A method and system for intersection vehicle left turn path planning and control

Abstract

The application discloses a kind of intersection vehicle left turn path planning and control method, system, method includes: establishing improved artificial potential field model, wherein, improved artificial potential field model at least includes gravitational potential field model and improved repulsive potential field model, improved repulsive potential field model at least includes the repulsive potential field model of mobile obstacle and boundary repulsive potential field model;Establish vehicle trajectory prediction model, based on vehicle trajectory prediction model, the vehicle parameters of surrounding vehicle are predicted in real time, wherein, vehicle parameters at least include speed, steering angle and position coordinate;Based on vehicle parameters, improved artificial potential field model is used to path planning for left turn vehicle, wherein, when the path planned by improved artificial potential field model falls into local optimum, adjust the potential field constant of gravitational potential field model, when the next step position of left turn vehicle collides with surrounding vehicle, adjust the potential field constant of improved repulsive potential field model, improve the overall traffic efficiency of intersection.

Description

Technical Field

[0001] This invention relates to the field of autonomous vehicle path planning technology, specifically to a method and system for planning and controlling left-turn paths of vehicles at intersections. Background Technology

[0002] Signalized intersections are key and bottleneck nodes in urban traffic networks, and are also prone to road traffic accidents, congestion, and high emissions. Due to factors such as traffic light phase timing and road geometry, left-turning behavior is more unstable and dynamic than straight-going behavior, with more complex influencing mechanisms and constraints, making left-turn trajectory planning and control research extremely challenging. However, by planning and controlling the trajectory of left-turning vehicles, traffic flow at intersections can be greatly improved, making them safer and more efficient.

[0003] However, due to factors such as traffic light phase timing, road geometry (e.g., turning angle, curb radius, number of exit lanes), and conflicts with other traffic participants, the left-turn trajectory of vehicles exhibits strong instability and high dynamic characteristics. Furthermore, left-turning behavior at intersections falls under the category of high-curvature driving conditions. Compared to straight-line trajectory planning, the complexity of parameter variables and constraints in this maneuvering trajectory planning increases exponentially, leading to a significant increase in solution difficulty. Therefore, left-turn trajectory planning and control for autonomous vehicles has become one of the most challenging problems in traffic engineering. However, by planning and controlling the trajectory of left-turning vehicles at intersections, not only can the spatiotemporal resources of the intersection be fully utilized, improving the overall traffic efficiency of the intersection, but it will also effectively enhance vehicle driving safety and energy utilization. Summary of the Invention

[0004] The purpose of this invention is to overcome the above-mentioned technical deficiencies and provide a method and system for planning and controlling the left-turn path of vehicles at intersections, thereby solving the technical problem of unstable vehicle left-turn trajectories in the prior art.

[0005] To achieve the above-mentioned technical objectives, the present invention adopts the following technical solution:

[0006] In a first aspect, the present invention provides a method for planning and controlling left-turn routes for vehicles at intersections, comprising the following steps:

[0007] An improved artificial potential field model is established, wherein the improved artificial potential field model includes at least a gravitational potential field model and an improved repulsive potential field model, and the improved repulsive potential field model includes at least a moving obstacle repulsive potential field model and a boundary repulsive potential field model.

[0008] A vehicle trajectory prediction model is established, and based on the vehicle trajectory prediction model, the vehicle parameters of surrounding vehicles are predicted in real time, wherein the vehicle parameters include at least speed, steering angle and position coordinates.

[0009] Based on the vehicle parameters, the improved artificial potential field model is used to plan the path for the left-turning vehicle. When the path planned by the improved artificial potential field model gets stuck in a local optimum, the potential field coefficient of the gravitational potential field model is adjusted to determine the magnitude and direction of the net force on the left-turning vehicle and to determine the next position of the left-turning vehicle. When the left-turning vehicle at the next position collides with the surrounding vehicles, the potential field coefficient of the repulsive potential field model is adjusted. This process is repeated until the left-turning vehicle reaches its destination.

[0010] In some embodiments, the gravitational potential field model includes:

[0011] Determine the distance from the left-turning vehicle to the target point, and establish a gravitational potential field model based on the distance.

[0012] In some embodiments, the gravitational potential field model is calculated as follows:

[0013]

[0014]

[0015] Among them, U att Let k be the gravitational potential field of the target point with respect to the left-turning vehicle. att r is a constant of the gravitational potential field. tv Let F be the distance between the target vehicle and the obstacle, a be the gravitational potential factor, and F be the distance between the target vehicle and the obstacle. att The gravitational force exerted by the target point on the vehicle is the negative gradient of the gravitational potential energy.

[0016] In some embodiments, the repulsive potential field model of the moving obstacle includes:

[0017] The relative position, angle, and speed between the left-turning vehicle and the obstacle are determined, and a repulsive potential field model of the moving obstacle is established based on the relative position, angle, and speed.

[0018] In some embodiments, the repulsive potential field model of the moving obstacle is calculated as follows:

[0019]

[0020] Among them, U obs k represents the repulsive potential field generated by the obstacle vehicle. obs Let v0 and v be the repulsive potential field constants of the obstacle vehicle. i The speeds of the vehicle turning left and the vehicle blocking the obstacle are θ0 and θ, respectively. i These are the angles between the left-turning vehicle and the obstacle vehicle, respectively, and the line connecting them to their centers of mass; r0 is the furthest distance the obstacle vehicle can travel. ovThis refers to the distance between a vehicle turning left and an obstacle vehicle.

