A trajectory reasoning data management system and method for tunnel blind spot location

Abstract

This invention discloses a trajectory reasoning data management system and method for tunnel blind zone location, relating to the field of trajectory reasoning technology. The invention determines whether a vehicle is approaching a tunnel blind zone; broadcasts a collaborative networking request signal via a communication module to initially form a potential vehicle list; elects a leader vehicle from the potential vehicle list, with the remaining vehicles serving as member vehicles; collects the leader vehicle's driving data upon entering the tunnel as an initial baseline; constructs a convoy trajectory reasoning mechanism: sets a broadcast period, the leader vehicle calculates its own driving data according to the broadcast period and broadcasts it, and member vehicles calculate their own driving data according to the broadcast; sets a re-election mechanism to re-elect a leader vehicle; recovers and collects the actual absolute coordinates from the BeiDou satellite signal, compares the absolute coordinates with the theoretical coordinates at the calculated vehicle tunnel entrance, calculates an error correction vector, and uses the error correction vector to update and optimize the convoy location reasoning mechanism.

Description

Technical Field

[0001] This invention relates to the field of trajectory reasoning technology, specifically to a trajectory reasoning data management system and method for tunnel blind spot locations. Background Technology

[0002] During vehicle operation, dashcams record relevant information, providing crucial data for traffic accident handling and vehicle status monitoring. With the widespread application of the BeiDou Navigation Satellite System, single-BeiDou dashcams are becoming increasingly common. However, when vehicles enter signal blind spots such as tunnels, satellite signals are blocked, preventing single-BeiDou dashcams from obtaining accurate real-time location information. This leads to missing or inaccurate vehicle trajectories within the tunnel. This not only affects the complete recording of the entire driving process but also complicates subsequent data analysis and applications based on the trajectories. For example, in intelligent traffic management, the inability to accurately grasp vehicle speed and dwell time within tunnels hinders precise traffic flow control and accident prevention. Furthermore, methods that calculate vehicle position based on initial speed and tunnel travel time and distance fail to account for changes within the tunnel environment. Environmental factors can affect the vehicle's position, making this inference method inaccurate and prone to errors. Therefore, designing a method to prevent dynamic changes in vehicle trajectory caused by external factors within tunnels is crucial. Summary of the Invention

[0003] The purpose of this invention is to provide a trajectory reasoning data management system and method for tunnel blind spot locations, in order to solve the problems raised in the prior art.

[0004] To achieve the above objectives, the present invention provides the following technical solution:

[0005] A method for trajectory reasoning data management for tunnel blind spot locations, the method comprising the following steps:

[0006] S100: During vehicle operation, the on-board main control chip is used to monitor the vehicle signal strength in real time to determine whether the vehicle is approaching the tunnel blind zone; the communication module broadcasts a collaborative networking request signal and receives networking request signals from other vehicles to initially form a list of potential vehicles.

[0007] Furthermore, the specific steps for initially forming a list of potential vehicles are as follows:

[0008] S101. Collect the time from when the signal strength begins to weaken to when the vehicle enters the tunnel in historical data, and calculate the average and standard deviation of all collected time data. Subtract the standard deviation from the average to obtain the signal loss time range. Set a time series (Time) within the signal loss time range. During vehicle operation, use the onboard main control chip to collect vehicle signal strength according to the time series and calculate the signal strength change trend. The formula is:

[0009] ;

[0010] In the formula, Qc represents the trend of vehicle signal strength change, Q(t) represents the signal strength of the vehicle signal at time t, Q(t-1) represents the signal strength of the vehicle signal at time t-1, and Δt represents the amount of time change.

[0011] In the time series, the vehicle signal strength change trend of two adjacent time points is calculated sequentially. When all calculated vehicle signal strength change trends are negative, it is determined that the vehicle is approaching the tunnel blind zone.

[0012] When the trend of change at all adjacent time points is negative, it is determined to be close to the blind zone. This method avoids the misjudgment problem of judging by a single signal strength threshold. For example, it will not mistakenly identify non-tunnel areas as blind zones due to instantaneous signal fluctuations, thus improving the accuracy of blind zone identification.

[0013] S102. When it is determined that a vehicle is approaching the tunnel blind spot, a cooperative networking request signal is broadcast through the communication module. The cooperative networking request signal includes the vehicle identifier ID, vehicle position coordinates, driving data, and a platooning request. The driving data includes, but is not limited to, vehicle speed, heading angle, and vehicle acceleration. At the same time, the communication module receives cooperative networking request signals broadcast by other vehicles, extracts the real-time tunnel length from the map, and filters the received cooperative networking request signals in each vehicle. It calculates the relative distance between its own vehicle and the receiving vehicle. If the relative distance is greater than the tunnel length, the vehicle is excluded. The vehicle's heading angle is compared with that of the receiving vehicle. If the difference between the heading angles of the two vehicles exceeds 45°, the vehicle is excluded.

[0014] After two rounds of elimination, a list of potential vehicles is constructed using the filtered vehicles.

[0015] From a distance perspective, this ensures that the selected vehicles are likely to enter the tunnel within the same time period, preventing vehicles with no overlap in their travel trajectories from joining. From a heading perspective, it ensures that the vehicles travel in roughly the same direction, reducing data interference in subsequent collaborative networking and laying a high-quality vehicle foundation for subsequent fleet collaborative reasoning.

[0016] S200. Extract the tunnel entry time and sensor data of each vehicle in the potential vehicle list, elect a leader vehicle from the potential vehicle list, and the remaining vehicles are member vehicles; collect the driving data of the leader vehicle when it enters the tunnel as the initial baseline.

