inability to complete projection, this embodiment no longer performs direction screening. Instead, it directly takes the vertical distance corresponding to the arc segment with the smallest vertical distance in the initial screening arc segment set as the projection distance of the current positioning point, ensuring that all positioning points can obtain the corresponding projection results.

[0072] It should be noted that the above values ​​are for illustrative purposes only and do not constitute a limitation on the present invention.

[0073] In summary, this step, by performing time alignment, abnormal positioning point removal, and road projection matching on the original BeiDou positioning data, and combining spatial distance constraints and directional consistency constraints, accurately correlates the positioning points with the highway network, effectively improving the quality of trajectory data and the accuracy of road matching. This provides a reliable data foundation for subsequent trajectory analysis, path completion, and toll auditing.

[0074] S200: Project the BeiDou positioning trajectory point sequence onto the highway network, and calculate the cumulative deviation of the average speed between adjacent points relative to the historical speed limit, as the speed deviation factor for each trajectory interval.

[0075] After obtaining the BeiDou positioning trajectory point sequence, the technical problem to be solved in this step is how to extract features that reflect the actual driving state of the vehicle from the trajectory data. Directly using positioning points for path matching is easily affected by positioning errors and local anomalies, leading to a decrease in trajectory recovery accuracy; while analyzing only instantaneous speed is difficult to accurately reflect the overall driving pattern of the vehicle. Therefore, it is necessary to combine highway network and historical speed limit information to perform speed deviation analysis on the trajectory data, extract speed deviation factors that can characterize the vehicle's driving characteristics, and provide a reliable basis for subsequent trajectory optimization and path completion.

[0076] Step S200 in the method provided in this embodiment of the invention includes:

[0077] Each trajectory point in the BeiDou positioning trajectory point sequence is projected onto the highway network to obtain the corresponding road network projection points, and arranged in chronological order to form a road network projection point sequence.

[0078] Extract the road network travel distance and time difference between two adjacent projection points from the road network projection point sequence, calculate the average speed between the two projection points, and record it as the interval average speed.

[0079] For each interval, the historical speed limit of the road segment where the interval is located is obtained, and the speed deviation rate of a single interval is determined by combining the absolute value of the difference between the average speed of the interval and the historical speed limit.

[0080] Starting from the first interval of the road network projection point sequence, multiple consecutive intervals are grouped according to the direction of travel, and the number of consecutive intervals in each group is equal to the interval grouping base. The interval grouping base is dynamically determined based on the road network length between the starting gantry and the ending gantry and the total number of points in the road network projection point sequence.

[0081] Calculate the sum of the single-interval speed deviation rates of all intervals within each group, and use this as the cumulative deviation within the group.

[0082] The cumulative deviation within the group is used as the velocity deviation factor for the corresponding trajectory interval of the group.

[0083] In this embodiment, the interval grouping base refers to the number of intervals used to divide a continuous trajectory interval into several analysis groups. It is dynamically determined based on the road network length between the starting gantry and the ending gantry and the total number of points in the road network projection point sequence, so as to achieve adaptive grouping under different road segment lengths and sampling densities.

[0084] The speed deviation factor is a comprehensive value obtained by aggregating and calculating the speed deviation within a certain continuous interval. It is used to reflect the degree of abnormality in the overall driving speed of a local road segment.

[0085] In this step, each trajectory point in the BeiDou positioning trajectory point sequence is first projected onto the highway network to obtain the corresponding road network projection points, which are then arranged in chronological order to form a road network projection point sequence. By converting the original spatial positioning points into projection points in the road network topology, the spatial drift effect caused by BeiDou positioning errors can be effectively eliminated, allowing subsequent speed calculations to be based on real road structure constraints, thereby improving the accuracy of speed analysis.

[0086] Subsequently, the travel distance and time difference between two adjacent projection points are extracted from the road network projection point sequence, and the interval average speed is calculated based on the ratio of the two. The interval average speed characterizes the actual average travel speed of a vehicle within a corresponding road network interval, and its calculation method is as follows:

[0087] Interval average speed = Road network travel distance between two adjacent road network projection points ÷ Time difference between two adjacent road network projection points.

[0088] The road network travel distance is the actual road length traveled from the corresponding position of the previous projection point to the corresponding position of the next projection point along the highway network topology, rather than the straight-line distance between the two projection points. Therefore, it can more realistically reflect the actual travel distance of the vehicle in the road network.

[0089] For example, when the travel distance calculated along the highway network between two adjacent road network projection points is 3000m and the corresponding time difference is 120s, the average speed of the interval can be obtained according to the above calculation method as 25m / s; if converted to commonly used traffic speed units, it is approximately 90km / h.

[0090] Furthermore, for each interval, the historical speed limit value corresponding to the road segment containing that interval is obtained, and the absolute value of the difference between the average speed of the interval and the historical speed limit is used to determine the speed deviation rate of a single interval. The speed deviation rate of a single interval is used to quantify the degree of deviation of the actual vehicle speed from the standard speed limit of the corresponding road segment, and its calculation method is as follows:

[0091] Single-section speed deviation rate = |Average speed of section - Historical speed limit| ÷ Historical speed limit.

[0092] When the average speed of a section is closer to the historical speed limit, the speed deviation rate of a single section is smaller, indicating that the vehicle's driving status is more in line with normal traffic patterns; when the difference between the average speed of a section and the historical speed limit is larger, the speed deviation rate of a single section is larger, indicating that there is a higher possibility of abnormal speed behavior in that section.

[0093] Furthermore, starting from the first interval of the road network projection point sequence, multiple consecutive intervals are grouped according to the actual driving direction of the vehicle. The number of consecutive intervals contained in each group is equal to the interval grouping cardinality, which is dynamically determined based on the road network length between the starting and ending gantry and the total number of points in the road network projection point sequence. Through this dynamic grouping mechanism, the analysis granularity can be adaptively adjusted according to different road segment lengths and trajectory sampling densities, making the speed deviation analysis both locally sensitive and overall stable.

[0094] For example, when the road network length between the starting gantry and the ending gantry is long and the road network projection points are relatively sparse, the interval grouping cardinality is increased accordingly to enhance statistical stability; when the road network length is short or the sampling points are dense, the interval grouping cardinality is decreased accordingly to improve the sensitivity to local abnormal speed changes.

[0095] Finally, the sum of the single-interval speed deviation rates of all intervals within each group is calculated as the cumulative deviation within that group, and this cumulative deviation is used as the speed deviation factor for the corresponding trajectory interval. By grouping and accumulating the speed deviations of consecutive intervals, the interference caused by single-point abnormal fluctuations can be effectively suppressed, while the overall speed anomaly characteristics within the local time window can be enhanced. This allows the speed deviation factor to more accurately characterize the comprehensive driving behavior characteristics of vehicles in specific road segments and provides a reliable basis for subsequent path matching and trajectory verification.

[0096] It should be noted that the above values ​​are for illustrative purposes only and do not constitute a limitation on the present invention.

[0097] In summary, this step, by projecting BeiDou positioning trajectory points onto the highway network, calculates the average speed and speed deviation rate of each interval, and dynamically groups the data based on the network length and trajectory sampling density to obtain the speed deviation factor for the corresponding trajectory interval, effectively reduces the impact of a single abnormal positioning point on the speed analysis results, improves the ability to identify abnormal local driving states of vehicles, and provides a reliable basis for subsequent path matching and path completion screening.

[0098] S300: Based on the positional relationship of each trajectory point relative to the starting gantry, calculate the path completion characterization value, and combine it with the road network projection distance to remove outlier trajectory points and obtain an optimized trajectory point sequence;

[0099] After obtaining the velocity deviation factors for each trajectory interval, the next technical challenge is to identify and eliminate abnormal trajectory points affected by factors such as BeiDou positioning drift and positioning jumps. Directly using the original trajectory for path recovery is prone to reduced trajectory continuity and distorted matching results due to local anomalies; while filtering based solely on spatial location is insufficient to accurately reflect the vehicle's actual driving process. Therefore, it is necessary to optimize the trajectory by combining the positional relationship of trajectory points relative to the starting gantry and road network projection information to obtain a more realistic and reliable optimized trajectory point sequence, providing a stable data foundation for subsequent path recovery.

