Vehicle trajectory pattern recognition method based on frenet coordinate system features

Abstract

The application relates to a vehicle trajectory mode recognition method based on Frenet coordinate system characteristics, which comprises the following steps: acquiring road left boundary coordinate data, road right boundary coordinate data, road center line coordinate data and road deflection angle through a map; establishing a Frenet coordinate system based on road data information and taking the road right boundary coordinate data information as a reference line; acquiring vehicle positioning information at a set time interval through a roadside sensing device; converting target vehicle positioning information sensed by the device into vehicle operation parameters in the Frenet coordinate system; forming a vehicle operation trajectory by correlating driving trajectory points of the same vehicle in real time through a trajectory correlation algorithm; generating a target vehicle trajectory operation dynamic vector in real time based on the vehicle operation parameters in the converted Frenet coordinate system and the correlated vehicle operation trajectory; and dividing the vehicle trajectory into two modes of normal trajectory and abnormal trajectory. The application can recognize more extensive vehicle abnormal trajectory modes and provide more comprehensive driving behavior analysis support.

Description

Technical Field

[0001] This invention relates to the field of intelligent transportation, and more specifically, to a vehicle trajectory pattern recognition method based on FRENET coordinate system features. Background Technology

[0002] In recent years, with social development, the number of motor vehicles has increased rapidly, and the corresponding traffic accident rate has also risen. Related research indicates that poor driving behavior is one of the main causes of traffic accidents. Since there are many causes of poor driving behavior, how to efficiently and accurately identify it has become a key focus. With the rapid development of sensor technology, various radar and video detection devices can acquire vehicle trajectory points with good spatiotemporal resolution. High-density, long-distance driving trajectory data can be used to mine microscopic driving trajectory patterns. Currently, some research on identifying driving behavior through trajectory mining has yielded certain results. Existing research mainly analyzes the single-dimensional information of the trajectory, without mining the high-dimensional information within the driving trajectory, and cannot analyze the specific impact of trajectory information. Furthermore, because abnormal trajectories of some vehicles are difficult to define, and high data acquisition accuracy is required, existing research on abnormal vehicle trajectory identification mainly targets specific violations, such as driving against traffic. Summary of the Invention

[0003] The technical problem to be solved by this invention is to provide a vehicle trajectory pattern recognition method based on FRENET coordinate system features. By combining existing sensor technology, it can acquire high-density, long-distance vehicle trajectory data, and realize in-depth mining and analysis of micro-driving trajectory patterns, identify a wider range of abnormal vehicle trajectory patterns, and provide more comprehensive driving behavior analysis support.

[0004] The technical solution adopted by this invention to solve its technical problem is: to construct a vehicle trajectory pattern recognition method based on FRENET coordinate system features, comprising the following steps:

[0005] S1. Obtain the coordinate data of the left boundary of the road, the coordinate data of the right boundary of the road, the coordinate data of the center line of the road, and the road deflection angle through the map;

[0006] S2. Based on road data information, establish a Frenet coordinate system using the right boundary coordinate data of the road as a reference line;

[0007] S3. Obtain vehicle location information at set time intervals using roadside sensing devices;

[0008] S4. Convert the target vehicle positioning information sensed by the device into vehicle operating parameters in the Frenet coordinate system;

[0009] S5. By using a trajectory association algorithm, the driving trajectory points of the same vehicle are associated in real time to form the vehicle's running trajectory;

[0010] S6. Based on the vehicle operating parameters in the transformed Frenet coordinate system and the associated vehicle operating trajectory, a dynamic vector of the target vehicle trajectory is generated in real time.

[0011] S7. Based on traffic rules and driving patterns, vehicle trajectories are divided into two modes: normal trajectory and abnormal trajectory.

[0012] According to the above scheme, in step S1, the map is created based on the geodetic coordinate system, and the coordinate data obtained is the absolute coordinates of latitude and longitude.

[0013] According to the above scheme, in step S2, the origin is taken as the starting point of the right boundary of the road and the right boundary of the road is taken as the vertical axis of the Frenet coordinate system.

