Method and system for analyzing time and place of vehicle accident based on EDR and T-BOX fusion

Abstract

The invention discloses a method and system for analyzing vehicle accident time and place based on EDR and T-BOX fusion. The method comprises the following steps: collecting EDR data and T-BOX data of a suspicious vehicle; performing multi-dimensional feature matching on the EDR data and the T-BOX data to realize time alignment of the EDR data and the T-BOX data; searching a time point corresponding to the collision event and spatial data corresponding to the time point in the T-BOX data, preliminarily judging whether the suspicious vehicle is a suspected hit-and-run vehicle, and if so, entering a final judgment step; and analyzing the collision point location of the collision event by using the EDR data and the T-BOX data, comparing the collision point location with the collision point location in the vehicle accident record, and if the collision point locations are consistent with the collision point location in the vehicle accident record, determining that the suspicious vehicle is a suspicious accident vehicle and reporting a conclusion. According to the method, the EDR data and the T-BOX data are subjected to feature matching, and linear interpolation and cubic spline interpolation algorithms are combined, so that high-precision space-time automatic positioning of the collision event is realized, and vehicle responsibility determination is assisted.

Description

Technical Field

[0001] This invention relates to the field of automotive electronic data forensics, and in particular to a method and system for analyzing the time and location of vehicle accidents based on the fusion of EDR and T-BOX. Background Technology

[0002] In existing technologies, when a passenger vehicle flees after a collision, it's impossible to directly determine whether it's the vehicle involved based on the collected EDR (Electronic Vehicle Determination) data. This is because EDR only records information such as changes in lateral and longitudinal velocity, mileage at the time of the collision, and vehicle system status, lacking absolute time and GPS positioning. While T-BOX modules continuously record vehicle speed, GPS trajectory, mileage, and heading angle, they do not directly record collision-related data. The current technical challenge in the industry is how to achieve precise spatiotemporal localization and vehicle identification of a collision event using multi-dimensional feature matching, leveraging T-BOX data, and supplementing the missing T-BOX information with EDR-recorded vehicle system status data to assist in determining the cause of the accident. Summary of the Invention

[0003] The purpose of this invention is to solve the problems in the prior art.

[0004] The technical solution adopted by this invention to solve its technical problem is: to provide a method for analyzing the time and location of vehicle accidents based on the fusion of EDR and T-BOX, including the following steps:

[0005] The data acquisition steps include collecting EDR data and T-BOX data of the suspected vehicle. The EDR data includes the first mileage data and the first speed data of the collision event, and the T-BOX includes time-aligned second mileage data, second speed data, and spatial data.

[0006] The data matching step involves multi-dimensional feature matching of EDR data and T-BOX data to achieve time alignment between the two.

[0007] The preliminary determination step involves searching the T-BOX data for the time point corresponding to the collision event and the spatial data corresponding to that time point to preliminarily determine whether the suspicious vehicle is the suspected vehicle involved in the accident. If so, proceed to the final determination step; otherwise, exclude the suspicious vehicle from the suspicion and report the conclusion.

[0008] The final determination step involves analyzing the collision point of the collision event using EDR data and T-BOX data, comparing it with the collision point in the vehicle accident record, and if the comparison is consistent, the suspicious vehicle is considered to be the vehicle suspected of causing the accident and the conclusion is reported.

[0009] Preferably, the data matching step includes the following steps:

[0010] The initial collision time was determined based on the first mileage data from EDR data and the second mileage data from T-BOX data.

[0011] Based on the initial collision time, the first velocity data in the EDR data and the second velocity data in the T-BOX data are matched, and the time point corresponding to the matched velocity is taken as the precise collision time.

[0012] Using the precise collision time, the spatial data in the T-BOX data is interpolated to obtain the collision location.

[0013] Preferably, the preliminary location of the collision time based on the first mileage data from the EDR data and the second mileage data from the T-BOX data includes the following steps:

[0014] The difference between each record of the second mileage data and the first mileage data is calculated and expressed as:

[0015] ;

[0016] in, This represents the difference between the i-th record of the second mileage data and the first mileage data. This represents the i-th record of the second mileage data. This indicates the first mileage data;

[0017] Find the smallest The corresponding time is used as the initial collision time.