[0100] Step S300 in the method provided in this embodiment of the invention includes:

[0101] The starting gantry is projected onto the highway network to obtain the starting projection point, and the cumulative driving distance corresponding to the starting projection point is set to zero.

[0102] The cumulative driving distance of the first projection point in the road network projection point sequence is set as the shortest path length from the starting projection point to the first projection point in the highway network.

[0103] Traverse the sequence of projection points of the road network in chronological order. Starting from the second projection point, calculate the shortest path length between each projection point and the previous projection point in the highway network. Add the shortest path length to the cumulative travel distance corresponding to the previous projection point to obtain the cumulative travel distance of the current projection point.

[0104] The cumulative driving distance of each projection point is extracted from the road network projection point sequence and used as the initial path completion characterization value of the corresponding trajectory point.

[0105] The initial path completion value sequence is monotonically increased and corrected, including:

[0106] If the initial path completion value of the next projection point is less than the initial path completion value of the previous projection point, then the path completion value of the next projection point is replaced with the path completion value of the previous projection point.

[0107] Otherwise, retain the initial path completion value of the next projection point, and use the corrected result as the path completion value of the corresponding trajectory point.

[0108] In this embodiment, the initial path completion characterization value refers to the cumulative driving distance corresponding to the projection point, which is used to characterize the path progress that the vehicle has completed at the current moment.

[0109] The path completion rating refers to the path completion value after monotonically increasing correction. It is used to eliminate path backtracking caused by positioning errors and ensure that the vehicle's path progress remains non-decreasing over time.

[0110] In this step, the starting gantry is first projected onto the highway network to obtain the starting projection point, and the cumulative travel distance corresponding to the starting projection point is set to zero to establish a unified path distance reference coordinate system. Since the actual driving process of the vehicle in the missing road segment all starts from the starting gantry, initializing its cumulative travel distance to zero ensures that the path completion degree of all subsequent trajectory points is calculated based on the same reference benchmark.

[0111] Subsequently, the cumulative travel distance of the first projection point in the road network projection point sequence is set as the shortest path length from the starting projection point to the first projection point in the highway network. The shortest path length is calculated based on the highway network topology, and the calculation method is as follows:

[0112] The cumulative travel distance to the first projection point = the shortest path length between the starting projection point and the first projection point.

[0113] Furthermore, the sequence of road network projection points is traversed sequentially over time. Starting from the second projection point, the shortest path length between the current projection point and the previous projection point in the highway network is calculated sequentially, and then added to the cumulative travel distance corresponding to the previous projection point to obtain the cumulative travel distance corresponding to the current projection point. The calculation method is as follows:

[0114] The cumulative travel distance at the current projection point = the cumulative travel distance at the previous projection point + the shortest path length between the current projection point and the previous projection point.

[0115] By using a recursive accumulation method, the accumulated driving distance can be continuously increased along the actual driving path of the vehicle, thus more accurately reflecting the actual driving progress of the vehicle in the missing road segment.

[0116] For example, if the shortest path length between the starting projection point and the first projection point is 350m, then the cumulative travel distance corresponding to the first projection point is 350m; if the shortest path length between the first projection point and the second projection point is 420m, then the cumulative travel distance corresponding to the second projection point is 770m; if the shortest path length between the second projection point and the third projection point is 510m, then the cumulative travel distance corresponding to the third projection point is 1280m, and so on to complete the calculation of the cumulative distance of the entire road network projection point sequence.

[0117] Furthermore, the cumulative driving distance corresponding to each projection point is extracted from the road network projection point sequence as the initial path completion characterization value for the corresponding trajectory point. Since BeiDou positioning errors, positioning drift, or road matching errors may cause a local decrease in the cumulative driving distance corresponding to individual positioning points, i.e., a regression in vehicle path completion, and since actual vehicles do not experience regression along the time direction during normal highway driving, a monotonically increasing correction is required for the initial path completion characterization value sequence.

[0118] Specifically, when the initial path completion value corresponding to the next projection point is less than the initial path completion value corresponding to the previous projection point, the path completion value corresponding to the next projection point is replaced with the path completion value corresponding to the previous projection point; otherwise, the initial path completion value corresponding to the next projection point is retained. The correction process can be expressed as follows: if the current value is less than the previous value, the current value is the previous value; if the current value is greater than or equal to the previous value, the current value remains unchanged.

[0119] For example, if the initial path completion values ​​corresponding to four consecutive projection points are 1000m, 1380m, 1350m and 1720m respectively, then the third projection point will experience path regression due to positioning error, and its value will be less than 1380m corresponding to the previous projection point. Therefore, the path completion value of the third projection point will be corrected to 1380m, and the final corrected path completion value sequence will be 1000m, 1380m, 1380m and 1720m, thus ensuring that the path completion always maintains a monotonically non-decreasing change.

[0120] Finally, the result after monotonically increasing correction is used as the path completion value of the corresponding trajectory point.

[0121] It should be noted that the above values ​​are for illustrative purposes only and do not constitute a limitation on the present invention.

[0122] Step S300 in the method provided in this embodiment of the invention further includes:

[0123] Based on the calculated path completion characterization value, the path completion characterization value corresponding to each positioning point in the BeiDou positioning trajectory point sequence is extracted, and the path completion increment between adjacent positioning points is calculated.

[0124] The projection distance of each positioning point in the Beidou positioning trajectory point sequence onto the highway network, as well as the straight-line distance and time difference between adjacent positioning points, are obtained.

[0125] Location points that simultaneously satisfy both the first and second conditions are identified as outlier trajectory points and are removed.

[0126] The first condition is that the projected distance of the positioning point is greater than half of the road network driving distance between adjacent projection points;

[0127] The second condition is that at least one of the following two conditions must be met:

[0128] The initial value of the path completion characterization value of the positioning point before monotonically increasing correction is less than the initial path completion characterization value of the previous positioning point.

[0129] The path completion increment corresponding to the location point is greater than twice the product of the historical speed limit and the time difference between adjacent location points;

[0130] The remaining location points after removing outlier trajectory points are arranged in chronological order to form an optimized trajectory point sequence.

[0131] In this embodiment, out-of-trajectory points refer to positioning points whose positions deviate significantly from the vehicle's actual driving trajectory due to reasons such as BeiDou positioning drift, signal anomalies, multipath reflection, or time synchronization errors.

[0132] An optimized trajectory point sequence refers to the set of trajectory points that are rearranged in chronological order after outlier trajectory points are removed, which can more accurately reflect the actual driving trajectory of the target vehicle.

[0133] In this step, based on the already calculated path completion characterization value, the path completion characterization value corresponding to each positioning point in the BeiDou positioning trajectory point sequence is first extracted, and the path completion increment between adjacent positioning points is calculated. The path completion increment is calculated as follows:

[0134] Path completion increment = Path completion value of the current location point - Path completion value of the previous location point.

[0135] By calculating the path completion increment, we can reflect the actual distance a vehicle travels along the highway network within adjacent sampling periods, providing a basis for subsequent identification of abnormal trajectory jumps.

[0136] Subsequently, the projected distance of each positioning point in the BeiDou positioning trajectory point sequence is obtained after being projected onto the highway network. Simultaneously, the straight-line distance and time difference between adjacent positioning points are acquired, and outlier detection is performed on the positioning points using these parameters. Specifically, the projected distance of the positioning point is used to evaluate the spatial deviation between the positioning result and the actual road, the time difference reflects the sampling period length, and the path completion increment is used to evaluate the continuity of vehicle movement along the road direction.

[0137] Furthermore, positioning points that simultaneously satisfy both the first and second conditions are identified as outlier trajectory points and eliminated. The first condition is that the projected distance of a positioning point is greater than half the road network travel distance between adjacent projected points.