[0017] Furthermore, the specific steps for collecting the driving data of the leader's vehicle when entering the tunnel as an initial baseline are as follows:

[0018] S201. Establish a leader election mechanism, including a primary mechanism and a secondary mechanism. The primary mechanism is to extract the tunnel entry time Tin of each vehicle in the potential vehicle list and select the vehicle with the latest tunnel entry time as the leader vehicle.

[0019] The secondary mechanism is as follows: when the time of vehicle entry into tunnel cannot be extracted, the sensor model of each vehicle is extracted, and the vehicle corresponding to the sensor with the highest accuracy is selected as the leader vehicle based on the accuracy marked when the different sensor models were manufactured.

[0020] After the leader's vehicle is elected, the remaining vehicles are defined as member vehicles, and the leader's vehicle and member vehicles are used to form a motorcade.

[0021] A two-tiered election method is adopted, consisting of a primary mechanism (selecting the vehicle that enters the tunnel latest) and a secondary mechanism (selecting the vehicle with the highest sensor accuracy). In the primary mechanism, the vehicle that enters the tunnel latest retains its satellite signal for a longer period before entering the tunnel, resulting in more accurate initial driving data and providing a more reliable initial baseline for the convoy. The secondary mechanism, when entry time cannot be obtained, ensures data quality through sensor accuracy, guaranteeing that the leader vehicle always has the optimal data foundation.

[0022] S202. The leader vehicle broadcasts the information that it has been successfully elected as the leader vehicle to all member vehicles in the convoy, and broadcasts the driving data of the leader convoy when it enters the tunnel to each member vehicle. The broadcast driving data is used as the initial reference as {P_leader(t0), V_leader(t0), θ_leader(t0)}, where P_leader(t0) represents the initial coordinates of the leader vehicle, V_leader(t0) represents the initial speed of the leader vehicle, and θ_leader(t0) represents the initial heading angle of the leader vehicle.

[0023] Using this as a reference starting point, the member vehicles avoid the problem of reference confusion caused by each vehicle setting its own reference, and reduce the initial error of subsequent trajectory reasoning.

[0024] S300, Construct a fleet trajectory reasoning mechanism: The leader vehicle sends election information to all member vehicles. After receiving the information, the member vehicles record their relative driving data with the leader vehicle. Set a broadcast period. The leader vehicle calculates its own driving data according to the broadcast period and broadcasts it. Member vehicles calculate their own driving data according to the broadcast.

[0025] Furthermore, the specific steps for member vehicles to calculate their own driving data based on the broadcast are as follows:

[0026] S301. When the initial reference sent by the leader vehicle when it enters the tunnel is received, each member vehicle calculates the initial relative distance D_j(t0) and the relative heading angle △θ_j(t0) between itself and the leader vehicle in the initial reference. The initial relative distance is calculated by using the Euclidean distance formula to calculate the coordinates of itself and the leader vehicle, and the relative heading angle is measured by the on-board wheel speed sensor.

[0027] S302. Set the inference period Δt. The leader vehicle continuously and dynamically calculates its own coordinates based on the inference period, using the following formula:

[0028] ;

[0029] Where P_leader(t) i P_leader(t) represents the coordinates of the leader vehicle at time t at time i. i-1 V_leader(t) represents the coordinates of the leader vehicle at time point i-1, including the vehicle's lateral and longitudinal coordinates. i θ_leader(t) represents the speed of the leader vehicle at time t at time i. i ) represents the heading angle of the leader vehicle at the i-th time point t; during the initial calculation, the initial coordinates of the leader vehicle are substituted into the output, and the coordinates after the interval time Δt are output; subsequently, the coordinate trajectory of the leader vehicle is continuously calculated according to the inference cycle; after each calculation of the new coordinates, the leader vehicle broadcasts the new coordinates to the other member vehicles. This represents the calculated lateral displacement of the leader vehicle. Indicates the longitudinal displacement of the leader's vehicle;

[0030] S303. When a member vehicle receives the new coordinates broadcast by the leader vehicle according to the inference cycle, it calculates its new relative distance and relative heading angle with the leader vehicle according to the method in S301. Each member vehicle collaboratively infers its own coordinates based on the relative distance and relative heading angle, using the following formula:

[0031] ;

[0032] In the formula, P_j_corrected(t) iD_j(t) represents the coordinates of the j-th member vehicle at time t, where D_j(t) is the coordinate of the j-th member vehicle at time t. i ) represents the relative distance between the j-th member vehicle and the leader vehicle at time t at time i, Δθ_j(t) i ) represents the relative heading angle between the j-th member vehicle and the leader vehicle at the i-th time point t; during the initial calculation, the initial relative distance D_j(t0) and the relative heading angle △θ_j(t0) between the member vehicle and the leader vehicle in the initial reference are substituted, and the coordinates after the interval time △t are output; subsequently, the coordinate trajectory of the member vehicle is continuously calculated according to the inference cycle.

[0033] The collaborative reasoning method based on relative position links the trajectories of all member vehicles with the leader vehicle's trajectory, avoiding trajectory deviations caused by independent reasoning of individual vehicles and ensuring the consistency of the entire convoy's trajectory.

[0034] S400: A re-election mechanism is set up so that when the leader vehicle malfunctions, a new leader vehicle is re-elected.

[0035] Furthermore, the specific steps for re-electing the leader's vehicle are as follows:

[0036] S401. Set up a re-election mechanism. When the leader vehicle broadcasts for more than twice the inference period Δt in the convoy and the number of broadcasts exceeds 3 times, the leader vehicle is judged to be abnormal and cannot serve as the leader. Re-elect the leader convoy according to the content in S201 and re-broadcast the initial baseline.

