Single-lane vehicle trajectory association method based on multiple geomagnetic sensors

Abstract

This invention discloses a single-lane vehicle trajectory association method based on multiple geomagnetic sensors, primarily addressing the problems of complex current vehicle trajectory association algorithms, high cost of sensors such as cameras and radar, and susceptibility to weather conditions. This invention deploys multiple geomagnetic sensors at equal intervals along one side of a single lane, detecting vehicle arrival timestamps and uploading them to a computing device. The computing device calculates the predicted timestamp of the vehicle's arrival at the geomagnetic sensor, constructs a bipartite graph model of the predicted vehicle timestamp and the geomagnetic sensor measurement timestamp, and uses the KM algorithm to solve the bipartite graph model to be associated. This invention achieves the association between vehicle prediction data and geomagnetic sensor measurement data, thereby enabling vehicle state updates and achieving real-time association of single-lane vehicle trajectories, obtaining accurate single-lane vehicle trajectory information.

Description

Technical Field

[0001] This invention belongs to the field of physical technology, and more specifically relates to a method for associating vehicle trajectories in a single lane based on multiple geomagnetic sensors within the field of data processing technology. This invention can be used in single-lane scenarios to detect the trajectory information of all vehicles at the same time using multiple geomagnetic sensors, thereby obtaining the associated trajectories of vehicles traveling in a single lane at each time point. Background Technology

[0002] Currently, acquiring vehicle trajectories in a single lane is an increasingly important issue in intelligent highway transportation. This helps traffic managers better understand vehicle movement on the road, providing valuable data for traffic flow prediction, traffic congestion warning, and safe driving. Currently, cameras and radar sensors are commonly used to correlate vehicle trajectories in single-lane scenarios. However, these sensors have specific deployment environments, typically requiring mounting on gantry cranes, making large-scale deployment difficult, costly, and their detection accuracy affected by weather. To improve Intelligent Traffic Systems (ITS) and enrich the types of data acquired, geomagnetic sensors are also being widely deployed in transportation systems to collect real-time vehicle trajectory information in a single lane. Geomagnetic sensors are generally not limited by scenario, have low cost, and their detection is unaffected by the environment. However, in terms of vehicle trajectory correlation, traditional data correlation schemes often have high computational complexity, low efficiency, and difficulty in achieving real-time trajectory correlation. This presents a unique challenge to using geomagnetic sensors to achieve real-time vehicle trajectory correlation. This trajectory correlation problem not only affects the acquisition of vehicle driving trajectories by the intelligent transportation system but also the subsequent determination of vehicle behavior status.

[0003] The Road Network Monitoring and Emergency Response Center of the Ministry of Transport proposed a single-lane vehicle trajectory association method based on radar and camera fusion in its jointly filed patent application, "Vehicle Trajectory Tracking Method and System Based on Radar and Video Fusion" (Application No.: 202210867558.4, Publication No.: CN 115327527A). This method achieves data association by synchronizing the timestamps of radar and camera equipment, thus linking the sensor data. However, this method has drawbacks. The deployment environment for sensors such as cameras and radar is unique, typically mounted on gantry frames, making large-scale deployment difficult and costly. Furthermore, the detection accuracy is affected by weather conditions; in harsh environments, vehicles may be difficult to detect, further impacting single-lane vehicle trajectory association.

[0004] In their paper "Vehicle Detection and Tracking in Adverse Weather Using a Deep Learning Framework" (IEEE Transactions on Intelligent Transportation Systems, 2021, pp. 4230-4242), M. Hassaballah et al. proposed a single-lane vehicle trajectory association method. The method involves first combining a traditional Gaussian mixture probability hypothesis density filter tracker with hierarchical data association (HDA), and then employing a multi-scale convolutional neural network (CNN) vehicle detection and tracking algorithm to achieve single-lane vehicle trajectory association. The drawback of this method is that the robust vehicle detection and tracking algorithm using a multi-scale deep CNN is complex and inefficient, making it unsuitable for real-time single-lane vehicle trajectory association scenarios in intelligent highway systems. Summary of the Invention

[0005] The purpose of this invention is to address the shortcomings of the existing technology by proposing a single-lane vehicle trajectory association method based on multiple geomagnetic sensors. This method solves the problems of complex vehicle trajectory association algorithms, high cost of sensors such as cameras and radar, susceptibility to weather conditions, difficulty in large-scale deployment, and difficulty in achieving data association using current geomagnetic sensor-based vehicle trajectory association methods.