[0138] The reason for adopting this judgment condition is that when a vehicle is driving normally along the road, the positioning points are usually distributed near the road, and their distance from the road centerline is much smaller than the actual distance the vehicle travels along the road in adjacent sampling periods. When the projected distance of a positioning point exceeds half of the road network travel distance between adjacent projected points, it indicates that the degree of deviation of the positioning point from the road has reached a significant proportion of the actual distance traveled by the vehicle. It is highly likely to be a pseudo positioning point caused by positioning drift or abnormal jumps. Therefore, using half as the spatial anomaly judgment threshold can effectively identify obvious drift points and avoid mistakenly deleting real trajectory points due to normal positioning errors.

[0139] Furthermore, the second condition requires that at least one of the following two conditions be met.

[0140] First, the initial value of the path completion characterization value corresponding to the location point before monotonically increasing correction is less than the initial path completion characterization value corresponding to the previous location point. Since the cumulative completed path of a vehicle during normal driving on a highway should continuously increase over time and there should be no regression in path completion, when the current path completion is less than the path completion at the previous moment, it can be determined that there is a significant anomaly at the location point.

[0141] Secondly, the path completion increment corresponding to the location point is greater than twice the product of the historical speed limit and the time difference between adjacent location points, that is:

[0142] The path completion increment is greater than 2 × historical speed limit × time difference between adjacent positioning points.

[0143] The product of the historical speed limit and the time difference between adjacent positioning points represents the theoretical maximum normal driving distance that a vehicle can theoretically complete under the legal maximum speed conditions. When the increase in path completion exceeds twice this theoretical distance, it indicates that the vehicle needs to complete the corresponding displacement at a speed far exceeding the road's permissible speed. This situation is almost impossible in actual highway operation and is usually caused by positioning jumps, positioning drift, or time synchronization anomalies. Therefore, using twice the theoretical maximum normal driving distance as the judgment threshold can effectively identify large path jumps caused by positioning anomalies, while retaining trajectory changes caused by normal driving behavior.

[0144] For example, if the historical speed limit for a certain highway section is 120 km / h and the time difference between adjacent positioning points is 5 seconds, then a vehicle can theoretically travel approximately 166.7 meters under the historical speed limit conditions, and twice that is approximately 333.4 meters. If the calculated path completion increment reaches 580 meters, it significantly exceeds the theoretical normal range, indicating an abnormal jump at this positioning point. If simultaneously, the projected distance of this positioning point reaches 210 meters, while the road network travel distance between corresponding adjacent projected points is only 360 meters, then the projected distance has exceeded half of the road network travel distance. Both conditions are met simultaneously, therefore this positioning point is identified as an outlier and removed.

[0145] Finally, the remaining location points after removing outlier trajectory points are rearranged in chronological order to form an optimized trajectory point sequence.

[0146] It should be noted that the above values ​​are for illustrative purposes only and do not constitute a limitation on the present invention.

[0147] In summary, this step constructs a path completion characterization value and performs monotonically increasing correction. Combined with multi-dimensional constraints such as the positioning point projection distance and path completion increment, it identifies and eliminates outlier trajectory points. This effectively eliminates the impact of abnormal factors such as BeiDou positioning drift and positioning jump on trajectory analysis, improves the continuity and authenticity of trajectory data, and provides reliable data support for subsequent candidate path generation and actual driving path recovery.

[0148] S400: Taking the starting gantry as the starting point and the ending gantry as the ending point, match each trajectory point in the optimized trajectory point sequence to an arc segment in the road network that satisfies the projection distance constraint and the direction consistency constraint, and connect them in the matching order to generate at least one candidate driving path;

[0149] After obtaining the path completion characterization value calculated based on the positional relationship of each trajectory point relative to the starting gantry, a consistency constraint analysis is performed on the trajectory points in conjunction with the road network projection distance to eliminate abnormal trajectory points that deviate from the road network structure or are caused by positioning noise, thereby obtaining an optimized trajectory point sequence with better spatial continuity. Furthermore, this optimized trajectory point sequence is still in discrete coordinate form and has not yet established a topological correspondence with specific road network arc segments. Therefore, it needs to be used as input, and under the conditions of satisfying projection distance constraints and direction consistency constraints, mapped to the corresponding arc segments in the road network to achieve matching of trajectory points to structured road segments, and connected according to the matching order to generate candidate driving paths.

[0150] Step S400 in the method provided in this embodiment of the invention includes:

[0151] Project each trajectory point in the optimized trajectory point sequence onto the highway network, and select arc segments that satisfy the projection distance constraint and whose arc direction and driving direction satisfy the direction consistency constraint as candidate arc segments of the current trajectory point;

[0152] The projection distance constraint is that the projection distance of the positioning point is no greater than one-third of the straight-line distance between adjacent trajectory points;

[0153] The directional consistency constraint is that the angle between the direction of the arc segment and the direction of the line connecting the current trajectory point to the next trajectory point is no greater than 90 degrees, and the angle between the direction of the connecting arc segment inserted in the shortest path and the direction of travel is no greater than 90 degrees.

[0154] For two adjacent trajectory points, iterate through all combinations of candidate arc segments of the previous trajectory point and candidate arc segments of the next trajectory point. If the two arc segments are directly connected in the highway network, then establish a connection relationship.

[0155] If the two arc segments are not directly connected, the shortest path length between the two arc segments is calculated. When the shortest path length does not exceed the upper limit determined based on the historical speed limit and the time difference between the adjacent trajectory points, and the connecting path satisfies the directional consistency constraint, a connection relationship is established.

[0156] The starting gantry and the ending gantry are projected onto the highway network, and the arc segment with the smallest projection distance is taken as the starting arc segment and the ending arc segment.

[0157] Starting from the initial arc segment, the established adjacent connections are used to connect them step by step in chronological order until the termination arc segment is reached, generating all connected paths from the initial arc segment to the termination arc segment as candidate travel paths.

[0158] In this embodiment, the projection distance constraint specifies that the projection distance between the designated point and the candidate arc segment should meet the constraint condition of not being greater than one-third of the straight-line distance between adjacent trajectory points, in order to ensure that the positioning point and the matching road have a sufficiently high spatial consistency.

[0159] The direction consistency constraint means that the direction of the candidate arc segment is consistent with the actual driving direction of the vehicle. That is, the angle between the direction of the arc segment and the direction of the line connecting the current trajectory point to the next trajectory point is no greater than 90 degrees. At the same time, the angle between the direction of all connecting arc segments inserted in the connecting path and the driving direction of the vehicle is no greater than 90 degrees. This is used to ensure that the path topology conforms to the normal driving law of the vehicle.

[0160] In this step, each trajectory point in the optimized trajectory point sequence is first projected onto the highway network, and road arcs that satisfy the projection distance constraint and the arc direction constraint satisfy the direction consistency constraint are selected as candidate arcs corresponding to the current trajectory point.

[0161] The projection distance constraint stipulates that the projection distance corresponding to the positioning point must not exceed one-third of the straight-line distance between adjacent trajectory points. This constraint is adopted because, when a vehicle is traveling normally along a highway, the BeiDou positioning error is usually much smaller than the vehicle's actual displacement within adjacent sampling periods. When the deviation distance between the positioning point and the road exceeds one-third of the displacement distance of adjacent trajectory points, it indicates that the positioning point is highly likely to be affected by factors such as positioning drift or multipath reflection. Continuing to participate in path matching can easily lead to mismatches. Therefore, this constraint can effectively ensure spatial matching accuracy.

[0162] Subsequently, directional consistency constraints are further applied to the candidate arc segments that satisfy the projection distance constraint for screening.

[0163] Specifically, the direction of the candidate arc segment must be no greater than 90 degrees from the direction of the line connecting the current trajectory point to the next trajectory point. Furthermore, the direction of all connecting arc segments inserted in the subsequent shortest path must be no greater than 90 degrees from the direction of the vehicle's travel direction. Since vehicles do not move in the opposite direction along the road during normal highway travel, the direction consistency constraint can effectively eliminate oncoming lanes, reverse ramps, and roads with significantly opposite directions, improving the reliability of the trajectory matching results.

[0164] Furthermore, for two adjacent trajectory points, all combinations between the candidate arc segments corresponding to the previous trajectory point and the candidate arc segments corresponding to the next trajectory point are traversed. When the two candidate arc segments are directly connected in the highway network, a connection relationship is directly established; when the two candidate arc segments are not directly connected, the shortest path length between the two candidate arc segments is further calculated, and it is determined whether the shortest path length satisfies the length constraint and the direction consistency constraint.