[0014] According to the above scheme, in step S3, the vehicle positioning information includes the vehicle's absolute latitude and longitude coordinates, vehicle heading angle, and vehicle speed; the roadside sensing equipment includes microwave radar and lidar; the detection time interval set by the sensing equipment is set to 200-400ms.

[0015] According to the above scheme, in step S4, the target vehicle positioning information sensed by the device is converted into vehicle operating parameters in the Frenet coordinate system, specifically as follows:

[0016] S401. Calculate the coordinate system deflection angle based on the coordinates of the calibration point in the sensing device coordinate system and the Cartesian coordinate system, and the formula.

[0017] S402. Calculate the target vehicle coordinates in the Cartesian coordinate system from the coordinates of the sensing device using a formula.

[0018] S403. Convert the target vehicle coordinates in the Cartesian coordinate system to the target vehicle coordinates in the Frenet coordinate system using the conversion formula.

[0019] According to the above scheme, in step S401, the origin of the Cartesian coordinate system is selected based on the actual operating area of ​​the vehicle, with the starting point of operation chosen as the origin. The calibration point is selected based on the detection range of the sensing device. The specific calculation formula and conversion process are as follows:

[0020] First, obtain the coordinates (m, n) and (m0, n0) of the calibration point and the sensing device in the Cartesian coordinate system, and the coordinates (m′, n′) of the calibration point in the sensing device's coordinate system, using the formula:

[0021]

[0022] Calculate the coordinate system deflection angle θ, where θ ranges from 0 to 360°; perform coordinate transformation based on the above formula and the calculated θ to obtain the coordinates (x, y) of the detection point in the Cartesian coordinate system, using the formula:

[0023]

[0024] Calculate the coordinates (x, y) of the detection point in the Cartesian coordinate system, where (x′, y′) are the coordinates of the detection point in the sensor coordinate system; retrieve the reference point (x′, y′) on the reference line that is closest to (x, y). f y f ), where x f That is, the ordinate S of the detection point in the Frenet coordinate system, and the road deflection angle of the reference point is α, which is obtained through the formula.

[0025]

[0026] Calculate the x-coordinate D of the detection point in the Frenet coordinate system to obtain the coordinates (S, D) of the detection point in the transformed Frenet coordinate system.

[0027] According to the above scheme, in step S5, the trajectory points of the same vehicle are linked in real time to form the vehicle's running trajectory through a trajectory association algorithm, specifically as follows:

[0028] S501. Calculate the predicted trajectory coordinates of the vehicle at the next moment based on the converted target vehicle operating parameters and the vehicle heading angle and speed in the target vehicle positioning information.

[0029] S502. Compare the current coordinates of the target vehicle with the predicted coordinates of the previous time. If |ΔS| < 1 and |ΔD| < 0.5, add the current coordinate data to the target vehicle trajectory data. Otherwise, create a new vehicle trajectory database and add the current coordinate data to the new trajectory database. ΔS and ΔD are the differences between the ordinates and abscissas of adjacent times in the Frenet coordinate system, respectively.

[0030] S503, Associate the trajectory points of each vehicle to generate a real-time trajectory, and use S... i The sequence of vehicle trajectory points is numbered according to a set detection time interval to indicate the specific spatial location of the vehicle trajectory.

[0031] According to the above scheme, in step S6, the target vehicle trajectory dynamic vector is generated in real time based on the vehicle running parameters in the transformed Frenet coordinate system and the associated vehicle running trajectory:

[0032] A O =(S i T i X, Y, Vx V y DHW, THW, ID)

[0033] In the formula: S i For the sequence of vehicle trajectory points, the target vehicle is numbered according to the set detection time interval to indicate the specific spatial location of the vehicle trajectory;

[0034] T i The timestamp of the vehicle's trajectory is calculated from the vehicle trajectory sequence pair and the set detection time interval.