[0006] The technical approach to achieving the objective of this invention is as follows: Multiple geomagnetic sensors of the same model are selected based on the required vehicle trajectory range. These sensors are deployed at equal intervals along one side of a single lane, with the deployment direction aligned with the vehicle's travel direction. GPS timing technology is used to provide unified time synchronization for all geomagnetic sensors. Because the geomagnetic sensors used in this invention are low-cost and unaffected by weather, they are easy to deploy on a large scale, solving the problems of high cost and severe weather dependence associated with existing technologies using cameras or radar sensors. This invention uses the same model of geomagnetic sensor and employs unified time synchronization technology, overcoming the data association errors that occur when using different sensors without precise time synchronization in existing technologies. This invention uses a low-computational-complexity bipartite graph matching KM data association algorithm to achieve real-time association of single-lane vehicle trajectories, thereby obtaining accurate single-lane vehicle trajectory information.

[0007] The specific steps to achieve the objective of this invention are as follows:

[0008] Step 1: Select M geomagnetic sensors of the same model according to the required vehicle trajectory range;

[0009] Step 2, Deploy the geomagnetic sensor and synchronize its time:

[0010] Step 2.1: Deploy M geomagnetic sensors at equal intervals of N meters on one side of the lane line of a single lane. The deployment direction of the geomagnetic sensors is the same as the driving direction of the vehicle. The value of N depends on the detection range of the geomagnetic sensors.

[0011] Step 2.2: Use GPS timing technology to provide unified timing for each geomagnetic sensor;

[0012] Step 3: Calculate the timestamp of each vehicle arriving at each geomagnetic sensor:

[0013] Using a vehicle detection algorithm, each geomagnetic sensor detects the timestamp of each vehicle entering the sensor and the timestamp of each vehicle leaving the sensor. The average of the two timestamps is taken as the timestamp of the vehicle arriving at the geomagnetic sensor. The timestamp of each vehicle arriving at each geomagnetic sensor and its corresponding geomagnetic sensor number are uploaded to the computing device.

[0014] Step 4: Construct the bipartite graph model to be associated for each geomagnetic sensor:

[0015] Step 4.1: The computing device categorizes the timestamps of all vehicles arriving at the same time from each geomagnetic sensor into one category;

[0016] Step 4.2, calculate the predicted timestamp for each vehicle's journey from the current geomagnetic sensor to the next geomagnetic sensor using the following formula:

[0017]

[0018] in, Let i be the predicted timestamp of the i-th vehicle arriving at the (k+1)-th geomagnetic sensor from the k-th geomagnetic sensor, where i = 1, 2, ..., I, and I is the total number of vehicles arriving at the k-th geomagnetic sensor, where k = 1, 2, ..., M-1. Let x be the measurement timestamp of the i-th vehicle arriving at the k-th geomagnetic sensor. k+1 Let x be the coordinate position of the (k+1)th geomagnetic sensor. k Let be the coordinates of the k-th geomagnetic sensor. Let be the speed of the i-th vehicle at the moment it passes the k-th geomagnetic sensor;

[0019] Step 4.3: Calculate the weight of the predicted timestamp of each vehicle and the corresponding vehicle measurement timestamp of the geomagnetic sensor based on its arrival time.

[0020]

[0021] Among them, a ij Represents the i-th vehicle Its corresponding (k+1)th geomagnetic sensor's j-th measurement timestamp The weights are j = 1, 2, ..., J, where J is the total number of measurement timestamps of the (k+1)th geomagnetic sensor when the vehicle arrives at the kth geomagnetic sensor.