[0165] The upper limit value corresponding to the length constraint is dynamically determined based on the time difference between the historical speed limit and adjacent trajectory points, and its calculation method is as follows:

[0166] Upper limit = 2 × historical speed limit × time difference between adjacent trajectory points.

[0167] The reason for adopting the above constraints is that the product of the historical speed limit and the time difference between adjacent trajectory points represents the maximum normal driving distance that a vehicle can theoretically complete under the legal maximum driving speed. Considering factors such as positioning sampling error, speed fluctuation, and road topology matching error, this embodiment further introduces a safety margin of twice as the upper limit of the connection length. When the calculated shortest path length exceeds this upper limit, it means that the vehicle needs to complete the corresponding distance at a speed significantly exceeding the road's allowed speed, which usually does not conform to actual driving patterns, so no corresponding connection relationship is established; conversely, it is considered that there is a reasonable possibility of connection between the two candidate arc segments, and a connection relationship can be established.

[0168] For example, when the historical speed limit of a certain highway section is 120 km / h and the time difference between adjacent trajectory points is 6 seconds, the theoretical maximum normal driving distance is approximately 200 m, and the upper limit of the connection length corresponding to this embodiment is approximately 400 m. If the calculated shortest path length between two candidate arc segments is 310 m, the length constraint is satisfied, and a connection relationship can be established; if the shortest path length reaches 580 m, it exceeds the preset upper limit value, indicating that the corresponding connection requires the vehicle to complete an abnormally long distance movement in a short period of time, and therefore a connection relationship is not established.

[0169] Finally, the starting gantry and the ending gantry are projected onto the highway network, and the arc segments with the smallest projection distance are selected as the starting arc segment and the ending arc segment, respectively. Then, starting from the starting arc segment, the path is expanded step by step according to the time sequence of the trajectory points, using the established adjacent connection relationship, until the ending arc segment is reached, generating all connected paths from the starting arc segment to the ending arc segment, and these paths are used as candidate driving paths.

[0170] It should be noted that the above values ​​are for illustrative purposes only and do not constitute a limitation on the present invention.

[0171] In summary, this step combines projection distance constraints, directional consistency constraints, and connection length constraints to filter and connect candidate arc segments corresponding to the optimized trajectory points, and generates candidate driving paths between the starting arc segment and the ending arc segment based on the road network topology. This effectively eliminates abnormal connections that do not conform to the actual driving patterns of vehicles, improves the authenticity and reliability of candidate paths, and provides a reliable foundation for subsequent path completion determination.

[0172] S500: Based on the time deviation, arc length, and velocity deviation of the corresponding trajectory interval in each candidate path, calculate the comprehensive consistency cost of each path and select the candidate path with the lowest cost as the completion path.

[0173] After generating multiple candidate driving paths that match the road network arcs, although each path satisfies the constraints in terms of spatial topology, deviations may still exist in terms of temporal continuity and driving state consistency. Therefore, it is necessary to further comprehensively evaluate the candidate paths from dimensions such as time deviation, arc length, and speed deviation, calculate the comprehensive consistency cost of each path, and select the optimal path accordingly.

[0174] Step S500 in the method provided in this embodiment of the invention includes:

[0175] For each candidate path, extract the arc segment sequence that constitutes the candidate path;

[0176] The arc length of each arc in the arc sequence is obtained, and the time deviation cost is determined by combining the actual travel time between the starting gantry and the ending gantry. The time deviation cost is positively correlated with the degree to which the arc sequence deviates from the expected travel time.

[0177] Based on the sum of the arc lengths of each arc in the arc sequence, and combined with the shortest path length from the starting gantry to the ending gantry in the highway network, the path length cost is determined, wherein the path length cost is positively correlated with the degree to which the sum of the arc lengths deviates from the shortest path length;

[0178] Extract the trajectory intervals corresponding to the optimized trajectory point sequence covered by the candidate path, obtain the speed deviation factor of the corresponding trajectory interval, and determine the speed deviation cost, wherein the speed deviation cost is positively correlated with the central tendency of the speed deviation factor;

[0179] By integrating the time deviation cost, the path length cost, and the speed deviation cost, the comprehensive consistency cost of the candidate path is obtained.

[0180] The overall consistency cost of all candidate paths is compared, and the candidate path with the smallest overall consistency cost is selected as the completion path.

[0181] In this step, firstly, for each candidate path, the arc segment sequence constituting the candidate path is extracted, and the arc length corresponding to each arc segment and the historical speed limit information of the road segment are obtained.

[0182] Furthermore, the actual travel time for the target vehicle is calculated based on the termination gantry transaction time and the initiation gantry transaction time. The calculation method is as follows:

[0183] Actual travel time = Ending gantry transaction time - Starting gantry transaction time.

[0184] Subsequently, the theoretical travel time is calculated for each arc segment in the candidate path. This is achieved by dividing the arc segment length by the historical speed limit of the corresponding road segment to obtain the expected travel time for each arc segment. The expected travel times for all arc segments are then summed to obtain the total expected travel time for the path. The calculation method is as follows:

[0185] Expected total time for the route = Σ (length of arc segment ÷ historical speed limit of the road segment containing the arc segment).

[0186] After obtaining the actual travel time and the expected total route time, the time deviation cost is calculated based on the degree of deviation between the two. The calculation method is as follows:

[0187] Time deviation cost = |Actual travel time - Expected total route time| ÷ Expected total route time.

[0188] The smaller the time deviation cost, the closer the theoretical travel time of the candidate path is to the actual travel time of the vehicle, and the higher the authenticity of the path; conversely, it indicates that there is a large difference between the candidate path and the actual vehicle operation.

[0189] Furthermore, the total length of the candidate path is calculated based on the lengths of all arc segments in the arc segment sequence, and the calculation method is as follows:

[0190] Total length of candidate paths = Σ arc length.

[0191] Subsequently, the total length of the candidate paths is compared with the shortest path length from the starting gantry to the ending gantry in the highway network, and the path length cost is calculated as follows:

[0192] Path length cost = Total length of candidate paths ÷ Shortest path length from starting gantry to ending gantry in the highway network.

[0193] The above calculations can be used to evaluate the degree of detour by the candidate path relative to the theoretical shortest path. When the path length cost is closer to 1, it indicates that the candidate path is closer to the shortest reasonable path; when the path length cost increases significantly, it indicates that there is significant detour behavior.

[0194] Subsequently, the trajectory intervals corresponding to the optimized trajectory point sequences covered by the candidate paths are extracted, and the calculated velocity deviation factors for the corresponding trajectory intervals are obtained. The velocity deviation cost is obtained by summing all velocity deviation factors and dividing by the number of covered trajectory intervals. The calculation method is as follows:

[0195] Speed ​​deviation cost = Sum of speed deviation factors corresponding to the trajectory intervals covered by the candidate path ÷ Number of trajectory intervals covered.

[0196] That is, the arithmetic mean of all speed deviation factors within the coverage area of ​​the candidate path is used as the speed deviation cost corresponding to the entire candidate path, thereby comprehensively reflecting the overall speed stability of the vehicle on the candidate path.

[0197] Finally, by integrating the time deviation cost, path length cost, and speed deviation cost, the comprehensive consistency cost corresponding to the candidate path is obtained, and its calculation method is as follows:

[0198] Overall consistency cost = time deviation cost + path length cost + speed deviation cost.

[0199] For example, if the actual travel time of the target vehicle through the starting and ending gantries is 600s, and a candidate path consists of three arc segments with theoretical travel times of 180s, 180s, and 252s respectively, then the expected total travel time is 612s. Based on the above calculation, the time deviation cost is approximately 0.0196 (|600 - 612| ÷ 612). Further, if the total length of the candidate path is 18000m, and the shortest path length from the starting to the ending gantries in the highway network is 17500m, then the path length cost is approximately 18000 ÷ 17500 ≈ 1.0286. If the candidate path covers four trajectory intervals with corresponding speed deviation factors of 0.12, 0.08, 0.15, and 0.09, then the speed deviation cost is approximately 0.11 ((0.12 + 0.08 + 0.15 + 0.09) ÷ 4). Finally, based on the above calculation method, the overall consistency cost is approximately 0.0196 + 1.0286 + 0.11 ≈ 1.1582.