[0037] The system quickly identified a new leader vehicle and broadcast a new initial baseline, preventing the entire trajectory reasoning system from crashing due to leader vehicle anomalies and ensuring the system's continuous operation within the tunnel.

[0038] S500: When a vehicle exits the tunnel, the BeiDou satellite signal recovers and collects the true absolute coordinates. The absolute coordinates are compared with the theoretical coordinates of the vehicle at the tunnel entrance, and an error correction vector is calculated. The error correction vector is then used to update and optimize the vehicle position inference mechanism.

[0039] Furthermore, the specific steps for updating and optimizing the vehicle position reasoning mechanism using the error correction vector are as follows:

[0040] S501. For each vehicle in the convoy, including the leader vehicle and member vehicles, after the satellite signal recovers upon exiting the tunnel, collect the true absolute coordinates P_GPS of each vehicle and infer the trajectory coordinates P_inferred(t) from the exit. end The error correction vector is calculated by comparing the absolute coordinates with the absolute coordinates, using the following formula:

[0041] ;

[0042] In the formula, △Error represents the error correction vector, which is used to calculate the error correction vector for each inference cycle when the vehicle is traveling in the tunnel. The formula is as follows:

[0043] ;

[0044] In the formula, C(t) i ) represents the error correction vector of the vehicle from entering the tunnel to the i-th time point, t i Let t represent the i-th time point, t0 represent the initial time point when the vehicle enters the tunnel, and t end Indicates the time point when the vehicle exits the tunnel;

[0045] S502. The vehicle position inference mechanism is updated and optimized using error correction vectors. Specifically, the inference coordinates of the vehicle at each time point in the inference cycle are used to obtain the optimized real vehicle coordinate trajectory by adding the error correction vector of the corresponding time interval to the inference coordinates. The updated trajectory data is then encapsulated and uploaded to the cloud for storage.

[0046] The error correction vector is superimposed onto the inference coordinates of the corresponding time interval to obtain the optimized true trajectory, which is then uploaded to the cloud for storage. On the one hand, the optimized trajectory data is more accurate and can serve as a reference for subsequent vehicle travel in the tunnel; on the other hand, the historical optimization data stored in the cloud can provide data support for adjusting the parameters of the inference mechanism (such as inference cycle, relative distance calculation weight, etc.), enabling continuous improvement of the system.

[0047] A trajectory reasoning data management system for tunnel blind spot locations includes a data acquisition module, a potential vehicle list composition module, an initial baseline determination module, a fleet trajectory reasoning module, a re-election module, and a storage and update module.

[0048] The data acquisition module is used to collect vehicle driving data using onboard sensors;

[0049] The potential vehicle list building module includes a communication module, which broadcasts a collaborative networking request signal and receives networking request signals from other vehicles to initially form a potential vehicle list.

[0050] The initial benchmark determination module is used to select a leader vehicle from the potential vehicle list, and the remaining vehicles are member vehicles; the driving data of the leader vehicle when entering the tunnel is collected as the initial benchmark.

[0051] The fleet trajectory reasoning module is used for the leader vehicle to calculate its own driving data according to the broadcast period and broadcast it, and for member vehicles to calculate their own driving data according to the broadcast.

[0052] The re-election module is used to set up a re-election mechanism, which re-elects a leader vehicle when the leader vehicle is abnormal.

[0053] The storage and update module is used to calculate the error correction vector, use the error correction vector to update and optimize the vehicle position inference mechanism, and upload it to the cloud for storage.

[0054] The potential vehicle list module includes a blind spot detection unit and a vehicle filtering unit;

[0055] The blind spot determination unit is used to calculate the vehicle signal strength change trend of two adjacent time points in the time series. When all calculated vehicle signal strength change trends are negative, it is determined that the vehicle is approaching the tunnel blind spot.

[0056] The vehicle filtering unit is used to construct a potential vehicle list after eliminating vehicles twice, by distance and heading angle.

[0057] The fleet trajectory reasoning module includes a leader vehicle reasoning unit and a member vehicle reasoning unit;

[0058] The leader vehicle reasoning unit is used to set the reasoning period Δt, and the leader vehicle continuously and dynamically calculates its own coordinates according to the reasoning period.

[0059] The member vehicle reasoning unit is used by each member vehicle to collaboratively reason its own coordinates based on relative distance and relative heading angle.

[0060] The storage and update module includes an error correction vector calculation unit, an update unit, and a storage unit;

[0061] The error correction vector calculation unit is used to collect the true absolute coordinates P_GPS of each vehicle and infer the trajectory coordinates P_inferred(t) from the exit. end The error correction vector is calculated by comparing the absolute coordinates with the absolute coordinates.

[0062] The update unit is used to update and optimize the vehicle position reasoning mechanism using the error correction vector.

[0063] The storage unit is used to encapsulate the updated trajectory data and upload it to cloud storage.

[0064] Compared with the prior art, the beneficial effects of the present invention are:

[0065] 1. Traditional positioning methods cannot achieve accurate positioning in tunnels due to the loss of satellite signals. This solution constructs a complete tunnel blind zone trajectory management system through vehicle collaborative networking, dynamic trajectory reasoning, and error correction, effectively solving this technical problem and providing key technical support for vehicle monitoring, scheduling, and safety management in tunnels.