[0022] Step 4.4: Construct a bipartite graph model to be associated between the predicted data of all vehicles and the measurement data of each geomagnetic sensor at the same time. In this model, X nodes represent vehicles, and Y nodes represent the measurement timestamps of the geomagnetic sensor. Any edge in the bipartite graph to be associated satisfies the following formula:

[0023] W(X i ,Y j ) <= L(X i )+L(Y j )

[0024] Among them, W(X) i ,Y j Let X be the node X of the i-th vehicle in the bipartite graph model to be associated for the (k+1)-th geomagnetic sensor. i Y with the j-th measurement timestamp j The weight of the edge between nodes is equal to a. ij , L(X i Let L(Y) be the node value of node Xi in the bipartite graph model to be associated. j ) represents node Y in the bipartite graph model to be associated. j The node value;

[0025] Step 5: Use the computationally inefficient KM algorithm to associate nodes in each bipartite graph model to be associated.

[0026] Step 5.1: Select one unmatched node from all X nodes in the bipartite graph to be associated;

[0027] Step 5.2: Determine whether an augmenting path for the selected node has been found. If yes, the selected node is successfully matched and proceed to step 5.1; otherwise, proceed to step 5.3.

[0028] Step 5.3: Determine whether the bipartite graph to be associated has reached the maximum matching condition. If so, the association of the bipartite graphs is completed, and the associated bipartite graph is obtained. Then proceed to step 5.5. Otherwise, proceed to step 5.4.

[0029] Step 5.4, modify the node values ​​of all nodes on the augmenting paths: modify the current L(X) values ​​of all nodes on the augmenting paths. i Subtract a constant ε from all L(Y) on the current augmenting path. j After adding a constant ε to each, proceed to step 5.1: ε = min(L(X) i )+L(Yj )-W(i,j));

[0030] Step 5.5: For each X node forming an augmenting path in the associated bipartite graph, update the corresponding vehicle's coordinates using the (k+1)th geomagnetic sensor's coordinates. use Update the corresponding vehicle Let be the speed of the i-th vehicle when it passes the (k+1)-th geomagnetic sensor;

[0031] Step 5.6: For each X node in the associated bipartite graph that has not formed an augmenting path, update the coordinates of the vehicle corresponding to that X node using the coordinates of the (k+1)th geomagnetic sensor. use Update the vehicle corresponding to the X node that has not formed an augmentation road.

[0032] Step 5.7: Add each Y node in the already associated bipartite graph that has not formed an augmenting path to the X node in the next bipartite graph to be associated, and update the coordinates of the geomagnetic sensor corresponding to Y to the coordinates of the vehicle corresponding to X. Update the speed of vehicle X to half the maximum speed allowed in a single lane. Complete the state update of the corresponding vehicle in the associated bipartite graph model;

[0033] Step 5.8: Determine if there are any unconnected bipartite graphs. If so, treat the bipartite graph as the one to be connected and proceed to step 5.1. Otherwise, proceed to step 6.

[0034] Step 6: The computing device transmits the updated vehicle trajectory information for all vehicles on a single lane to the intelligent transportation cloud platform.

[0035] Compared with the prior art, the present invention has the following advantages:

[0036] First, the geomagnetic sensor used in this invention is not limited by the scene, has a low cost, is not affected by the environment, has lower power consumption, and is easy to install. It overcomes the problems of existing technologies that use cameras or radar sensors, which need to be installed on a gantry, are difficult to deploy on a large scale, have high costs, and whose detection accuracy is affected by the weather. This invention enables large-scale deployment and achieves trajectory association of vehicles in a single lane at low cost and without being affected by the environment.

[0037] Secondly, because the present invention uses the same type of geomagnetic sensor and employs GPS timing technology to uniformly synchronize the time of all geomagnetic sensors, it overcomes the problem of potential clock deviations when using different sensors in the prior art; thus enabling the present invention to correctly obtain vehicle detection information on a single lane and improve the accuracy of single-lane vehicle trajectory association.