[0200] It should be noted that the above values ​​are for illustrative purposes only and do not constitute a limitation on the present invention.

[0201] In summary, this step comprehensively considers the time deviation, path length, and speed deviation of candidate paths to construct a comprehensive consistency cost to evaluate each candidate path, and selects the candidate path with the minimum comprehensive consistency cost as the completion path. This achieves accurate restoration of the vehicle's actual driving trajectory, improves the authenticity of the path completion results, and enhances the accuracy of billing and auditing.

[0202] S600: Output the cumulative mileage of the completed path as the actual mileage of the target vehicle in the missing road segment, which is used to replace the default path to complete the billing audit.

[0203] Step S600 in the method provided in this embodiment of the invention includes:

[0204] Extract the arc segment sequence contained in the completed path, and sum the lengths of each arc segment to obtain the cumulative mileage of the completed path;

[0205] Extract the path completion characterization values ​​of the first and last trajectory points in the optimized trajectory point sequence, and calculate the difference between the path completion characterization value of the last trajectory point and the path completion characterization value of the first trajectory point, which is used as the trajectory reference mileage;

[0206] Compare the cumulative mileage of the completed path with the trajectory reference mileage. If the absolute value of the difference between the two does not exceed the deviation range determined based on the straight-line distance between adjacent trajectory points in the optimized trajectory point sequence, then the cumulative mileage of the completed path is taken as the actual driving mileage.

[0207] Otherwise, the reference mileage of the trajectory will be taken as the actual mileage traveled;

[0208] Replace the default shortest path mileage between the starting gantry and the ending gantry with the actual mileage to complete the initial billing audit.

[0209] The original toll mileage of the target vehicle between the starting gantry and the ending gantry is obtained, and the ratio of the absolute value of the difference between the actual mileage and the toll mileage to the toll mileage is calculated as the mileage difference rate.

[0210] When the mileage difference rate does not exceed the difference threshold determined based on the road network length between the starting gantry and the ending gantry, the billing mileage of the target vehicle is updated with the actual mileage traveled to complete the secondary audit correction.

[0211] When the mileage difference rate exceeds the difference threshold, the path completion value sequence and speed deviation factor sequence corresponding to the optimized trajectory point sequence are extracted. The path completion value sequence, the speed deviation factor sequence and the completed path are used to generate an audit report, which is then submitted for manual review.

[0212] In this embodiment, the difference threshold refers to the allowable deviation range determined based on the road network length between the starting gantry and the ending gantry. It is obtained by multiplying the road network length by a specified proportional coefficient, where the specified proportional coefficient is a fixed value used to balance automatic processing efficiency and anomaly identification sensitivity.

[0213] In this step, the arc segment sequence contained in the completed path is first extracted, and the lengths of each arc segment are summed to obtain the cumulative mileage of the completed path. The calculation method is as follows:

[0214] The total mileage of the completed path is equal to the length of the arc segment.

[0215] Subsequently, the path completion characterization values ​​corresponding to the first and last trajectory points in the optimized trajectory point sequence are extracted, and the difference between the two is calculated as the trajectory reference mileage. The calculation method is as follows:

[0216] Trajectory reference mileage = Path completion value of the last trajectory point - Path completion value of the first trajectory point.

[0217] Furthermore, the straight-line distances between all adjacent trajectory points in the optimized trajectory point sequence are statistically analyzed, and the median is taken as the deviation range. Using the median instead of the average can reduce the impact of a small number of abnormal trajectory points on the results, making the deviation range more stable and reliable.

[0218] Subsequently, the cumulative mileage of the completed path is compared with the trajectory reference mileage. If the absolute value of the difference between the two does not exceed the aforementioned deviation range, the completed path is considered to have a high degree of consistency with the actual trajectory, and the cumulative mileage of the completed path is taken as the actual mileage; otherwise, the trajectory reference mileage is taken as the actual mileage. The determination relationship can be expressed as follows:

[0219] If |Completed Path Cumulative Mileage - Trajectory Reference Mileage| ≤ Deviation Range, then Actual Mileage = Completed Path Cumulative Mileage; otherwise, Actual Mileage = Trajectory Reference Mileage.

[0220] For example, when the cumulative mileage of the completed path is 20560m, the trajectory reference mileage is 20510m, and the median straight-line distance between adjacent trajectory points in the optimized trajectory point sequence is 80m, since the difference between the two is 50m, which is less than the deviation range of 80m, 20560m is used as the actual driving mileage.

[0221] Furthermore, the determined actual mileage is used to replace the billing mileage corresponding to the default shortest path between the starting and ending gantries, completing the initial billing audit; subsequently, the original billing mileage of the target vehicle is obtained, and the mileage difference rate between the actual mileage and the billing mileage is calculated, as follows:

[0222] Mileage difference rate = |Actual mileage - Billing mileage| ÷ Billing mileage.

[0223] Meanwhile, the difference threshold is calculated based on the road network length between the starting gantry and the ending gantry, and the calculation method is as follows:

[0224] Difference threshold = Network length between the starting gantry and the ending gantry × Specified scaling factor.

[0225] The specified proportional coefficient uses a fixed value, such as 5%. Since the longer the road network, the greater the absolute value of the cumulative error that can exist under normal circumstances, the combination of road network length and fixed proportional coefficient can adapt to the tolerance range of mileage deviation under different missing road segment lengths, thereby improving the rationality of automatic auditing.

[0226] Finally, the calculated mileage difference rate is compared with the difference threshold. When the mileage difference rate does not exceed the difference threshold, the billing mileage of the target vehicle is updated using the actual mileage traveled, completing the secondary audit correction. When the mileage difference rate exceeds the difference threshold, the path completion characterization value sequence and speed deviation factor sequence corresponding to the optimized trajectory point sequence are extracted, and combined with the completed path to generate an audit report, which is then submitted for manual review.

[0227] It should be noted that the above values ​​are for illustrative purposes only and do not constitute a limitation on the present invention.

[0228] In summary, this step verifies the consistency between the cumulative mileage of the completed path and the reference mileage of the trajectory, and determines the actual mileage by combining the deviation range. On this basis, it further utilizes the mileage difference rate to complete automatic billing audit and secondary correction. For abnormal situations exceeding the difference threshold, an audit report is generated and submitted for manual review. This realizes a billing correction mechanism that combines automatic audit and manual review, improving the accuracy of actual mileage determination and the reliability of billing audit results.

[0229] Example 2, as Figure 3 As shown, this invention provides a vehicle trajectory auditing system based on the fusion of BeiDou positioning and ETC data. The system includes:

[0230] The trajectory data acquisition module 11 is used to extract the Beidou positioning trajectory point sequence of the target vehicle between the starting gantry and the ending gantry when there is a missing ETC gantry transaction record in the road segment through which the target vehicle passes.

[0231] The speed deviation analysis module 12 is used to project the Beidou positioning trajectory point sequence onto the highway network and calculate the cumulative deviation of the average speed between adjacent points relative to the historical speed limit, as the speed deviation factor for each trajectory interval.

[0232] The trajectory optimization module 13 is used to calculate the path completion characterization value based on the positional relationship of each trajectory point relative to the starting gantry, and combine it with the road network projection distance to remove outlier trajectory points and obtain an optimized trajectory point sequence.

[0233] The candidate path generation module 14 is used to match each trajectory point in the optimized trajectory point sequence to an arc segment in the road network that satisfies the projection distance constraint and the direction consistency constraint, with the starting gantry as the starting point and the ending gantry as the ending point, and connect them in the matching order to generate at least one candidate driving path.

[0234] The path completion module 15 is used to calculate the comprehensive consistency cost of each path based on the time deviation, arc length and speed deviation of the corresponding trajectory interval in each candidate path, and select the candidate path with the minimum cost as the completion path.

[0235] The billing audit module 16 is used to output the cumulative mileage of the completed path as the actual mileage of the target vehicle in the missing road segment, and is used to replace the default path to complete the billing audit.