[0035] X and Y represent the vehicle's lateral and longitudinal travel distances, respectively, which are derived from the vehicle's operating parameters in the transformed Frenet coordinate system.

[0036] V x V y These are the vehicle's lateral and longitudinal speeds, calculated from the difference in lateral and longitudinal distances between adjacent sequences of the target vehicle and the set detection time interval.

[0037] The ID is the lane number where the vehicle is located, determined by the target vehicle's lateral travel distance and the actual lane width, starting with the leftmost or rightmost lane. The target vehicle's normal lane-changing behavior is determined by its lane ID. DHW and THW represent the target vehicle's position information relative to the vehicle in front, including headway and time distance.

[0038] DHW=Y F -Y O

[0039] THW = DHW / V Oy

[0040] In the formula, Y F Y represents the longitudinal distance traveled by the vehicle in front. O C represents the longitudinal distance traveled by the target vehicle. Oy The longitudinal speed of the target vehicle is THW. If THW < 1s, it indicates that there is a potential collision risk between the target vehicle and the vehicle in front. When the longitudinal distance between the target vehicle and the vehicle in front exceeds the set distance, the position information between the target vehicle and the vehicle in front will not be displayed. Set DHW = ∞ and THW = ∞.

[0041] According to the above scheme, in step S7, based on changes in the speed and direction of the running object, traffic laws, and vehicle motion patterns, the abnormal trajectory patterns include five types: sudden acceleration trajectory, running stop trajectory, collision trajectory, reverse driving trajectory, and road deviation trajectory; among which...

[0042]

[0043] This indicates that the abnormal trajectory is a trajectory of sudden acceleration.

[0044]

[0045] This indicates that the abnormal trajectory is a stopped trajectory;

[0046] |DnW|<4.5

[0047] This indicates that the abnormal trajectory is a collision trajectory;

[0048] X i -X i+1 <0, i∈[1,n]

[0049] This indicates that the abnormal trajectory is a reverse driving trajectory;

[0050] P{Y<0∪Y>d}=1

[0051] This indicates that the abnormal trajectory is an off-road trajectory, where d is the actual width of the road.

[0052] The multi-dimensional dynamic vectors of the target vehicle and the vehicle in front of it are fused to form the trajectory pattern expression of the target vehicle as follows:

[0053] T′ i =(A O A F )

[0054] T′=(T′1, T′2, T′3,..., T′ n ), i∈[1,n]

[0055] In the formula: A O Let A be the trajectory state vector of the target vehicle. F The trajectory state vector of the vehicle in front of the target vehicle;

[0056] The determination criteria are calculated using the trajectory pattern expression to determine whether the target vehicle belongs to a normal trajectory or an abnormal trajectory.

[0057] The vehicle trajectory pattern recognition method based on FRENET coordinate system features of the present invention has the following beneficial effects:

[0058] This invention, by combining existing sensor technologies, can acquire high-density, long-distance vehicle trajectory data and achieve in-depth mining and analysis of microscopic driving trajectory patterns. It converts and calculates the sensor's perception information and combines it with the correlated vehicle running trajectory to construct a high-dimensional trajectory information pattern of the vehicle and identify its trajectory pattern. Compared with existing research that focuses on the identification of specific vehicle violations, it can identify a wider range of abnormal vehicle trajectory patterns, providing more comprehensive driving behavior analysis support. It is applicable to various roads and vehicle targets and has a wider range of applications than previous related technologies. Attached Figure Description

[0059] The present invention will be further described below with reference to the accompanying drawings and embodiments. In the accompanying drawings:

[0060] Figure 1 This is a flowchart of the vehicle trajectory pattern recognition method based on FRENET coordinate system features of the present invention;

[0061] Figure 2 This is a schematic diagram of the Cartesian coordinate system to Frenet coordinate system conversion of the present invention;

[0062] Figure 3 This is a flowchart of the trajectory association algorithm of the present invention;

[0063] Figure 4 This is a schematic diagram illustrating the vehicle trajectory association of the present invention;

[0064] Figure 5 This is a trajectory pattern extraction map of target vehicles and preceding vehicles on a target road segment within a certain time period according to the present invention. Detailed Implementation

[0065] To provide a clearer understanding of the technical features, objectives, and effects of the present invention, specific embodiments of the present invention will now be described in detail with reference to the accompanying drawings.