[0038] Third, because the present invention uses the KM algorithm with low computational complexity to associate nodes in each bipartite graph model to be associated, the algorithm is highly efficient and will not put computational pressure on the computing device; it overcomes the problems of complex and inefficient vehicle trajectory association algorithms in the prior art, enabling the present invention to achieve real-time association of single-lane vehicle trajectories and obtain accurate single-lane vehicle trajectory information. Attached Figure Description

[0039] Figure 1 This is a flowchart of the present invention;

[0040] Figure 2 This is a schematic diagram of the deployment of the geomagnetic sensor according to an embodiment of the present invention;

[0041] Figure 3 This is a schematic diagram of the weighted bipartite graph model of the present invention;

[0042] Figure 4 This is a flowchart illustrating how the KM algorithm is used to associate nodes in each bipartite graph model to be associated, as described in this invention. Detailed Implementation

[0043] The embodiments of the present invention will now be clearly and completely described with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of the present invention, and not all of them. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without creative effort are within the scope of protection of the present invention.

[0044] Reference Figure 1 The implementation steps of the embodiments of the present invention will be further described below.

[0045] Step 1: Select M geomagnetic sensors of the same model according to the required vehicle trajectory range. In the embodiment of the present invention, the geomagnetic sensor selected is model RM3100. The value of M depends on the vehicle trajectory range to be associated. In the embodiment of the present invention, M = 7.

[0046] Step 2: Deploy the geomagnetic sensor and synchronize the time.

[0047] Reference Figure 2 The present invention will further describe an embodiment of deploying a geomagnetic sensor on one side of a single lane.

[0048] Step 2.1: Deploy M geomagnetic sensors at equal intervals of N meters on one side of the lane line of a single lane. The deployment direction of the geomagnetic sensors is the same as the driving direction of the vehicle. In the embodiment of the present invention, M=7 and N=10.

[0049] Step 2.2: Use GPS timing technology to provide unified timing for each geomagnetic sensor.

[0050] Step 3: Calculate the timestamp of each vehicle arriving at each geomagnetic sensor.

[0051] Using a vehicle detection algorithm, each geomagnetic sensor detects the timestamp of each vehicle entering the sensor and the timestamp of each vehicle leaving the sensor. The average of the two timestamps is taken as the timestamp of the vehicle arriving at the geomagnetic sensor. The timestamp of each vehicle arriving at each geomagnetic sensor and its corresponding geomagnetic sensor number are uploaded to the computing device.

[0052] The vehicle detection algorithm refers to calculating and fusing the X, Y, and Z-axis magnetic field strengths detected by the geomagnetic sensor at each moment, resulting in the following fused magnetic field strength:

[0053]

[0054] Where F(k) represents the magnetic field strength after fusing the triaxial magnetic field disturbances detected by the geomagnetic sensor at the sensor deployment location at the k-th time, F X (k), F Y (k), F Z (k) represents the magnetic field strength of the X, Y, and Z axes detected by each geomagnetic sensor at the sensor deployment location at time k.

[0055] The magnetic field strength of each geomagnetic sensor after detecting a vehicle passing by is compared with the vehicle detection threshold Th obtained from actual testing. If there are K consecutive magnetic field strength values ​​greater than Th, it is determined that a vehicle has entered the detection range of the geomagnetic sensor and the timestamp data of the detected vehicle is recorded. If, after determining that a vehicle has entered the detection range of the geomagnetic sensor, there are L consecutive magnetic field strength values ​​less than Th, it is determined that the vehicle has left the detection range of the geomagnetic sensor. The parameters Th, K, and L are determined according to the type of vehicle and the vehicle speed traveling on the actual deployment road.

[0056] The vehicle detection thresholds used in the embodiments of the present invention are: Th = 50nT, K = 5, L = 5.

[0057] Step 4: Construct the bipartite graph model to be associated for the (k+1)th geomagnetic sensor.

[0058] Reference Figure 3 The following is a further description of an embodiment of the bipartite graph model to be associated generated by the (k+1)th geomagnetic sensor of the present invention.

[0059] Step 4.1: The computing device categorizes the measurement timestamps of the same time period reaching the (k+1)th geomagnetic sensor into one category. In the embodiment of the present invention, the time period is selected as 2 seconds to achieve real-time purpose.

[0060] Reference Figure 3 Y1 and Y2 represent the measurement timestamps of the (k+1)th geomagnetic sensor during this time period.