[0236] In this embodiment of the invention, the trajectory data acquisition module 11 is further configured to:

[0237] A time window is defined based on the transaction time of the starting gantry and the transaction time of the ending gantry, and the original BeiDou positioning data of the target vehicle within the time window is extracted.

[0238] Redundant positioning points with inverted or duplicate timestamps in the original BeiDou positioning data are removed, and the times of the remaining positioning points are uniformly transformed to the time reference of the gantry device.

[0239] The average velocity between adjacent points and the straight-line distance between adjacent points are calculated sequentially for the time-aligned positioning point sequence, and each positioning point is projected onto the highway network to obtain the projected distance of each point;

[0240] Location points that meet any of the following conditions will be removed: the average speed between two adjacent points is greater than twice the historical speed limit, or the projected distance of the location point is greater than one-third of the straight-line distance between two adjacent points.

[0241] The remaining positioning points are arranged in chronological order to form the BeiDou positioning trajectory point sequence.

[0242] The process of projecting each location point onto the highway network to obtain the projected distance of each point includes:

[0243] Based on the highway network between the starting gantry and the ending gantry, all arc segments are extracted to form a candidate arc segment set.

[0244] For each positioning point, the straight-line distance between the positioning point and the adjacent positioning point is taken as the search radius. In the candidate arc segment set, arc segments whose vertical distance to the positioning point does not exceed the search radius are selected to form an initial set of arc segments.

[0245] In the initial screening arc segment set, calculate the direction angle between the direction of each arc segment and the direction of the line connecting two adjacent points of the positioning point, remove arc segments with a direction angle greater than 90 degrees, and take the value with the smallest vertical distance among the remaining arc segments as the projected distance of the positioning point.

[0246] If the directional angle of all arc segments in the initial screening arc segment set is greater than 90 degrees, then the minimum vertical distance in the initial screening arc segment set is taken as the projected distance of the positioning point.

[0247] In this embodiment of the invention, the speed deviation analysis module 12 is further used for:

[0248] Each trajectory point in the BeiDou positioning trajectory point sequence is projected onto the highway network to obtain the corresponding road network projection points, and arranged in chronological order to form a road network projection point sequence.

[0249] Extract the road network travel distance and time difference between two adjacent projection points from the road network projection point sequence, calculate the average speed between the two projection points, and record it as the interval average speed.

[0250] For each interval, the historical speed limit of the road segment where the interval is located is obtained, and the speed deviation rate of a single interval is determined by combining the absolute value of the difference between the average speed of the interval and the historical speed limit.

[0251] Starting from the first interval of the road network projection point sequence, multiple consecutive intervals are grouped according to the direction of travel, and the number of consecutive intervals in each group is equal to the interval grouping base. The interval grouping base is dynamically determined based on the road network length between the starting gantry and the ending gantry and the total number of points in the road network projection point sequence.

[0252] Calculate the sum of the single-interval speed deviation rates of all intervals within each group, and use this as the cumulative deviation within the group.

[0253] The cumulative deviation within the group is used as the velocity deviation factor for the corresponding trajectory interval of the group.

[0254] In this embodiment of the invention, the trajectory optimization module 13 is further configured to:

[0255] The starting gantry is projected onto the highway network to obtain the starting projection point, and the cumulative driving distance corresponding to the starting projection point is set to zero.

[0256] The cumulative driving distance of the first projection point in the road network projection point sequence is set as the shortest path length from the starting projection point to the first projection point in the highway network.

[0257] Traverse the sequence of projection points of the road network in chronological order. Starting from the second projection point, calculate the shortest path length between each projection point and the previous projection point in the highway network. Add the shortest path length to the cumulative travel distance corresponding to the previous projection point to obtain the cumulative travel distance of the current projection point.

[0258] The cumulative driving distance of each projection point is extracted from the road network projection point sequence and used as the initial path completion characterization value of the corresponding trajectory point.

[0259] The initial path completion value sequence is monotonically increased and corrected, including:

[0260] If the initial path completion value of the next projection point is less than the initial path completion value of the previous projection point, then the path completion value of the next projection point is replaced with the path completion value of the previous projection point.

[0261] Otherwise, retain the initial path completion value of the next projection point, and use the corrected result as the path completion value of the corresponding trajectory point.

[0262] The steps for obtaining the optimized trajectory point sequence include:

[0263] Based on the calculated path completion characterization value, the path completion characterization value corresponding to each positioning point in the BeiDou positioning trajectory point sequence is extracted, and the path completion increment between adjacent positioning points is calculated.

[0264] The projection distance of each positioning point in the Beidou positioning trajectory point sequence onto the highway network, as well as the straight-line distance and time difference between adjacent positioning points, are obtained.

[0265] Location points that simultaneously satisfy both the first and second conditions are identified as outlier trajectory points and are removed.

[0266] The first condition is that the projected distance of the positioning point is greater than half of the road network driving distance between adjacent projection points;

[0267] The second condition is that at least one of the following two conditions must be met:

[0268] The initial value of the path completion characterization value of the positioning point before monotonically increasing correction is less than the initial path completion characterization value of the previous positioning point.

[0269] The path completion increment corresponding to the location point is greater than twice the product of the historical speed limit and the time difference between adjacent location points;

[0270] The remaining location points after removing outlier trajectory points are arranged in chronological order to form an optimized trajectory point sequence.

[0271] In this embodiment of the invention, the candidate path generation module 14 is further configured to:

[0272] Project each trajectory point in the optimized trajectory point sequence onto the highway network, and select arc segments that satisfy the projection distance constraint and whose arc direction and driving direction satisfy the direction consistency constraint as candidate arc segments of the current trajectory point;

[0273] The projection distance constraint is that the projection distance of the positioning point is no greater than one-third of the straight-line distance between adjacent trajectory points;

[0274] The directional consistency constraint is that the angle between the direction of the arc segment and the direction of the line connecting the current trajectory point to the next trajectory point is no greater than 90 degrees, and the angle between the direction of the connecting arc segment inserted in the shortest path and the direction of travel is no greater than 90 degrees.

[0275] For two adjacent trajectory points, iterate through all combinations of candidate arc segments of the previous trajectory point and candidate arc segments of the next trajectory point. If the two arc segments are directly connected in the highway network, then establish a connection relationship.

[0276] If the two arc segments are not directly connected, the shortest path length between the two arc segments is calculated. When the shortest path length does not exceed the upper limit determined based on the historical speed limit and the time difference between the adjacent trajectory points, and the connecting path satisfies the directional consistency constraint, a connection relationship is established.

[0277] The starting gantry and the ending gantry are projected onto the highway network, and the arc segment with the smallest projection distance is taken as the starting arc segment and the ending arc segment.

[0278] Starting from the initial arc segment, the established adjacent connections are used to connect them step by step in chronological order until the termination arc segment is reached, generating all connected paths from the initial arc segment to the termination arc segment as candidate travel paths.

[0279] In this embodiment of the invention, the path completion module 15 is further used for:

[0280] For each candidate path, extract the arc segment sequence that constitutes the candidate path;

[0281] The arc length of each arc in the arc sequence is obtained, and the time deviation cost is determined by combining the actual travel time between the starting gantry and the ending gantry. The time deviation cost is positively correlated with the degree to which the arc sequence deviates from the expected travel time.

[0282] Based on the sum of the arc lengths of each arc in the arc sequence, and combined with the shortest path length from the starting gantry to the ending gantry in the highway network, the path length cost is determined, wherein the path length cost is positively correlated with the degree to which the sum of the arc lengths deviates from the shortest path length;

[0283] Extract the trajectory intervals corresponding to the optimized trajectory point sequence covered by the candidate path, obtain the speed deviation factor of the corresponding trajectory interval, and determine the speed deviation cost, wherein the speed deviation cost is positively correlated with the central tendency of the speed deviation factor;

[0284] By integrating the time deviation cost, the path length cost, and the speed deviation cost, the comprehensive consistency cost of the candidate path is obtained.

[0285] The overall consistency cost of all candidate paths is compared, and the candidate path with the smallest overall consistency cost is selected as the completion path.