[0066] like Figure 1-5 As shown, the vehicle trajectory pattern recognition method based on FRENET coordinate system features of the present invention includes the following steps:

[0067] S1. Obtain the coordinates of the left and right boundaries of the road, the centerline of the road, and the road deflection angle using a high-precision map. This map is created based on a geodetic coordinate system, and the obtained coordinate data are absolute latitude and longitude coordinates.

[0068] S2. Based on road data, establish a Frenet coordinate system using the right boundary coordinates as a reference line. The system is established with the right boundary as the origin and the right boundary itself as the vertical axis.

[0069] S3. Obtain vehicle positioning information at set time intervals using roadside sensing devices. The target vehicles are mainly motor vehicles, and the vehicle positioning information includes, but is not limited to, the vehicle's absolute latitude and longitude coordinates, heading angle, and speed; the roadside sensing devices are not limited to microwave radar and lidar; the detection time interval set by the sensing devices is set, but is not limited to, 200–400 ms.

[0070] S4. Convert the target vehicle positioning information sensed by the device into vehicle operating parameters in the Frenet coordinate system. This includes the following steps:

[0071] S401. Calculate the coordinate system deflection angle based on the coordinates of the calibration point in the sensing device coordinate system and the Cartesian coordinate system, and the formula.

[0072] S402. Calculate the target vehicle coordinates in the Cartesian coordinate system from the coordinates of the sensing device using a formula.

[0073] S403. Convert the target vehicle coordinates in the Cartesian coordinate system to the Frenet coordinate system using the transformation formula. The coordinate transformation is as follows: Figure 2 As shown;

[0074] The origin of the Cartesian coordinate system can be selected based on the actual operating area of ​​the vehicle. Generally, the starting point of the operation is chosen as the origin. In this embodiment, the starting point of the right boundary of the road is selected as the origin of the Cartesian coordinate system. The calibration point is selected based on the detection range of the sensing device. The specific calculation formula and conversion process are as follows.

[0075] First, obtain the coordinates (m, n) and (m0, n0) of the calibration point and the sensing device in the Cartesian coordinate system, and the coordinates (m′, n′) of the calibration point in the sensing device's coordinate system; then, using the formula:

[0076]

[0077] Calculate the coordinate system deflection angle θ, where θ ranges from 0 to 360°; perform coordinate transformation based on the above formula and the calculated θ to obtain the coordinates (x, y) of the detection point in the Cartesian coordinate system, and then use the formula...

[0078]

[0079] Calculate the coordinates (x, y) of the detection point in the Cartesian coordinate system, where (x′, y′) are the coordinates of the detection point in the sensor coordinate system; retrieve the reference point (x′, y′) on the reference line that is closest to (x, y). f y f ), where x f That is, the ordinate S of the detection point in the Frenet coordinate system, and the road deflection angle of the reference point is α, which is obtained through the formula.

[0080]

[0081] Calculate the x-coordinate D of the detection point in the Frenet coordinate system to obtain the coordinates (S, D) of the detection point in the transformed Frenet coordinate system; wherein, the transformed vehicle operating parameters include, but are not limited to, the coordinates of the target vehicle in the Frenet coordinate system.

[0082] S5. Through trajectory association algorithms, the driving trajectory points of the same vehicle are associated in real time to form the vehicle's running trajectory; specifically:

[0083] S501. Calculate the predicted trajectory coordinates of the vehicle at the next moment based on the converted target vehicle operating parameters and the vehicle heading angle and speed in the target vehicle positioning information.