[0061] Step 4.2, calculate the predicted timestamp for each vehicle's journey from the current geomagnetic sensor to the next geomagnetic sensor using the following formula:

[0062]

[0063] in, Let i be the predicted timestamp of the i-th vehicle arriving at the (k+1)-th geomagnetic sensor from the k-th geomagnetic sensor, where i = 1, 2, ..., I, and I is the total number of vehicles arriving at the k-th geomagnetic sensor, where k = 1, 2, ..., M-1. Let x be the measurement timestamp of the i-th vehicle arriving at the k-th geomagnetic sensor. k+1 Let x be the coordinate position of the (k+1)th geomagnetic sensor. k Let be the coordinates of the k-th geomagnetic sensor. Let be the speed of the i-th vehicle at the moment it passes the k-th geomagnetic sensor. (Refer to...) Figure 3 There are two vehicles in this time period. X1 and X2 represent the predicted timestamps of the two vehicles from the k-th geomagnetic sensor to the (k+1)-th geomagnetic sensor.

[0064] Step 4.3, calculate the weight of the predicted timestamp of each vehicle and the corresponding vehicle measurement timestamp of the geomagnetic sensor according to the following formula:

[0065]

[0066] Among them, a ij Represents the i-th vehicle Its corresponding (k+1)th geomagnetic sensor's j-th measurement timestamp The weights are j = 1, 2, ..., J, where J is the total number of measurement timestamps of the (k+1)th geomagnetic sensor when the vehicle arrives at the kth geomagnetic sensor.

[0067] Figure 3 This is a schematic diagram of the bipartite graph model of the first vehicle in the (k+1)th geomagnetic sensor in this embodiment of the invention. Wherein, a 11 Let a represent the weight of the predicted arrival time of the first vehicle and the first measurement time in the (k+1)th geomagnetic sensor. 12 Let a represent the weight of the predicted arrival time of the first vehicle and the measurement time of the second vehicle in the (k+1)th geomagnetic sensor. 21 Let a represent the weight of the predicted arrival time of the second vehicle from the (k+1)th geomagnetic sensor relative to the first measurement time. 22This represents the weight of the predicted arrival time of the second vehicle and the second measurement time in the (k+1)th geomagnetic sensor.

[0068] Step 4.4: Construct a bipartite graph model to be associated between the predicted data of all vehicles and the measurement data of each geomagnetic sensor at the same time. In this model, X nodes represent vehicles, and Y nodes represent the measurement timestamps of the geomagnetic sensor. Any edge in the bipartite graph to be associated satisfies the following formula:

[0069] W(X i ,Y j ) <= L(X i )+L(Y j )

[0070] Among them, W(X) i ,Y j Let X be the node X of the i-th vehicle in the bipartite graph model to be associated for the (k+1)-th geomagnetic sensor. i Y with the j-th measurement timestamp j The weight of the edge between nodes is equal to a. ij , L(X i ) represents node X in the bipartite graph model to be associated. i The node value, L(Y) j ) represents node Y in the bipartite graph model to be associated. j The node value.

[0071] Step 5: Use the KM algorithm to associate each node in the bipartite graph model to be associated.

[0072] Reference Figure 4 The present invention further describes an embodiment of the invention that uses the KM algorithm to associate nodes in each bipartite graph model to be associated.

[0073] Step 5.1: Following the KM algorithm, select one unmatched node from all X nodes in the bipartite graph to be associated.

[0074] Step 5.2: Determine whether an augmenting path for the selected node has been found. If yes, the selected node is successfully matched and step 5.1 is executed; otherwise, step 5.3 is executed.

[0075] Step 5.3: Determine whether the bipartite graph to be associated has reached the maximum matching condition. If so, the association of the bipartite graphs is completed, and the associated bipartite graph is obtained. Then proceed to step 5.5. Otherwise, proceed to step 5.4.

[0076] The maximum matching condition refers to a situation where one of the following conditions is met:

[0077] Condition 1: All X nodes in the bipartite graph to be associated form augmenting paths.

[0078] Condition 2: All Y nodes in the bipartite graph to be associated form augmenting paths.

[0079] Step 5.4, modify the node values ​​of all nodes on the augmenting paths: modify the current L(X) values ​​of all nodes on the augmenting paths. i Subtract a constant ε from all L(Y) on the current augmenting path. j After adding a constant ε to each, proceed to step 5.1: ε = min(L(X) i )+L(Y j )-W(i,j)).