[0018] Preferably, if If the collision occurs between two T-BOX records and the difference between the two records is equal, then linear interpolation is used to estimate the absolute time of the collision.

[0019] ;

[0020] in, , These are the timestamps of the two T-BOX records, one before and one after.

[0021] Preferably, the step of matching the first velocity data in the EDR data and the second velocity data in the T-BOX data based on the initially located time, and using the time point corresponding to the matched velocity as the precise collision time, includes the following steps:

[0022] Using the initial positioning time as the center point, T-BOX data from several seconds before and after the initial positioning time are extracted for acceleration calculation, which is expressed as:

[0023] ;

[0024] in, This represents the acceleration of the T-BOX at time t. This refers to the velocity change in the T-BOX data. For time intervals;

[0025] The acceleration can be decomposed into its components as follows:

[0026] ;

[0027] ;

[0028] in, Let T-BOX be the orientation angle at time t;

[0029] Set window size and sliding step size For each sliding window position within the search range The similarity is calculated and expressed as:

[0030] ;

[0031] in, express Location similarity; and These represent the horizontal weighting coefficient and the vertical weighting coefficient, respectively. and These represent the lateral and longitudinal accelerations in the EDR data, respectively; corr is the formula for calculating the Pearson correlation coefficient.

[0032] Find similarity The moment corresponding to the maximum value is taken as the precise collision moment.

[0033] Preferably, the collision position is obtained by interpolating the spatial data in the T-BOX data using the precise collision time, and is represented as follows:

[0034] ;

[0035] ;

[0036] in, The precise moment of collision; and This indicates the time between the two data records before and after the tbox, and < < ; and for The dimensions and longitude of time; and for The latitude and longitude of the moment.

[0037] Preferably, the step of interpolating the spatial data in the T-BOX data to obtain the collision location using the precise collision time includes the following steps:

[0038] A smooth cubic polynomial function is constructed using multiple GPS sampling points. and ;

[0039] The latitude and longitude of the collision location are calculated based on the cubic polynomial function, and are expressed as follows:

[0040] ;

[0041] ;

[0042] in, and These are the latitude and precision of the collision location, respectively. The precise moment of collision is given.

[0043] Preferably, the preliminary determination step specifically involves comparing the precise collision time and collision location with the accident time and location in the vehicle accident record. If the comparison matches, the suspicious vehicle is preliminarily determined to be the suspected vehicle involved in the accident and the final determination step is initiated. Otherwise, the suspicion of the suspicious vehicle is eliminated and the conclusion is reported.

[0044] Preferably, the final determination step specifically involves: confirming the vehicle body location where the collision event occurred based on the lateral and longitudinal acceleration, comparing it with the collision location in the vehicle accident record, and if the comparison is consistent, then the suspicious vehicle is considered to be the vehicle suspected of causing the accident and the conclusion is reported.

[0045] This invention also provides a system for analyzing the time and location of vehicle accidents based on the fusion of EDR and T-BOX, comprising:

[0046] The data acquisition module collects EDR data and T-BOX data of suspicious vehicles. The EDR data includes first mileage data and first speed data of the collision event, and the T-BOX includes time-aligned second mileage data, second speed data, and spatial data.

[0047] The data matching module performs multi-dimensional feature matching on EDR data and T-BOX data to achieve time alignment between the two.

[0048] The preliminary judgment module searches the T-BOX data for the time point corresponding to the collision event and the spatial data corresponding to that time point to preliminarily determine whether the suspicious vehicle is the suspected vehicle involved in the accident. If so, it proceeds to the final judgment step; otherwise, the suspicion of the suspicious vehicle is ruled out and the conclusion is reported.

[0049] The final determination module analyzes the collision point of the collision event using EDR data and T-BOX data, compares it with the collision point in the vehicle accident record, and if the comparison is consistent, the suspicious vehicle is considered to be the vehicle suspected of causing the accident and the conclusion is reported.