[0066] 2. In this invention, member vehicles do not rely on their own easily interfered independent positioning data. Instead, they deduce their own trajectories by calculating the "relative distance + relative heading angle" with the leader vehicle and combining this with the leader vehicle's real-time broadcast dynamic coordinates. Through a "leader-centric collaborative mode," the trajectories of all member vehicles are strongly correlated with the leader vehicle's trajectory, avoiding trajectory drift caused by sensor errors or environmental interference (such as electromagnetic interference in tunnels) in a single vehicle, thus ensuring the consistency and accuracy of the entire convoy's trajectory.

[0067] 3. In this invention, the leader vehicle uniformly calculates and broadcasts its own dynamic coordinates. Member vehicles only need to perform simple deductions based on the broadcast data and their own relative data, without having to repeatedly perform complex independent positioning calculations. Through the "centralized calculation + distributed deduction" model, the amount of redundant data calculations for the entire fleet is reduced, the computing power consumption of the onboard main control chip is lowered, the real-time performance of trajectory reasoning is improved, and it is more adaptable to the computing power resource limitations of onboard equipment. Attached Figure Description

[0068] Figure 1 This is a module distribution diagram of a trajectory reasoning data management system for tunnel blind spot locations according to the present invention;

[0069] Figure 2 This is a schematic diagram illustrating the steps of a trajectory reasoning data management method for tunnel blind spot locations according to the present invention. Detailed Implementation

[0070] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0071] Example: Figures 1-2 As shown, the present invention provides a technical solution.

[0072] A method for trajectory reasoning data management for tunnel blind spot locations, the method comprising the following steps:

[0073] S100: During vehicle operation, the on-board main control chip is used to monitor the vehicle signal strength in real time to determine whether the vehicle is approaching the tunnel blind zone; the communication module broadcasts a collaborative networking request signal and receives networking request signals from other vehicles to initially form a list of potential vehicles.

[0074] The specific steps for initially forming a list of potential vehicles are as follows:

[0075] S101. Collect the time from when the signal strength begins to weaken to when the vehicle enters the tunnel in historical data, and calculate the average and standard deviation of all collected time data. Subtract the standard deviation from the average to obtain the signal loss time range. Set a time series (Time) within the signal loss time range. During vehicle operation, use the onboard main control chip to collect vehicle signal strength according to the time series and calculate the signal strength change trend. The formula is:

[0076] ;

[0077] In the formula, Qc represents the trend of vehicle signal strength change, Q(t) represents the signal strength of the vehicle signal at time t, Q(t-1) represents the signal strength of the vehicle signal at time t-1, and Δt represents the amount of time change.

[0078] In the time series, the vehicle signal strength change trend of two adjacent time points is calculated sequentially. When all calculated vehicle signal strength change trends are negative, it is determined that the vehicle is approaching the tunnel blind zone.

[0079] When the trend of change at all adjacent time points is negative, it is determined to be close to the blind zone. This method avoids the misjudgment problem of judging by a single signal strength threshold. For example, it will not mistakenly identify non-tunnel areas as blind zones due to instantaneous signal fluctuations, thus improving the accuracy of blind zone identification.

[0080] S102. When it is determined that a vehicle is approaching the tunnel blind spot, a cooperative networking request signal is broadcast through the communication module. The cooperative networking request signal includes the vehicle identifier ID, vehicle position coordinates, driving data, and a platooning request. The driving data includes, but is not limited to, vehicle speed, heading angle, and vehicle acceleration. At the same time, the communication module receives cooperative networking request signals broadcast by other vehicles, extracts the real-time tunnel length from the map, and filters the received cooperative networking request signals in each vehicle. It calculates the relative distance between its own vehicle and the receiving vehicle. If the relative distance is greater than the tunnel length, the vehicle is excluded. The vehicle's heading angle is compared with that of the receiving vehicle. If the difference between the heading angles of the two vehicles exceeds 45°, the vehicle is excluded.

[0081] After two rounds of elimination, a list of potential vehicles is constructed using the filtered vehicles.

[0082] From a distance perspective, this ensures that the selected vehicles are likely to enter the tunnel within the same time period, preventing vehicles with no overlap in their travel trajectories from joining. From a heading perspective, it ensures that the vehicles travel in roughly the same direction, reducing data interference in subsequent collaborative networking and laying a high-quality vehicle foundation for subsequent fleet collaborative reasoning.

[0083] S200. Extract the tunnel entry time and sensor data of each vehicle in the potential vehicle list, elect a leader vehicle from the potential vehicle list, and the remaining vehicles are member vehicles; collect the driving data of the leader vehicle when it enters the tunnel as the initial baseline.

[0084] The specific steps for collecting the driving data of the leader's vehicle when entering the tunnel as an initial baseline are as follows:

[0085] S201. Establish a leader election mechanism, including a primary mechanism and a secondary mechanism. The primary mechanism is to extract the tunnel entry time Tin of each vehicle in the potential vehicle list and select the vehicle with the latest tunnel entry time as the leader vehicle.

[0086] The secondary mechanism is as follows: when the time of vehicle entry into tunnel cannot be extracted, the sensor model of each vehicle is extracted, and the vehicle corresponding to the sensor with the highest accuracy is selected as the leader vehicle based on the accuracy marked when the different sensor models were manufactured.

[0087] After the leader's vehicle is elected, the remaining vehicles are defined as member vehicles, and the leader's vehicle and member vehicles are used to form a motorcade.

[0088] A two-tiered election method is adopted, consisting of a primary mechanism (selecting the vehicle that enters the tunnel latest) and a secondary mechanism (selecting the vehicle with the highest sensor accuracy). In the primary mechanism, the vehicle that enters the tunnel latest retains its satellite signal for a longer period before entering the tunnel, resulting in more accurate initial driving data and providing a more reliable initial baseline for the convoy. The secondary mechanism, when entry time cannot be obtained, ensures data quality through sensor accuracy, guaranteeing that the leader vehicle always has the optimal data foundation.