[0286] In this embodiment of the invention, the billing audit module 16 is further configured to:

[0287] Extract the arc segment sequence contained in the completed path, and sum the lengths of each arc segment to obtain the cumulative mileage of the completed path;

[0288] Extract the path completion characterization values ​​of the first and last trajectory points in the optimized trajectory point sequence, and calculate the difference between the path completion characterization value of the last trajectory point and the path completion characterization value of the first trajectory point, which is used as the trajectory reference mileage;

[0289] Compare the cumulative mileage of the completed path with the trajectory reference mileage. If the absolute value of the difference between the two does not exceed the deviation range determined based on the straight-line distance between adjacent trajectory points in the optimized trajectory point sequence, then the cumulative mileage of the completed path is taken as the actual driving mileage.

[0290] Otherwise, the reference mileage of the trajectory will be taken as the actual mileage traveled;

[0291] Replace the default shortest path mileage between the starting gantry and the ending gantry with the actual mileage to complete the initial billing audit.

[0292] The original toll mileage of the target vehicle between the starting gantry and the ending gantry is obtained, and the ratio of the absolute value of the difference between the actual mileage and the toll mileage to the toll mileage is calculated as the mileage difference rate.

[0293] When the mileage difference rate does not exceed the difference threshold determined based on the road network length between the starting gantry and the ending gantry, the billing mileage of the target vehicle is updated with the actual mileage traveled to complete the secondary audit correction.

[0294] When the mileage difference rate exceeds the difference threshold, the path completion value sequence and speed deviation factor sequence corresponding to the optimized trajectory point sequence are extracted. The path completion value sequence, the speed deviation factor sequence and the completed path are used to generate an audit report, which is then submitted for manual review.

[0295] In summary, the vehicle trajectory auditing system based on the fusion of BeiDou positioning and ETC data provided by this invention completes trajectory data acquisition, speed deviation analysis, trajectory optimization, candidate path generation, path completion, and billing auditing through the collaborative efforts of various modules. It achieves accurate recovery of the actual driving trajectory of the vehicle and intelligent correction of the billing results in scenarios where ETC transaction records are missing, thereby improving the accuracy of trajectory completion, billing accuracy, and anomaly identification capabilities.

[0296] It should be noted that the processes depicted in the accompanying drawings do not necessarily require a specific or sequential order to achieve the desired result. In some implementations, multitasking and parallel processing are possible or may be advantageous.

[0297] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. 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.

[0298] This specification and accompanying drawings are merely illustrative examples of the invention and are intended to cover any and all modifications, variations, combinations, or equivalents within the scope of the invention. Clearly, those skilled in the art can make various alterations and modifications to the invention without departing from its scope. Therefore, if such modifications and modifications fall within the scope of the invention and its equivalents, the invention is intended to include these modifications and modifications.

Claims

1. A vehicle trajectory auditing method based on the fusion of BeiDou positioning and ETC data, characterized in that, The method includes: When there are missing ETC gantry transaction records on the road segment through which the target vehicle passes, the gantry with a transaction record at the last upstream position of the missing road segment is taken as the starting gantry and the gantry with a transaction record at the first downstream position is taken as the ending gantry. The Beidou positioning trajectory point sequence of the target vehicle between the starting gantry and the ending gantry is extracted. The BeiDou positioning trajectory point sequence is projected onto the highway network, and the cumulative deviation of the average speed between adjacent points relative to the historical speed limit is calculated as the speed deviation factor for each trajectory interval. Based on the positional relationship of each trajectory point relative to the starting gantry, the path completion characterization value is calculated, and combined with the road network projection distance, outlier trajectory points are removed to obtain an optimized trajectory point sequence; Starting from the starting gantry and ending from the ending gantry, each trajectory point in the optimized trajectory point sequence is matched to an arc segment in the road network that satisfies the projection distance constraint and the direction consistency constraint, and connected in the matching order to generate at least one candidate driving path. Based on the time deviation, arc length, and velocity deviation of the corresponding trajectory intervals in each candidate path, the comprehensive consistency cost of each path is calculated, and the candidate path with the lowest cost is selected as the completion path. The cumulative mileage of the completed path is output as the actual mileage of the target vehicle in the missing road segment, which is used to replace the default path to complete the billing audit.

2. The vehicle trajectory auditing method based on the fusion of BeiDou positioning and ETC data according to claim 1, characterized in that, The steps for obtaining the BeiDou positioning trajectory point sequence include: A time window is defined based on the transaction time of the starting gantry and the transaction time of the ending gantry, and the original BeiDou positioning data of the target vehicle within the time window is extracted. Redundant positioning points with inverted or duplicate timestamps in the original BeiDou positioning data are removed, and the times of the remaining positioning points are uniformly transformed to the time reference of the gantry device. The average velocity between adjacent points and the straight-line distance between adjacent points are calculated sequentially for the time-aligned positioning point sequence, and each positioning point is projected onto the highway network to obtain the projected distance of each point; Location points that meet any of the following conditions will be removed: the average speed between two adjacent points is greater than twice the historical speed limit, or the projected distance of the location point is greater than one-third of the straight-line distance between two adjacent points. The remaining positioning points are arranged in chronological order to form the BeiDou positioning trajectory point sequence.

3. The vehicle trajectory verification method based on the fusion of BeiDou positioning and ETC data according to claim 2, characterized in that, Projecting each location point onto the highway network yields the projected distances for each point, including: Based on the highway network between the starting gantry and the ending gantry, all arc segments are extracted to form a candidate arc segment set. For each positioning point, the straight-line distance between the positioning point and the adjacent positioning point is taken as the search radius. In the candidate arc segment set, arc segments whose vertical distance to the positioning point does not exceed the search radius are selected to form an initial set of arc segments. In the initial screening arc segment set, calculate the direction angle between the direction of each arc segment and the direction of the line connecting two adjacent points of the positioning point, remove arc segments with a direction angle greater than 90 degrees, and take the value with the smallest vertical distance among the remaining arc segments as the projected distance of the positioning point. If the directional angle of all arc segments in the initial screening arc segment set is greater than 90 degrees, then the minimum vertical distance in the initial screening arc segment set is taken as the projected distance of the positioning point.

4. The vehicle trajectory verification method based on the fusion of BeiDou positioning and ETC data according to claim 1, characterized in that, The BeiDou positioning trajectory point sequence is projected onto the highway network, and the cumulative deviation of the average speed between adjacent points relative to the historical speed limit is calculated as the speed deviation factor for each trajectory interval, including: Each trajectory point in the BeiDou positioning trajectory point sequence is projected onto the highway network to obtain the corresponding road network projection points, and arranged in chronological order to form a road network projection point sequence. Extract the road network travel distance and time difference between two adjacent projection points from the road network projection point sequence, calculate the average speed between the two projection points, and record it as the interval average speed. For each interval, the historical speed limit of the road segment where the interval is located is obtained, and the speed deviation rate of a single interval is determined by combining the absolute value of the difference between the average speed of the interval and the historical speed limit. Starting from the first interval of the road network projection point sequence, multiple consecutive intervals are grouped according to the direction of travel, and the number of consecutive intervals in each group is equal to the interval grouping base. The interval grouping base is dynamically determined based on the road network length between the starting gantry and the ending gantry and the total number of points in the road network projection point sequence. Calculate the sum of the single-interval speed deviation rates of all intervals within each group, and use this as the cumulative deviation within the group. The cumulative deviation within the group is used as the velocity deviation factor for the corresponding trajectory interval of the group.

5. The vehicle trajectory auditing method based on the fusion of BeiDou positioning and ETC data according to claim 1, characterized in that, Based on the positional relationship of each trajectory point relative to the starting gantry, the path completion characterization value is calculated, including: The starting gantry is projected onto the highway network to obtain the starting projection point, and the cumulative driving distance corresponding to the starting projection point is set to zero. The cumulative driving distance of the first projection point in the road network projection point sequence is set as the shortest path length from the starting projection point to the first projection point in the highway network. Traverse the sequence of projection points of the road network in chronological order. Starting from the second projection point, calculate the shortest path length between each projection point and the previous projection point in the highway network. Add the shortest path length to the cumulative travel distance corresponding to the previous projection point to obtain the cumulative travel distance of the current projection point. The cumulative driving distance of each projection point is extracted from the road network projection point sequence and used as the initial path completion characterization value of the corresponding trajectory point. The initial path completion value sequence is monotonically increased and corrected, including: If the initial path completion value of the next projection point is less than the initial path completion value of the previous projection point, then the path completion value of the next projection point is replaced with the path completion value of the previous projection point. Otherwise, retain the initial path completion value of the next projection point, and use the corrected result as the path completion value of the corresponding trajectory point.