[0050] This invention offers the following advantages: It achieves multi-dimensional fusion of EDR collision event features and T-BOX operational data, improving the accuracy and reliability of spatiotemporal localization of collision events through feature matching such as mileage and changes in lateral and longitudinal acceleration. By fully utilizing high-frequency EDR data and spatial information from the T-BOX through a multi-dimensional feature matching algorithm, high-precision spatiotemporal automatic localization of collision events is achieved. Addressing the difference in sampling frequencies between the T-BOX and EDR, linear interpolation and cubic spline interpolation algorithms are employed to improve the accuracy of spatial reconstruction, adapting to different sampling frequencies and vehicle motion states. Combining the lateral and longitudinal acceleration curves of the EDR with relevant case data, the vehicle collision point can be further calculated, accurately locating the vehicle involved. Based on this, the collected vehicle system information can provide auxiliary support for accident cause analysis. This method has good versatility and engineering feasibility, applicable to data evidence collection scenarios after collisions involving various new energy passenger vehicles and commercial vehicles.

[0051] The present invention will be further described in detail below with reference to the accompanying drawings and embodiments, but the present invention is not limited to the embodiments. Attached Figure Description

[0052] Figure 1 This is a diagram illustrating the method steps of an embodiment of the present invention;

[0053] Figure 2 This is a detailed flowchart of an embodiment of the present invention;

[0054] Figure 3 This is a schematic diagram illustrating the final determination using Jeep EDR data in an embodiment of the present invention;

[0055] Figure 4 This is a system structure diagram of an embodiment of the present invention. Detailed Implementation

[0056] See Figure 1 and Figure 2 The diagram shown is a flowchart and detailed steps of a method according to an embodiment of the present invention, including the following steps:

[0057] S101, Data acquisition step: Acquire EDR data and T-BOX data of the suspicious vehicle. The EDR data includes first mileage data and first speed data of the collision event. The T-BOX includes time-aligned second mileage data, second speed data, and spatial data.

[0058] S102, Data matching step, performs multi-dimensional feature matching on EDR data and T-BOX data to achieve time alignment between the two;

[0059] S103, Preliminary determination step: Search the T-BOX data for the time point corresponding to the collision event and the spatial data corresponding to the time point, and preliminarily determine whether the suspicious vehicle is the suspected vehicle involved in the accident. If yes, proceed to the final determination step; otherwise, exclude the suspicious vehicle from the suspicion and report the conclusion.

[0060] S104, Final determination step: Analyze the collision point of the collision event using EDR data and T-BOX data, and compare it with the collision point in the vehicle accident record. If the comparison is consistent, the suspicious vehicle is considered to be the vehicle suspected of causing the accident and the conclusion is reported.

[0061] Specifically, in step S101, professional automotive forensics tools are used to collect EDR module data from the suspected vehicle, including: mileage at the time of the collision, changes in lateral and longitudinal speeds, and vehicle system status data (driver and passenger seatbelt status, seat status, etc.). The vehicle's T-BOX module is then used to collect vehicle operation data, including: time, speed, GPS trajectory, steering angle, and mileage.

[0062] Specifically, the multi-dimensional feature matching and time alignment in S102 includes the following:

[0063] (1) Mileage-based matching algorithm. Input the mileage at the time of collision recorded by the EDR. and T-BOX mileage and corresponding timestamp The goal is to determine the absolute time of the collision event, including the following steps:

[0064] Calculate the difference between each T-BOX record and the EDR crash mileage:

[0065]

[0066] in, This represents the difference between the mileage at the time of the i-th T-BOX record and the EDR collision. Let i be the mileage of the i-th T-BOX record. The mileage at the moment of collision recorded by EDR.

[0067] Traverse all T-BOX records and find smallest According to the corresponding time As a preliminary positioning time .

[0068] like It falls exactly in the middle of the two T-BOX records (i.e.) < < ,and = Then, linear interpolation is used to estimate the absolute time of the collision:

[0069]

[0070] in, , These are the timestamps of the preceding and following T-BOX records, respectively. If no suitable match is found, an error message is output, and the process ends.

[0071] (2) Feature matching algorithm based on lateral and longitudinal acceleration and T-BOX velocity and orientation angle. Input the lateral and longitudinal velocity change sequence of EDR, the velocity and orientation angle sequence of T-BOX, and calculate... Using a central point as the focal point, extract and analyze T-BOX data from several seconds before and after the points; this includes the following steps:

[0072] T-BOX acceleration calculation:

[0073]

[0074] in, This represents the acceleration of the T-BOX at time t. For the change in velocity, As a time interval, this embodiment uses Centered on, shifted forward or backward by a certain number of seconds as Generally, the data recording frequency of the tbox is referenced, such as 1 second or 10 seconds.