[0089] S202. The leader vehicle broadcasts the information that it has been successfully elected as the leader vehicle to all member vehicles in the convoy, and broadcasts the driving data of the leader convoy when it enters the tunnel to each member vehicle. The broadcast driving data is used as the initial reference as {P_leader(t0), V_leader(t0), θ_leader(t0)}, where P_leader(t0) represents the initial coordinates of the leader vehicle, V_leader(t0) represents the initial speed of the leader vehicle, and θ_leader(t0) represents the initial heading angle of the leader vehicle.

[0090] Using this as a reference starting point, the member vehicles avoid the problem of reference confusion caused by each vehicle setting its own reference, and reduce the initial error of subsequent trajectory reasoning.

[0091] S300, Construct a fleet trajectory reasoning mechanism: The leader vehicle sends election information to all member vehicles. After receiving the information, the member vehicles record their relative driving data with the leader vehicle. Set a broadcast period. The leader vehicle calculates its own driving data according to the broadcast period and broadcasts it. Member vehicles calculate their own driving data according to the broadcast.

[0092] The specific steps for each passenger vehicle to calculate its own driving data based on the broadcast are as follows:

[0093] S301. When the initial reference sent by the leader vehicle when it enters the tunnel is received, each member vehicle calculates the initial relative distance D_j(t0) and the relative heading angle △θ_j(t0) between itself and the leader vehicle in the initial reference. The initial relative distance is calculated by using the Euclidean distance formula to calculate the coordinates of itself and the leader vehicle, and the relative heading angle is measured by the on-board wheel speed sensor.

[0094] S302. Set the inference period Δt. The leader vehicle continuously and dynamically calculates its own coordinates based on the inference period, using the following formula:

[0095] ;

[0096] Where P_leader(t) i P_leader(t) represents the coordinates of the leader vehicle at time t at time i. i-1 V_leader(t) represents the coordinates of the leader vehicle at time point i-1, including the vehicle's lateral and longitudinal coordinates. i θ_leader(t) represents the speed of the leader vehicle at time t at time i. i ) represents the heading angle of the leader vehicle at the i-th time point t; during the initial calculation, the initial coordinates of the leader vehicle are substituted into the output, and the coordinates after the interval time Δt are output; subsequently, the coordinate trajectory of the leader vehicle is continuously calculated according to the inference cycle; after each calculation of the new coordinates, the leader vehicle broadcasts the new coordinates to the other member vehicles. This represents the calculated lateral displacement of the leader vehicle. Indicates the longitudinal displacement of the leader's vehicle;

[0097] S303. When a member vehicle receives the new coordinates broadcast by the leader vehicle according to the inference cycle, it calculates its new relative distance and relative heading angle with the leader vehicle according to the method in S301. Each member vehicle collaboratively infers its own coordinates based on the relative distance and relative heading angle, using the following formula:

[0098] ;

[0099] In the formula, P_j_corrected(t) iD_j(t) represents the coordinates of the j-th member vehicle at time t, where D_j(t) is the coordinate of the j-th member vehicle at time t. i ) represents the relative distance between the j-th member vehicle and the leader vehicle at time t at time i, Δθ_j(t) i ) represents the relative heading angle between the j-th member vehicle and the leader vehicle at the i-th time point t; during the initial calculation, the initial relative distance D_j(t0) and the relative heading angle △θ_j(t0) between the member vehicle and the leader vehicle in the initial reference are substituted, and the coordinates after the interval time △t are output; subsequently, the coordinate trajectory of the member vehicle is continuously calculated according to the inference cycle.

[0100] The collaborative reasoning method based on relative position links the trajectories of all member vehicles with the leader vehicle's trajectory, avoiding trajectory deviations caused by independent reasoning of individual vehicles and ensuring the consistency of the entire convoy's trajectory.

[0101] S400: A re-election mechanism is set up so that when the leader vehicle malfunctions, a new leader vehicle is re-elected.

[0102] The specific steps for re-electing the leader's vehicle are as follows:

[0103] S401. Set up a re-election mechanism. When the leader vehicle broadcasts for more than twice the inference period Δt in the convoy and the number of broadcasts exceeds 3 times, the leader vehicle is judged to be abnormal and cannot serve as the leader. Re-elect the leader convoy according to the content in S201 and re-broadcast the initial baseline.

[0104] The system quickly identified a new leader vehicle and broadcast a new initial baseline, preventing the entire trajectory reasoning system from crashing due to leader vehicle anomalies and ensuring the system's continuous operation within the tunnel.

[0105] S500: When a vehicle exits the tunnel, the BeiDou satellite signal recovers and collects the true absolute coordinates. The absolute coordinates are compared with the theoretical coordinates of the vehicle at the tunnel entrance, and an error correction vector is calculated. The error correction vector is then used to update and optimize the vehicle position inference mechanism.

[0106] The specific steps for updating and optimizing the vehicle position inference mechanism using error correction vectors are as follows:

[0107] S501. For each vehicle in the convoy, including the leader vehicle and member vehicles, after the satellite signal recovers upon exiting the tunnel, collect the true absolute coordinates P_GPS of each vehicle and infer the trajectory coordinates P_inferred(t) from the exit. end The error correction vector is calculated by comparing the absolute coordinates with the absolute coordinates, using the following formula:

[0108] ;

[0109] In the formula, △Error represents the error correction vector, which is used to calculate the error correction vector for each inference cycle when the vehicle is traveling in the tunnel. The formula is as follows:

[0110] ;

[0111] In the formula, C(t) i ) represents the error correction vector of the vehicle from entering the tunnel to the i-th time point, t i Let t represent the i-th time point, t0 represent the initial time point when the vehicle enters the tunnel, and t end Indicates the time point when the vehicle exits the tunnel;

[0112] S502. The vehicle position inference mechanism is updated and optimized using error correction vectors. Specifically, the inference coordinates of the vehicle at each time point in the inference cycle are used to obtain the optimized real vehicle coordinate trajectory by adding the error correction vector of the corresponding time interval to the inference coordinates. The updated trajectory data is then encapsulated and uploaded to the cloud for storage.