6. The vehicle trajectory verification method based on the fusion of BeiDou positioning and ETC data according to claim 1, characterized in that, The steps for obtaining the optimized trajectory point sequence include: Based on the calculated path completion characterization value, the path completion characterization value corresponding to each positioning point in the BeiDou positioning trajectory point sequence is extracted, and the path completion increment between adjacent positioning points is calculated. The projection distance of each positioning point in the Beidou positioning trajectory point sequence onto the highway network, as well as the straight-line distance and time difference between adjacent positioning points, are obtained. Location points that simultaneously satisfy both the first and second conditions are identified as outlier trajectory points and are removed. The first condition is that the projected distance of the positioning point is greater than half of the road network driving distance between adjacent projection points; The second condition is that at least one of the following two conditions must be met: The initial value of the path completion characterization value of the positioning point before monotonically increasing correction is less than the initial path completion characterization value of the previous positioning point. The path completion increment corresponding to the location point is greater than twice the product of the historical speed limit and the time difference between adjacent location points; The remaining location points after removing outlier trajectory points are arranged in chronological order to form an optimized trajectory point sequence.

7. The vehicle trajectory auditing method based on the fusion of BeiDou positioning and ETC data according to claim 1, characterized in that, Starting from the initial gantry and ending at the final gantry, each trajectory point in the optimized trajectory point sequence is matched to an arc segment in the road network that satisfies the projection distance constraint and the direction consistency constraint. These arc segments are then connected in the matching order to generate at least one candidate driving path, including: Project each trajectory point in the optimized trajectory point sequence onto the highway network, and select arc segments that satisfy the projection distance constraint and whose arc direction and driving direction satisfy the direction consistency constraint as candidate arc segments of the current trajectory point; The projection distance constraint is that the projection distance of the positioning point is no greater than one-third of the straight-line distance between adjacent trajectory points; The directional consistency constraint is that the angle between the direction of the arc segment and the direction of the line connecting the current trajectory point to the next trajectory point is no greater than 90 degrees, and the angle between the direction of the connecting arc segment inserted in the shortest path and the direction of travel is no greater than 90 degrees. For two adjacent trajectory points, iterate through all combinations of candidate arc segments of the previous trajectory point and candidate arc segments of the next trajectory point. If the two arc segments are directly connected in the highway network, then establish a connection relationship. If the two arc segments are not directly connected, the shortest path length between the two arc segments is calculated. When the shortest path length does not exceed the upper limit determined based on the historical speed limit and the time difference between the adjacent trajectory points, and the connecting path satisfies the directional consistency constraint, a connection relationship is established. The starting gantry and the ending gantry are projected onto the highway network, and the arc segment with the smallest projection distance is taken as the starting arc segment and the ending arc segment. Starting from the initial arc segment, the established adjacent connections are used to connect them step by step in chronological order until the termination arc segment is reached, generating all connected paths from the initial arc segment to the termination arc segment as candidate travel paths.

8. The vehicle trajectory auditing method based on the fusion of BeiDou positioning and ETC data according to claim 1, characterized in that, Based on the time deviation, arc length, and velocity deviation of the corresponding trajectory intervals in each candidate path, the comprehensive consistency cost of each path is calculated, and the candidate path with the lowest cost is selected as the completion path, including: For each candidate path, extract the arc segment sequence that constitutes the candidate path; The arc length of each arc in the arc sequence is obtained, and the time deviation cost is determined by combining the actual travel time between the starting gantry and the ending gantry. The time deviation cost is positively correlated with the degree to which the arc sequence deviates from the expected travel time. Based on the sum of the arc lengths of each arc in the arc sequence, and combined with the shortest path length from the starting gantry to the ending gantry in the highway network, the path length cost is determined, wherein the path length cost is positively correlated with the degree to which the sum of the arc lengths deviates from the shortest path length; Extract the trajectory intervals corresponding to the optimized trajectory point sequence covered by the candidate path, obtain the speed deviation factor of the corresponding trajectory interval, and determine the speed deviation cost, wherein the speed deviation cost is positively correlated with the central tendency of the speed deviation factor; By integrating the time deviation cost, the path length cost, and the speed deviation cost, the comprehensive consistency cost of the candidate path is obtained. The overall consistency cost of all candidate paths is compared, and the candidate path with the smallest overall consistency cost is selected as the completion path.

9. The vehicle trajectory auditing method based on the fusion of BeiDou positioning and ETC data according to claim 1, characterized in that, The cumulative mileage of the completed path is output as the actual mileage traveled by the target vehicle within the missing road segment, used to replace the default path for billing auditing, including: Extract the arc segment sequence contained in the completed path, and sum the lengths of each arc segment to obtain the cumulative mileage of the completed path; Extract the path completion characterization values ​​of the first and last trajectory points in the optimized trajectory point sequence, and calculate the difference between the path completion characterization value of the last trajectory point and the path completion characterization value of the first trajectory point, which is used as the trajectory reference mileage; Compare the cumulative mileage of the completed path with the trajectory reference mileage. If the absolute value of the difference between the two does not exceed the deviation range determined based on the straight-line distance between adjacent trajectory points in the optimized trajectory point sequence, then the cumulative mileage of the completed path is taken as the actual driving mileage. Otherwise, the reference mileage of the trajectory will be taken as the actual mileage traveled; Replace the default shortest path mileage between the starting gantry and the ending gantry with the actual mileage to complete the initial billing audit. The original toll mileage of the target vehicle between the starting gantry and the ending gantry is obtained, and the ratio of the absolute value of the difference between the actual mileage and the toll mileage to the toll mileage is calculated as the mileage difference rate. When the mileage difference rate does not exceed the difference threshold determined based on the road network length between the starting gantry and the ending gantry, the billing mileage of the target vehicle is updated with the actual mileage traveled to complete the secondary audit correction. When the mileage difference rate exceeds the difference threshold, the path completion value sequence and speed deviation factor sequence corresponding to the optimized trajectory point sequence are extracted. The path completion value sequence, the speed deviation factor sequence and the completed path are used to generate an audit report, which is then submitted for manual review.

10. A vehicle trajectory auditing system based on the fusion of BeiDou positioning and ETC data, characterized in that, A vehicle trajectory verification method based on the fusion of BeiDou positioning and ETC data as described in any one of claims 1 to 9 includes: The trajectory data acquisition module is used to extract the BeiDou positioning trajectory point sequence of the target vehicle between the starting gantry and the ending gantry when there is a missing ETC gantry transaction record in the road segment through which the target vehicle passes. The speed deviation analysis module is used to project the Beidou positioning trajectory point sequence onto the highway network and calculate the cumulative deviation of the average speed between adjacent points relative to the historical speed limit, which serves as the speed deviation factor for each trajectory interval. The trajectory optimization module is used to calculate the path completion characterization value based on the positional relationship of each trajectory point relative to the starting gantry, and combine it with the road network projection distance to remove outlier trajectory points and obtain an optimized trajectory point sequence. The candidate path generation module is used to match each trajectory point in the optimized trajectory point sequence to an arc segment in the road network that satisfies the projection distance constraint and the direction consistency constraint, with the starting gantry as the starting point and the ending gantry as the ending point, and connect them in the matching order to generate at least one candidate driving path. The path completion module is used to calculate the comprehensive consistency cost of each path based on the time deviation, arc length, and speed deviation of the corresponding trajectory interval in each candidate path, and select the candidate path with the lowest cost as the completion path. The billing audit module is used to output the cumulative mileage of the completed path as the actual mileage of the target vehicle in the missing road segment, and is used to replace the default path to complete the billing audit.

Images