[0080] Step 5.5: For each X node forming an augmenting path in the associated bipartite graph, update the corresponding vehicle's coordinates using the (k+1)th geomagnetic sensor's coordinates. use Update the corresponding vehicle Let be the speed of the i-th vehicle when it passes the (k+1)-th geomagnetic sensor.

[0081] Step 5.6: For each X node in the associated bipartite graph that has not formed an augmenting path, update the coordinates of the vehicle corresponding to that X node using the coordinates of the (k+1)th geomagnetic sensor. use Update the vehicle corresponding to the X node that has not formed an augmentation road.

[0082] Step 5.7: Add each Y node in the already associated bipartite graph that has not formed an augmenting path to the X node in the next bipartite graph to be associated, and update the coordinates of the geomagnetic sensor corresponding to Y to the coordinates of the vehicle corresponding to X. Update the speed of vehicle X to half the maximum speed allowed in a single lane. Complete the state update of the corresponding vehicle in the associated bipartite graph model.

[0083] Step 5.8: Determine if there are any unconnected bipartite graphs. If so, treat the bipartite graph as the one to be connected and proceed to step 5.1. Otherwise, proceed to step 6.

[0084] Step 6: The computing device transmits the updated vehicle trajectory information for all vehicles on a single lane to the intelligent transportation cloud platform.

Claims

1. A method for associating vehicle trajectories in a single lane based on multiple geomagnetic sensors, characterized in that, Multiple geomagnetic sensors are used to associate vehicle trajectories in a single lane, and the KM algorithm for bipartite graph matching is used to obtain the trajectories of vehicles traveling in a single lane. The specific steps of this detection method include the following: Step 1: Select M geomagnetic sensors of the same model according to the required vehicle trajectory range; Step 2, Deploy the geomagnetic sensor and synchronize its time: Step 2.1: Deploy M geomagnetic sensors at equal intervals of N meters on one side of the lane line of a single lane. The deployment direction of the geomagnetic sensors is the same as the driving direction of the vehicle. The value of N depends on the detection range of the geomagnetic sensors. Step 2.2: Use GPS timing technology to provide unified timing for each geomagnetic sensor; Step 3: Calculate the timestamp of each vehicle arriving at each geomagnetic sensor: Using a vehicle detection algorithm, each geomagnetic sensor detects the timestamp of each vehicle entering the sensor and the timestamp of each vehicle leaving the sensor. The average of the two timestamps is taken as the timestamp of the vehicle arriving at the geomagnetic sensor. The timestamp of each vehicle arriving at each geomagnetic sensor and its corresponding geomagnetic sensor number are uploaded to the computing device. Step 4: Construct the bipartite graph model to be associated for each geomagnetic sensor: Step 4.1: The computing device categorizes the timestamps of all vehicles arriving at the same moment from each geomagnetic sensor into one category; Step 4.2, calculate the predicted timestamp for each vehicle's arrival at the next geomagnetic sensor from the current geomagnetic sensor using the following formula: in, Let i be the predicted timestamp of the i-th vehicle arriving at the (k+1)-th geomagnetic sensor from the k-th geomagnetic sensor, where i = 1, 2, ..., I, and I is the total number of vehicles arriving at the k-th geomagnetic sensor, where k = 1, 2, ..., M-1. Let x be the measurement timestamp of the i-th vehicle arriving at the k-th geomagnetic sensor. k+1 Let x be the coordinate position of the (k+1)th geomagnetic sensor. k Let be the coordinates of the k-th geomagnetic sensor. Let be the speed of the i-th vehicle at the moment it passes the k-th geomagnetic sensor; Step 4.3: Calculate the weight of the predicted timestamp of each vehicle and the corresponding vehicle measurement timestamp of the geomagnetic sensor based on its arrival time. Among them, a ij Represents the i-th vehicle Its corresponding (k+1)th geomagnetic sensor's j-th measurement timestamp The weights are j = 1, 2, ..., J, where J is the total number of measurement timestamps of the (k+1)th geomagnetic sensor when the vehicle arrives at the kth geomagnetic sensor. Step 4.4: Construct a bipartite graph model to be associated between the predicted data of all vehicles and the measurement data of each geomagnetic sensor at the same time. In this model, X nodes represent vehicles, and Y nodes represent the measurement timestamps of the geomagnetic sensor. Any edge in the bipartite graph to be associated satisfies the following formula: W(X i ,Y j )<=L(X i )+L(Y j ) Among them, W(X) i ,Y j Let X be the node X of the i-th vehicle in the bipartite graph model to be associated for the (k+1)-th geomagnetic sensor. i Y with the j-th measurement timestamp j The weight of the edge between nodes is equal to a. ij , L(X i ) represents node X in the bipartite graph model to be associated. i The node value, L(Y) j ) represents node Y in the bipartite graph model to be associated. j The node value; Step 5: Use the computationally inefficient KM algorithm to associate nodes in each bipartite graph model to be associated. Step 5.1: Select one unmatched node from all X nodes in the bipartite graph to be associated; Step 5.2: Determine whether an augmenting path for the selected node has been found. If yes, the selected node is successfully matched and proceed to step 5.1; otherwise, proceed to step 5.