[0113] The error correction vector is superimposed onto the inference coordinates of the corresponding time interval to obtain the optimized true trajectory, which is then uploaded to the cloud for storage. On the one hand, the optimized trajectory data is more accurate and can serve as a reference for subsequent vehicle travel in the tunnel; on the other hand, the historical optimization data stored in the cloud can provide data support for adjusting the parameters of the inference mechanism (such as inference cycle, relative distance calculation weight, etc.), enabling continuous improvement of the system.

[0114] A trajectory reasoning data management system for tunnel blind spot locations includes a data acquisition module, a potential vehicle list composition module, an initial baseline determination module, a fleet trajectory reasoning module, a re-election module, and a storage and update module.

[0115] The data acquisition module is used to collect vehicle driving data using onboard sensors;

[0116] The potential vehicle list building module includes a communication module, which broadcasts a collaborative networking request signal and receives networking request signals from other vehicles to initially form a potential vehicle list.

[0117] The initial benchmark determination module is used to select a leader vehicle from the potential vehicle list, and the remaining vehicles are member vehicles; the driving data of the leader vehicle when entering the tunnel is collected as the initial benchmark.

[0118] The fleet trajectory reasoning module is used for the leader vehicle to calculate its own driving data according to the broadcast period and broadcast it, and for member vehicles to calculate their own driving data according to the broadcast.

[0119] The re-election module is used to set up a re-election mechanism, which re-elects a leader vehicle when the leader vehicle is abnormal.

[0120] The storage and update module is used to calculate the error correction vector, use the error correction vector to update and optimize the vehicle position inference mechanism, and upload it to the cloud for storage.

[0121] The potential vehicle list module includes a blind spot detection unit and a vehicle filtering unit;

[0122] The blind spot determination unit is used to calculate the vehicle signal strength change trend of two adjacent time points in the time series. When all calculated vehicle signal strength change trends are negative, it is determined that the vehicle is approaching the tunnel blind spot.

[0123] The vehicle filtering unit is used to construct a potential vehicle list after eliminating vehicles twice, by distance and heading angle.

[0124] The fleet trajectory reasoning module includes a leader vehicle reasoning unit and a member vehicle reasoning unit;

[0125] The leader vehicle reasoning unit is used to set the reasoning period Δt, and the leader vehicle continuously and dynamically calculates its own coordinates according to the reasoning period.

[0126] The member vehicle reasoning unit is used by each member vehicle to collaboratively reason its own coordinates based on relative distance and relative heading angle.

[0127] The storage and update module includes an error correction vector calculation unit, an update unit, and a storage unit;

[0128] The error correction vector calculation unit is used to collect the true absolute coordinates P_GPS of each vehicle and infer the trajectory coordinates P_inferred(t) from the exit. end The error correction vector is calculated by comparing the absolute coordinates with the absolute coordinates.

[0129] The update unit is used to update and optimize the vehicle position reasoning mechanism using the error correction vector.

[0130] The storage unit is used to encapsulate the updated trajectory data and upload it to cloud storage.

[0131] Example: Qinglongshan Tunnel, one-way two lanes, known length L=2000 meters, speed limit 80km / h (≈22.22m / s);

[0132] Vehicles A and B are initially traveling one after the other towards a tunnel at a speed of 72 km / h (20 m / s). The initial distance between the two vehicles is D_B(t0) = 50 meters, and their heading angles are the same.

[0133] Let t0 = 0 seconds, t_end = 102 seconds, and the initial coordinates of vehicle A be (450.000, 1000.000); vehicle B is located 50m behind vehicle A, with coordinates (500.000, 1000.000).

[0134] Both vehicles have an initial speed of 20 m / s and an initial heading angle of 90 degrees.

[0135] Through an election mechanism, vehicle A is the leader's vehicle, vehicle B is the member's vehicle, and the vehicles together form a convoy.

[0136] Assuming the inference cycle is 4s, calculate the coordinates of vehicle A after the first inference cycle. According to the formula, (450.000,1000.000)+[80,0]=(530.000,1000.000);

[0137] At this moment, the relative distance between vehicle B and vehicle A is 30m, and the relative heading angle is 0 degrees;

[0138] The coordinates of vehicle B at this time are deduced to be (530.000, 1000.000) + [30, 0] = (560.000, 1000.000).

[0139] It will be apparent to those skilled in the art that the present invention is not limited to the details of the exemplary embodiments described above, and that the invention can be implemented in other specific forms without departing from its spirit or essential characteristics. Therefore, the embodiments should be considered in all respects as exemplary and non-limiting, and the scope of the invention is defined by the appended claims rather than the foregoing description. Thus, all variations falling within the meaning and scope of equivalents of the claims are intended to be included within the present invention. No reference numerals in the claims should be construed as limiting the scope of the claims.