[0084] S502. Compare the current coordinates of the target vehicle with the predicted coordinates of the previous time step. If |ΔS| < 1 and |ΔD| < 0.5, add the current coordinate data to the target vehicle trajectory data; otherwise, create a new vehicle trajectory database and add the current coordinate data to the new trajectory database. Here, ΔS and ΔD are the differences between the ordinates and abscissas of adjacent time steps in the Frenet coordinate system, respectively. The flowchart of the trajectory association principle is as follows: Figure 3 As shown;

[0085] S503, Associate the trajectory points of each vehicle to generate a real-time trajectory, and use S... i The sequence of vehicle trajectory points is numbered according to a set detection time interval, representing the specific spatial location of the vehicle trajectory. A specific vehicle trajectory is expressed as follows: Figure 4 As shown.

[0086] S6. Based on the vehicle operating parameters in the transformed Frenet coordinate system and the associated vehicle operating trajectory, the target vehicle trajectory dynamic vector is generated in real time as follows:

[0087] A O =(S i T i X, Y, V x V y DHW, THW, ID)

[0088] In the formula: S i For the sequence of vehicle trajectory points, the target vehicle is numbered according to the set detection time interval to indicate the specific spatial location of the vehicle trajectory;

[0089] T i The timestamp of the vehicle's trajectory is calculated from the vehicle trajectory sequence pair and the set detection time interval.

[0090] X and Y represent the vehicle's lateral and longitudinal travel distances, respectively, which are derived from the vehicle's operating parameters in the transformed Frenet coordinate system.

[0091] V x V y These are the vehicle's lateral and longitudinal speeds, calculated from the difference in lateral and longitudinal distances between adjacent sequences of the target vehicle and the set detection time interval.

[0092] The ID is the lane number where the vehicle is located, determined by the target vehicle's lateral travel distance and the actual lane width. Generally, the starting lane number is one of the left or rightmost lanes; in this embodiment, the rightmost lane is used. The target vehicle's normal lane-changing behavior can be determined by its lane ID. DHW and THW represent the positional information of the target vehicle's interaction with the vehicle in front, including the headway and time distance between vehicles.

[0093] DHW=Y F -Y O

[0094] THW = DHW / V Oy

[0095] In the formula, Y F Y represents the longitudinal distance traveled by the vehicle in front. O V represents the longitudinal distance traveled by the target vehicle. Oy The longitudinal speed of the target vehicle is THW. If THW < 1s, it indicates that there is a potential collision risk between the target vehicle and the vehicle in front. When the longitudinal distance between the target vehicle and the vehicle in front exceeds the set distance, the position information between the target vehicle and the vehicle in front will not be displayed. Set DHW = ∞ and THW = ∞.

[0096] The set distance is related to the vehicle's speed and the accuracy of the roadside sensing equipment. When the vehicle speed is high and the sensing equipment accuracy is high, the set distance is longer. When the vehicle speed is low or the sensing equipment accuracy is low, the set distance is shorter. The shortest set distance is 150m.

[0097] S7. Based on traffic rules and driving patterns, vehicle trajectories are divided into two modes: normal trajectory and abnormal trajectory. Among these, based on changes in the speed and direction of the moving object, as well as traffic laws and vehicle motion patterns, abnormal trajectory modes are further divided into five types: sudden acceleration trajectory, stop trajectory, collision trajectory, reverse driving trajectory, and road deviation trajectory.

[0098]

[0099] This indicates that the abnormal trajectory is a trajectory of sudden acceleration.

[0100]

[0101] This indicates that the abnormal trajectory is a stopped trajectory;

[0102] |DHW|<4.5

[0103] This indicates that the abnormal trajectory is a collision trajectory;

[0104] X i -X i+1 <0, i∈[1,n]

[0105] This indicates that the abnormal trajectory is a reverse driving trajectory;

[0106] P{Y<0∪Y>d}=1

[0107] This indicates that the abnormal trajectory is an off-road trajectory, where d is the actual width of the road.