[0075] Accelerometer decomposition:

[0076]

[0077]

[0078] in, Let be the direction angle of T-BOX at time t.

[0079] Calculate the similarity of the sliding windows.

[0080] Specifically, the formula for calculating the similarity of sliding windows is:

[0081] ;

[0082] ;

[0083] The time window size is Typically, the duration of the collision event is taken as 200-300 milliseconds; sliding step size A value of 0.05 seconds is used to meet the accuracy requirement of ±0.1 seconds. and These are the lateral and longitudinal accelerations recorded by the EDR module; corr is the Pearson correlation coefficient, used to measure the similarity between the EDR and T-BOX acceleration sequences; , Acceleration sequences corresponding to EDR and TBOX , , , The average values ​​of the corresponding EDR and TBOX acceleration sequences.

[0084] Specifically, in practical applications, when calculating the similarity of sliding windows, the window size is set. =0.3 seconds (EDR lateral and longitudinal velocity change recording time range is 300 milliseconds), sliding step size =0.05 seconds, search range Seconds. For each sliding window position The similarity is expressed as:

[0085] ;

[0086] Among them, the weighting coefficient = =0.5 (indicating that lateral and longitudinal accelerations are equally important in a collision event). Find similarity. The time corresponding to the maximum value This is used for subsequent spatial interpolation and positioning calculations.

[0087] Specifically, the preliminary determination in S103 is based on event spatiotemporal alignment and spatial reconstruction. The precise collision time is input. The goal is to obtain the spatial location of the collision event using GPS data from the T-BOX. Depending on the path conditions, either linear interpolation or cubic spline interpolation is selected to calculate the location.

[0088] Specifically, the linear interpolation method is applicable to scenarios where the T-BOX sampling frequency is 5-10 seconds / strip, the vehicle motion state is simple, and it is suitable for straight driving and slow turning scenarios. The error is: straight driving <5 meters, slow turning 5-15 meters, and sharp turning >20 meters.

[0089]

[0090]

[0091] in, = The precise collision time obtained from the aforementioned steps. < < , , , , The coordinates of the preceding and following GPS points are their latitude and longitude. and This indicates the time between the two data records before and after the tbox. Depending on the tbox data recording frequency, t2-t1 could be 5 seconds or 10 seconds.

[0092] Specifically, the cubic spline interpolation method is applicable in the following scenarios: T-BOX sampling frequency of 1 second / strip, vehicle motion involving complex trajectories, sharp turns, and emergency avoidance, with high accuracy requirements (error controlled within the original GPS error range). The error range is: <3 meters for straight-line driving, 3-8 meters for slow turns, and 8-15 meters for sharp turns.

[0093]

[0094] ;

[0095] in, = The precise collision time obtained from the aforementioned steps. , These are cubic spline interpolation functions constructed from several GPS sampling points.

[0096] The principle of cubic spline interpolation algorithm: It utilizes multiple GPS sampling points to construct a smooth cubic polynomial function. , where i represents the i-th known sampling point among N sampling points; the trajectory is made continuous at the sampling points and the first and second derivatives are continuous. The coefficients are determined by solving the linear equation system, and a complete cubic spline interpolation function is constructed.

[0097] The final output is the spatial location of the collision event. , This algorithm is applicable to collision moments obtained in any way.

[0098] Specifically, the final determination in S104 involves matching the collision time and location with the approximate information in the case record, and simultaneously calculating whether the collision type matches the trace based on the lateral and longitudinal velocity values ​​collected by the EDR module. That is, the maximum longitudinal acceleration value in the collision event record is selected. The time corresponding to this value From the lateral and longitudinal acceleration curves and Find the corresponding lateral acceleration value The collision location is confirmed according to EDR data recording rules. Trace detection is used to confirm whether the suspicious vehicle is the one involved in the incident. Further analysis of vehicle system information collected in the EDR helps in analyzing the cause of the collision. For example, analyzing the collision angle or lateral and longitudinal pulse data collected by the EDR can determine the vehicle's collision point, which can be further verified using traditional trace detection. Finally, the collected EDR and T-BOX data, along with the accident time, location, and cause, are compiled into a standardized report.