Claims

1. A method for managing trajectory reasoning data for tunnel blind spot locations, characterized in that: The method includes the following steps: S100: During vehicle operation, the on-board main control chip is used to monitor the vehicle signal strength in real time to determine whether the vehicle is approaching the tunnel blind zone; the communication module broadcasts a collaborative networking request signal and receives networking request signals from other vehicles to initially form a list of potential vehicles. S200. Extract the tunnel entry time and sensor data of each vehicle in the potential vehicle list, elect a leader vehicle from the potential vehicle list, and the remaining vehicles are member vehicles; collect the driving data of the leader vehicle when it enters the tunnel as the initial baseline. S300, Construct a fleet trajectory reasoning mechanism: The leader vehicle sends election information to all member vehicles. After receiving the information, the member vehicles record their relative driving data with the leader vehicle. Set a broadcast period. The leader vehicle calculates its own driving data according to the broadcast period and broadcasts it. Member vehicles calculate their own driving data according to the broadcast. S400: A re-election mechanism is set up so that when the leader vehicle malfunctions, a new leader vehicle is re-elected. S500: When a vehicle exits the tunnel, the BeiDou satellite signal recovers and collects the true absolute coordinates. The absolute coordinates are compared with the theoretical coordinates of the vehicle at the tunnel entrance, and an error correction vector is calculated. The error correction vector is then used to update and optimize the vehicle position inference mechanism.

2. The method for trajectory reasoning data management for tunnel blind spot locations according to claim 1, characterized in that: The specific steps for initially forming the potential vehicle list in S100 are as follows: S101. Collect the time from when the signal strength begins to weaken to when the vehicle enters the tunnel in historical data, and calculate the average and standard deviation of all collected time data. Subtract the standard deviation from the average to obtain the signal loss time range. Set a time series (Time) within the signal loss time range. During vehicle operation, use the onboard main control chip to collect vehicle signal strength according to the time series and calculate the signal strength change trend. The formula is: ； In the formula, Qc represents the trend of vehicle signal strength change, Q(t) represents the signal strength of the vehicle signal at time t, Q(t-1) represents the signal strength of the vehicle signal at time t-1, and Δt represents the amount of time change. In the time series, the vehicle signal strength change trend of two adjacent time points is calculated sequentially. When all calculated vehicle signal strength change trends are negative, it is determined that the vehicle is approaching the tunnel blind zone. S102. When it is determined that a vehicle is approaching the tunnel blind spot, a cooperative networking request signal is broadcast through the communication module. The cooperative networking request signal contains the vehicle identifier ID, vehicle position coordinates, driving data, and a platooning request. The driving data includes vehicle speed, heading angle, and vehicle acceleration. At the same time, the communication module receives cooperative networking request signals broadcast by other vehicles, extracts the real-time tunnel length from the map, and filters the received cooperative networking request signals in each vehicle. It calculates the relative distance between its own vehicle and the receiving vehicle. If the relative distance is greater than the tunnel length, the vehicle is excluded. The vehicle's heading angle is compared with that of the receiving vehicle. If the difference between the heading angles of the two vehicles exceeds 45°, the vehicle is excluded. After two rounds of elimination, a list of potential vehicles is constructed using the filtered vehicles.

3. The method for trajectory reasoning data management for tunnel blind spot locations according to claim 2, characterized in that: The specific steps for collecting the driving data of the leader vehicle when entering the tunnel in S200 as the initial reference are as follows: S201. Establish a leader election mechanism, including a primary mechanism and a secondary mechanism. The primary mechanism is to extract the tunnel entry time Tin of each vehicle in the potential vehicle list and select the vehicle with the latest tunnel entry time as the leader vehicle. The secondary mechanism is as follows: when the time of vehicle entry into tunnel cannot be extracted, the sensor model of each vehicle is extracted, and the vehicle corresponding to the sensor with the highest accuracy is selected as the leader vehicle based on the accuracy marked when the different sensor models were manufactured. After the leader's vehicle is elected, the remaining vehicles are defined as member vehicles, and the leader's vehicle and member vehicles are used to form a motorcade. S202. The leader vehicle broadcasts the information that it has been successfully elected as the leader vehicle to all member vehicles in the convoy, and broadcasts the driving data of the leader convoy when it enters the tunnel to each member vehicle. The broadcast driving data is used as the initial reference as {P_leader(t0), V_leader(t0), θ_leader(t0)}, where P_leader(t0) represents the initial coordinates of the leader vehicle, V_leader(t0) represents the initial speed of the leader vehicle, and θ_leader(t0) represents the initial heading angle of the leader vehicle.

4. The method for trajectory reasoning data management for tunnel blind spot locations according to claim 3, characterized in that: The specific steps for the passenger vehicles in S300 to calculate their own driving data based on the broadcast are as follows: S301. When the initial reference sent by the leader vehicle when it enters the tunnel is received, each member vehicle calculates the initial relative distance D_j(t0) and the relative heading angle △θ_j(t0) between itself and the leader vehicle in the initial reference. The initial relative distance is calculated by using the Euclidean distance formula to calculate the coordinates of itself and the leader vehicle, and the relative heading angle is measured by the on-board wheel speed sensor. S302. Set the inference period Δt. The leader vehicle continuously and dynamically calculates its own coordinates based on the inference period, using the following formula: ； Where P_leader(t) i P_leader(t) represents the coordinates of the leader vehicle at time t at time i. i-1 V_leader(t) represents the coordinates of the leader vehicle at time point i-1, including the vehicle's lateral and longitudinal coordinates. i θ_leader(t) represents the speed of the leader vehicle at time t at time i. i ) represents the heading angle of the leader vehicle at the i-th time point t; during the initial calculation, the initial coordinates of the leader vehicle are substituted into the calculation, and the coordinates after the interval time △t are output; subsequently, the coordinate trajectory of the leader vehicle is continuously calculated according to the inference cycle. Each time a new coordinate is calculated, the leader vehicle broadcasts the new coordinates to the other member vehicles; This represents the calculated lateral displacement of the leader vehicle. Indicates the longitudinal displacement of the leader's vehicle; S303. When a member vehicle receives the new coordinates broadcast by the leader vehicle according to the inference cycle, it calculates its new relative distance and relative heading angle with the leader vehicle according to the method in S301. Each member vehicle collaboratively infers its own coordinates based on the relative distance and relative heading angle, using the following formula: ； In the formula, P_j_corrected(t) i D_j(t) represents the coordinates of the j-th member vehicle at time t, where t is the coordinate of the j-th member vehicle. i ) represents the relative distance between the j-th member vehicle and the leader vehicle at time t, Δθ_j(t) i ) represents the relative heading angle between the j-th member vehicle and the leader vehicle at time t at time i; during the initial calculation, the initial relative distance D_j(t0) and the relative heading angle △θ_j(t0) between the member vehicle and the leader vehicle in the initial reference are substituted, and the coordinates after the interval time △t are output; subsequently, the coordinate trajectory of the member vehicle is continuously calculated according to the inference cycle.