[0108] The multi-dimensional dynamic vectors of the target vehicle and the vehicle in front of it are fused to form the trajectory pattern expression of the target vehicle as follows:

[0109] T′ i =(A O A F )

[0110] T′=(T′1, T′2, T′3,..., T′ n ), i∈[1,n]

[0111] In the formula: A O Let A be the trajectory state vector of the target vehicle. F The trajectory state vector of the vehicle in front of the target vehicle;

[0112] The determination criteria are calculated using the above trajectory pattern expression to determine whether the target vehicle belongs to a normal trajectory or an abnormal trajectory.

[0113] The embodiments of the present invention have been described above with reference to the accompanying drawings. However, the present invention is not limited to the specific embodiments described above. The specific embodiments described above are merely illustrative and not restrictive. Those skilled in the art can make many other forms under the guidance of the present invention without departing from the spirit and scope of the claims. All of these forms are within the protection scope of the present invention.

Claims

1. A vehicle trajectory pattern recognition method based on FRENET coordinate system features, characterized in that, Includes the following steps: S1. Obtain the coordinate data of the left boundary of the road, the coordinate data of the right boundary of the road, the coordinate data of the center line of the road, and the road deflection angle through the map; S2. Based on road data information, establish a Frenet coordinate system using the right boundary coordinate data of the road as a reference line; S3. Obtain vehicle location information at set time intervals using roadside sensing devices; S4. Convert the target vehicle positioning information sensed by the device into vehicle operating parameters in the Frenet coordinate system; S5. By using a trajectory association algorithm, the driving trajectory points of the same vehicle are associated in real time to form the vehicle's running trajectory; S6. Based on the vehicle operating parameters in the transformed Frenet coordinate system and the associated vehicle operating trajectory, a dynamic vector of the target vehicle trajectory is generated in real time. Based on the vehicle operating parameters in the transformed Frenet coordinate system and the associated vehicle operating trajectory, the target vehicle trajectory dynamic vector is generated in real time as follows: In the formula: For the sequence of vehicle trajectory points, the target vehicle is numbered according to the set detection time interval to indicate the specific spatial location of the vehicle trajectory; The timestamp of the vehicle's trajectory is calculated from the vehicle trajectory sequence pair and the set detection time interval. These represent the vehicle's horizontal and vertical travel distances, respectively, derived from the vehicle's operating parameters in the transformed Frenet coordinate system. These are the vehicle's lateral and longitudinal speeds, calculated from the difference in lateral and longitudinal distances between adjacent sequences of the target vehicle and the set detection time interval. The lane number is determined based on the target vehicle's lateral travel distance and the actual lane width, with the leftmost or rightmost lane as the starting lane number. The target vehicle's normal lane-changing behavior involves passing through the lane the vehicle is in. The number determines, and The location information of the interaction between the target vehicle and the vehicle in front includes the distance between the front ends of the vehicles and the time distance between the front ends; in In the formula, The longitudinal distance traveled by the vehicle in front. The longitudinal distance traveled by the target vehicle. Let the longitudinal speed of the target vehicle be... This indicates a potential collision risk between the target vehicle and the vehicle in front. If the longitudinal distance between the target vehicle and the vehicle in front exceeds a set distance, the positional information between the target vehicle and the vehicle in front will not be displayed. ; S7. Based on traffic rules and driving patterns, vehicle trajectories are divided into two modes: normal trajectory and abnormal trajectory. Based on the changes in the speed and direction of the object being driven, as well as traffic laws and vehicle motion patterns, the abnormal trajectory patterns include five types: sudden speed change trajectory, running stop trajectory, collision trajectory, reverse driving trajectory, and road deviation trajectory. in This indicates that the abnormal trajectory is a trajectory of sudden acceleration. This indicates that the abnormal trajectory is a stopped trajectory; This indicates that the abnormal trajectory is a collision trajectory; This indicates that the abnormal trajectory is a reverse driving trajectory; This indicates that the abnormal trajectory is an off-road trajectory, where d is the actual width of the road. The multi-dimensional dynamic vectors of the target vehicle and the vehicle ahead of it are fused to form the trajectory pattern expression of the target vehicle as follows: In the formula: Let be the trajectory state vector of the target vehicle. The trajectory state vector of the vehicle in front of the target vehicle; The determination criteria are calculated using the trajectory pattern expression to determine whether the target vehicle belongs to a normal trajectory or an abnormal trajectory.