[0099] See Figure 3 As shown, using Jeep EDR data as an example, the maximum longitudinal velocity change was found to be -34 km / h, corresponding to 188 milliseconds, as shown in Figure (a); the lateral velocity corresponding to 188 milliseconds was found to be 0 km / h, as shown in Figure (b). Jeep EDR defines longitudinal velocity values ​​as positive in the vehicle's forward direction and lateral velocity values ​​as positive from left to right. Therefore, based on the collected longitudinal and lateral acceleration values, it can be inferred that the vehicle collision point was directly in front of the vehicle. Combined with vehicle tracks, this further verifies whether it is the vehicle involved in the case.

[0100] See Figure 4 The diagram shown is a system structure diagram according to an embodiment of the present invention, including:

[0101] The data acquisition module 401 acquires EDR data and T-BOX data of the suspicious vehicle. The EDR data includes first mileage data and first speed data of the collision event, and the T-BOX includes time-aligned second mileage data, second speed data, and spatial data.

[0102] Data matching module 402 performs multi-dimensional feature matching on EDR data and T-BOX data to achieve time alignment between the two;

[0103] The preliminary judgment module 403 searches the T-BOX data for the time point corresponding to the collision event and the spatial data corresponding to the time point, and makes a preliminary judgment on whether the suspicious vehicle is the suspected vehicle involved in the accident. If so, it proceeds to the final judgment step; otherwise, it excludes the suspicious vehicle from the suspicion and reports the conclusion.

[0104] The final determination module 404 analyzes the collision point of the collision event using EDR data and T-BOX data, compares it with the collision point in the vehicle accident record, and if the comparison is consistent, it considers the suspicious vehicle to be the suspected offender and reports the conclusion.

[0105] As can be seen, this invention achieves highly automatic spatiotemporal localization of collision events by matching features such as mileage and changes in lateral and longitudinal acceleration, combined with linear interpolation and cubic spline interpolation algorithms. By combining registered case information (approximate time and location), it determines whether a collision event occurred, and utilizes vehicle system information and collision type recorded by EDR to assist in analyzing the cause of the accident and determining vehicle liability.

[0106] The above are merely preferred embodiments of the present invention and are 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.

Claims

1. A method for analyzing the time and location of vehicle accidents based on the fusion of EDR and T-BOX, characterized in that, Includes the following steps: The data acquisition steps include collecting EDR data and T-BOX data of the suspected vehicle. The EDR data includes the first mileage data and the first speed data of the collision event, and the T-BOX includes time-aligned second mileage data, second speed data, and spatial data. The data matching step involves multi-dimensional feature matching of EDR data and T-BOX data to achieve time alignment between the two. The preliminary determination step involves searching the T-BOX data for the time point corresponding to the collision event and the spatial data corresponding to that time point to preliminarily determine whether the suspicious vehicle is the suspected vehicle involved in the accident. If so, proceed to the final determination step; otherwise, exclude the suspicious vehicle from the suspicion and report the conclusion. The final determination step involves analyzing the collision point of the collision event using EDR data and T-BOX data, comparing it with the collision point in the vehicle accident record, and if the comparison is consistent, the suspicious vehicle is considered to be the vehicle suspected of causing the accident and the conclusion is reported.

2. The method for analyzing vehicle accident time and location based on the fusion of EDR and T-BOX as described in claim 1, characterized in that, The data matching step includes the following steps: The initial collision time was determined based on the first mileage data from EDR data and the second mileage data from T-BOX data. Based on the initial collision time, the first velocity data in the EDR data and the second velocity data in the T-BOX data are matched, and the time point corresponding to the matched velocity is taken as the precise collision time. Using the precise collision time, the spatial data in the T-BOX data is interpolated to obtain the collision location.

3. The method for analyzing vehicle accident time and location based on the fusion of EDR and T-BOX according to claim 2, characterized in that, The preliminary location of the collision time based on the first mileage data from EDR data and the second mileage data from T-BOX data includes the following steps: The difference between each record of the second mileage data and the first mileage data is calculated and expressed as: ； in, This represents the difference between the i-th record of the second mileage data and the first mileage data. This represents the i-th record of the second mileage data. This indicates the first mileage data; Find the smallest The corresponding time is used as the initial collision time.