5. The method for trajectory reasoning data management for tunnel blind spot locations according to claim 4, characterized in that: The specific steps for re-electing the leader vehicle in S400 are as follows: S401. Set up a re-election mechanism. When the leader vehicle broadcasts for more than twice the inference period Δt in the convoy and the number of broadcasts exceeds 3 times, the leader vehicle is judged to be abnormal and cannot serve as the leader. Re-elect the leader convoy according to the content in S201 and re-broadcast the initial baseline.

6. The method for trajectory reasoning data management for tunnel blind spot locations according to claim 4, characterized in that: The specific steps of the vehicle position inference mechanism updated and optimized using the error correction vector in S500 are as follows: S501. For each vehicle in the convoy, including the leader vehicle and member vehicles, after the satellite signal recovers upon exiting the tunnel, collect the true absolute coordinates P_GPS of each vehicle and infer the trajectory coordinates P_inferred(t) from the exit. end The error correction vector is calculated by comparing the absolute coordinates with the absolute coordinates, using the following formula: ； In the formula, △Error represents the error correction vector, which is used to calculate the error correction vector for each inference cycle when the vehicle is traveling in the tunnel. The formula is as follows: ； In the formula, C(t) i ) represents the error correction vector of the vehicle from entering the tunnel to the i-th time point, t i Let t represent the i-th time point, t0 represent the initial time point when the vehicle enters the tunnel, and t end Indicates the time point when the vehicle exits the tunnel; S502. The vehicle position inference mechanism is updated and optimized using error correction vectors. Specifically, the inference coordinates of the vehicle at each time point in the inference cycle are used to obtain the optimized real vehicle coordinate trajectory by adding the error correction vector of the corresponding time interval to the inference coordinates. The updated trajectory data is then encapsulated and uploaded to the cloud for storage.

7. A trajectory reasoning data management system for tunnel blind spot locations, characterized in that: The trajectory reasoning data management system includes a data acquisition module, a potential vehicle list composition module, an initial baseline determination module, a fleet trajectory reasoning module, a re-election module, and a storage and update module. The data acquisition module is used to collect vehicle driving data using onboard sensors; The potential vehicle list building module includes a communication module, which broadcasts a collaborative networking request signal and receives networking request signals from other vehicles to initially form a potential vehicle list. The initial benchmark determination module is used to elect a leader vehicle from the potential vehicle list, and the remaining vehicles are member vehicles. The driving data of the leader's vehicle when entering the tunnel was collected as an initial baseline; The fleet trajectory reasoning module is used for the leader vehicle to calculate its own driving data according to the broadcast period and broadcast it, and for member vehicles to calculate their own driving data according to the broadcast. The re-election module is used to set up a re-election mechanism, which re-elects a leader vehicle when the leader vehicle is abnormal. The storage and update module is used to calculate the error correction vector, use the error correction vector to update and optimize the vehicle position inference mechanism, and upload it to the cloud for storage.

8. A trajectory reasoning data management system for tunnel blind spot location according to claim 7, characterized in that: The potential vehicle list module includes a blind spot judgment unit and a vehicle filtering unit; The blind spot determination unit is used to calculate the vehicle signal strength change trend of two adjacent time points in the time series. When all calculated vehicle signal strength change trends are negative, it is determined that the vehicle is approaching the tunnel blind spot. The vehicle filtering unit is used to construct a potential vehicle list after eliminating vehicles twice, by distance and heading angle.

9. A trajectory reasoning data management system for tunnel blind spot location according to claim 7, characterized in that: The fleet trajectory reasoning module includes a leader vehicle reasoning unit and a member vehicle reasoning unit; The leader vehicle reasoning unit is used to set the reasoning period Δt, and the leader vehicle continuously and dynamically calculates its own coordinates according to the reasoning period. The member vehicle reasoning unit is used by each member vehicle to collaboratively reason its own coordinates based on relative distance and relative heading angle.

10. A trajectory reasoning data management system for tunnel blind spot location according to claim 7, characterized in that: The storage and update module includes an error correction vector calculation unit, an update unit, and a storage unit; The error correction vector calculation unit is used to collect the true absolute coordinates P_GPS of each vehicle and infer the trajectory coordinates P_inferred(t) from the exit. end The error correction vector is calculated by comparing the absolute coordinates with the absolute coordinates. The update unit is used to update and optimize the vehicle position reasoning mechanism using the error correction vector. The storage unit is used to encapsulate the updated trajectory data and upload it to cloud storage.

Images