2. The vehicle trajectory pattern recognition method based on FRENET coordinate system features according to claim 1, characterized in that, In step S1, the map is created based on a geodetic coordinate system, and the coordinate data obtained are absolute latitude and longitude coordinates.

3. The vehicle trajectory pattern recognition method based on FRENET coordinate system features according to claim 1, characterized in that, In step S2, the system is established with the starting point of the right boundary of the road as the origin and the right boundary of the road as the vertical axis of the Frenet coordinate system.

4. The vehicle trajectory pattern recognition method based on FRENET coordinate system features according to claim 1, characterized in that, In step S3, the vehicle positioning information includes the vehicle's absolute latitude and longitude coordinates, vehicle heading angle, and vehicle speed; the roadside sensing equipment includes microwave radar and lidar; the detection time interval set by the sensing equipment is 200~400ms.

5. The vehicle trajectory pattern recognition method based on FRENET coordinate system features according to claim 1, characterized in that, In step S4, the target vehicle positioning information sensed by the device is converted into vehicle operating parameters in the Frenet coordinate system, specifically as follows: S401. Calculate the coordinate system deflection angle based on the coordinates of the calibration point in the sensing device coordinate system and the Cartesian coordinate system, and the formula. S402. Calculate the target vehicle coordinates in the Cartesian coordinate system from the coordinates of the sensing device using a formula. S403. Convert the target vehicle coordinates in the Cartesian coordinate system to the target vehicle coordinates in the Frenet coordinate system using the conversion formula.

6. The vehicle trajectory pattern recognition method based on FRENET coordinate system features according to claim 5, characterized in that, In step S401, the origin of the Cartesian coordinate system is selected based on the actual operating area of ​​the vehicle, with the starting point of operation chosen as the origin. The calibration point is selected based on the detection range of the sensing device. The specific calculation formula and conversion process are as follows: First, obtain the coordinates of the calibration point and the sensing device in the Cartesian coordinate system. , and the coordinates of the calibration point in the coordinate system of the sensing device Through the formula: Calculate the coordinate system deflection angle The size of the value, where The range is 0-360°; based on the above formula and calculation Perform coordinate transformation to obtain the coordinates of the detection point in the Cartesian coordinate system. Through the formula: Calculate the coordinates of the detection point in the Cartesian coordinate system ,in To determine the coordinates of the detection point in the sensing coordinate system; to retrieve the distance from the reference line. Recent reference point ,in That is, the ordinate of the detection point in the Frenet coordinate system. The road deflection angle at the reference point is Through formula Calculate the x-coordinate of the detection point in the Frenet coordinate system. The coordinates of the detection points in the transformed Frenet coordinate system are obtained. .

7. The vehicle trajectory pattern recognition method based on FRENET coordinate system features according to claim 1, characterized in that, In step S5, the trajectory points of the same vehicle are linked in real time to form the vehicle's running trajectory using a trajectory association algorithm. Specifically: S501. Calculate the predicted trajectory coordinates of the vehicle at the next moment based on the converted target vehicle operating parameters and the vehicle heading angle and speed in the target vehicle positioning information. S502. Compare the current coordinates of the target vehicle with the predicted coordinates from the previous time step. If the current coordinates are correct, add them to the target vehicle trajectory data; otherwise, create a new vehicle trajectory database and add the current coordinates to the new trajectory database. and These represent the differences between the ordinate and abscissa at adjacent times in the Frenet coordinate system. S503, Associate the trajectory points of each vehicle to generate a real-time trajectory, and use... The sequence of vehicle trajectory points is numbered according to a set detection time interval to indicate the specific spatial location of the vehicle trajectory.

Images