4. The method for analyzing vehicle accident time and location based on the fusion of EDR and T-BOX according to claim 3, characterized in that, like If the collision occurs between two T-BOX records and the difference between the two records is equal, then linear interpolation is used to estimate the absolute time of the collision. ； in, , These are the timestamps of the two T-BOX records, one before and one after.

5. The method for analyzing vehicle accident time and location based on the fusion of EDR and T-BOX according to claim 2, characterized in that, The process of matching the first velocity data from the EDR data and the second velocity data from the T-BOX data based on the initially located time, and using the time point corresponding to the matched velocity as the precise collision time, includes the following steps: Using the initial positioning time as the center point, T-BOX data from several seconds before and after the initial positioning time are extracted for acceleration calculation, which is expressed as: ； in, This represents the acceleration of the T-BOX at time t. This refers to the velocity change in the T-BOX data. For time intervals; The acceleration can be decomposed into its components as follows: ； ； in, Let T-BOX be the orientation angle at time t; Set window size and sliding step size For each sliding window position within the search range The similarity is calculated and expressed as: ； in, express Location similarity; and These represent the horizontal weighting coefficient and the vertical weighting coefficient, respectively. and These represent the lateral and longitudinal accelerations in the EDR data, respectively; corr is the formula for calculating the Pearson correlation coefficient. Find similarity The moment corresponding to the maximum value is taken as the precise collision moment.

6. The method for analyzing vehicle accident time and location based on the fusion of EDR and T-BOX according to claim 2, characterized in that, The collision position is obtained by interpolating the spatial data in the T-BOX data using the precise collision time, and is represented as follows: ； ； in, The precise moment of collision; and This indicates the time between the two data records before and after the tbox, and < < ; and for The dimensions and longitude of time; and for The latitude and longitude of the moment.

7. The method for analyzing vehicle accident time and location based on the fusion of EDR and T-BOX according to claim 2, characterized in that, The step of interpolating the spatial data in the T-BOX data to obtain the collision location using the precise collision time includes the following steps: A smooth cubic polynomial function is constructed using multiple GPS sampling points. and ; The latitude and longitude of the collision location are calculated based on the cubic polynomial function, and are expressed as follows: ； ； in, and These are the latitude and precision of the collision location, respectively. The precise moment of collision is given.

8. The method for analyzing vehicle accident time and location based on the fusion of EDR and T-BOX according to claim 1, characterized in that, The preliminary determination step specifically involves comparing the precise collision time and location with the accident time and location in the vehicle accident record. If the comparison matches, the suspicious vehicle is preliminarily determined to be the vehicle suspected of causing the accident and proceeds to the final determination step; otherwise, the suspicion of the suspicious vehicle is ruled out and the conclusion is reported.

9. The method for analyzing vehicle accident time and location based on the fusion of EDR and T-BOX according to claim 1, characterized in that, The final determination step is as follows: based on the lateral and longitudinal acceleration, the vehicle body location where the collision event occurred is confirmed, and compared with the collision location in the vehicle accident record. If the comparison is consistent, the suspicious vehicle is considered to be the vehicle suspected of causing the accident, and the conclusion is reported.

10. A system for analyzing the time and location of vehicle accidents based on the fusion of EDR and T-BOX, characterized in that, include: The data acquisition module collects EDR data and T-BOX data of suspicious vehicles. The EDR data includes first mileage data and first speed data of the collision event, and the T-BOX includes time-aligned second mileage data, second speed data, and spatial data. The data matching module performs multi-dimensional feature matching on EDR data and T-BOX data to achieve time alignment between the two. The preliminary judgment module searches the T-BOX data for the time point corresponding to the collision event and the spatial data corresponding to that time point to preliminarily determine whether the suspicious vehicle is the suspected vehicle involved in the accident. If so, it proceeds to the final judgment step; otherwise, the suspicion of the suspicious vehicle is ruled out and the conclusion is reported. The final determination module analyzes the collision point of the collision event using EDR data and T-BOX data, compares it with the collision point in the vehicle accident record, and if the comparison is consistent, the suspicious vehicle is considered to be the vehicle suspected of causing the accident and the conclusion is reported.

Images