3. Step 5.3: Determine whether the bipartite graph to be associated has reached the maximum matching condition. If so, the association of the bipartite graphs is completed, and the associated bipartite graph is obtained. Then proceed to step 5.

5. Otherwise, proceed to step 5.

4. Step 5.4, modify the node values ​​of all nodes on the augmenting paths: modify the current L(X) values ​​of all nodes on the augmenting paths. i Subtract a constant ε from all L(Y) on the current augmenting path. j After adding a constant ε to each, proceed to step 5.1: ε = min(L(X) i )+L(Y j )-W(i,j)); Step 5.5: For each X node forming an augmenting path in the associated bipartite graph, update the corresponding vehicle's coordinates using the (k+1)th geomagnetic sensor's coordinates. use Update the corresponding vehicle Let be the speed of the i-th vehicle when it passes the (k+1)-th geomagnetic sensor; Step 5.6: For each X node in the associated bipartite graph that has not formed an augmenting path, update the coordinates of the vehicle corresponding to that X node using the coordinates of the (k+1)th geomagnetic sensor. use Update the vehicle corresponding to the X node that has not formed an augmentation road. Step 5.7: Add each Y node in the already associated bipartite graph that has not formed an augmenting path to the X node in the next bipartite graph to be associated, and update the coordinates of the geomagnetic sensor corresponding to Y to the coordinates of the vehicle corresponding to X. Update the speed of vehicle X to half the maximum speed allowed in a single lane. Complete the state update of the corresponding vehicle in the associated bipartite graph model; Step 5.8: Determine if there are any unconnected bipartite graphs. If so, treat the bipartite graph as the one to be connected and proceed to step 5.

1. Otherwise, proceed to step 6. Step 6: The computing device transmits the updated vehicle trajectory information for all vehicles on a single lane to the intelligent transportation cloud platform.

2. The single-lane vehicle trajectory association method based on multiple geomagnetic sensors according to claim 1, characterized in that, The vehicle detection algorithm described in step 3 refers to comparing the disturbance magnetic field value detected by each geomagnetic sensor when a vehicle passes by with the vehicle detection threshold Th obtained from actual testing. If there are K consecutive magnetic field strength values ​​greater than Th, it is determined that a vehicle has entered the detection range of the geomagnetic sensor and the timestamp data of the detected vehicle is recorded. If, after a vehicle has entered the detection range of the geomagnetic sensor, there are L consecutive magnetic field strength values ​​less than Th, it is determined that the vehicle has left the detection range of the geomagnetic sensor. The parameters Th, K, and L are determined based on the type of vehicle and the vehicle speed traveling on the actual deployed road.

3. The single-lane vehicle trajectory association method based on multiple geomagnetic sensors according to claim 1, characterized in that, The maximum matching condition mentioned in step 5.3 refers to the situation where one of the following conditions is met: Condition 1: All X nodes in the bipartite graph to be associated form augmenting paths; Condition 2: All Y nodes in the bipartite graph to be associated form augmenting paths.

Images