[0021] In some embodiments, the boundary repulsive potential field model includes:

[0022] Determine the boundary line of the left-turn comfort zone, determine the distance of the left-turning vehicle to the boundary line based on the boundary line, and establish a boundary repulsive potential field model based on the distance of the left-turning vehicle to the boundary line.

[0023] In some embodiments, the boundary repulsive potential field model is calculated as follows:

[0024]

[0025]

[0026]

[0027] Among them, U DE The repulsive potential field generated by the boundary DE, U CD The repulsive potential field generated by the boundary CD, U AB Let be the repulsive potential field generated by the boundary line AB, k1, k2, and k3 be the constants of the boundary potential fields of the boundaries DE, CD, and AB, respectively, x, y, and z be the distances of the left-turning vehicle to the boundaries DE, CD, and AB, and x0, y0, and z0 be the distances of the left-turning vehicle from the boundary repulsive potential field set.

[0028] In some embodiments, the step of planning the left-turning vehicle path using the improved artificial potential field model based on the vehicle parameters includes:

[0029] Step 1: Set the coordinates of the target point for vehicles turning left;

[0030] Step 2: Predict vehicle parameters of surrounding vehicles in real time;

[0031] Step 3: Determine the path planning for left-turning vehicles. When the left-turning vehicle has not completed the path planning, determine the magnitude and direction of the attractive force, repulsive force, and resultant force acting on the left-turning vehicle.

[0032] Step 4: When the net force on the left-turning vehicle is 0, adjust the potential field coefficient of the gravitational potential field model and recalculate the magnitude and direction of the net force on the left-turning vehicle in the artificial potential field until the net force is not 0.

[0033] Step 5: Determine the acceleration and steering angle of the left-turning vehicle. When the steering angle of the left-turning vehicle exceeds the kinematic constraints, use the maximum acceleration and maximum steering angle to calculate the next position of the left-turning vehicle and update the vehicle position coordinates.

[0034] Step 6: Based on the vehicle position coordinates of the left-turning vehicle and the vehicle parameters of the surrounding vehicles, determine the collision risk between the left-turning vehicle and the surrounding vehicles. When the left-turning vehicle collides with the surrounding vehicles at the next position, adjust the potential field coefficient of the repulsive potential field model and repeat steps 1 to 6 until the left-turning vehicle reaches the target point.

[0035] In some embodiments, the speed of the left-turning vehicle is calculated as follows:

[0036]

[0037] v n+1 =v n +a n t,

[0038] Among them, a n F is the acceleration vector of the vehicle turning left. SUM Let m be the resultant force of the left-turning vehicle, t be the mass of the left-turning vehicle, and v be the set time step for trajectory updates. n v is the speed of a vehicle turning left. n+1 This indicates the next speed for vehicles turning left.

[0039] Secondly, the present invention also provides a system for planning and controlling left-turn routes for vehicles at intersections, comprising:

[0040] An artificial potential field model establishment module is used to establish an improved artificial potential field model, wherein the improved artificial potential field model includes at least a gravitational potential field model and an improved repulsive potential field model, and the improved repulsive potential field model includes at least a moving obstacle repulsive potential field model and a boundary repulsive potential field model.

[0041] The surrounding vehicle parameter prediction module is used to establish a vehicle trajectory prediction model and, based on the vehicle trajectory prediction model, predict the vehicle parameters of surrounding vehicles in real time. The vehicle parameters include at least speed, steering angle, and position coordinates.

[0042] The left-turning vehicle path planning module is used to plan the left-turning vehicle path based on the vehicle parameters and using the improved artificial potential field model. When the path planned by the improved artificial potential field model gets stuck in a local optimum, the potential field coefficient of the gravitational potential field model is adjusted to determine the magnitude and direction of the net force on the left-turning vehicle and determine the next position of the left-turning vehicle. When the left-turning vehicle at the next position collides with the surrounding vehicles, the potential field coefficient of the repulsive potential field model is adjusted. The process is iterated until the left-turning vehicle reaches the destination point.

[0043] Compared with existing technologies, the method and system for planning and controlling left-turn paths of vehicles at intersections provided by this invention first establishes an improved artificial potential field model, which includes at least a gravitational potential field model and an improved repulsive potential field model, and the improved repulsive potential field model includes at least a moving obstacle repulsive potential field model and a boundary repulsive potential field model; secondly, it establishes a vehicle trajectory prediction model, and based on the vehicle trajectory prediction model, predicts the vehicle parameters of surrounding vehicles in real time, wherein the vehicle parameters include at least speed, steering angle, and position coordinates; finally, based on the vehicle parameters... An improved artificial potential field model is used to plan the path of left-turning vehicles. When the path planned by the improved artificial potential field model gets stuck in a local optimum, the potential field coefficient of the gravitational potential field model is adjusted to determine the magnitude and direction of the net force on the left-turning vehicle and determine the next position of the left-turning vehicle. When the left-turning vehicle at the next position collides with surrounding vehicles, the potential field coefficient of the repulsive potential field model is adjusted. This process is iterated until the left-turning vehicle reaches its destination. This increases the safety of vehicle driving and can significantly reduce overall vehicle delays and average number of stops, improve the overall traffic service level of the city, and alleviate urban traffic congestion. Attached Figure Description

[0044] Figure 1 This is a flowchart of an embodiment of the method for planning and controlling left-turn routes of vehicles at intersections provided by the present invention;

[0045] Figure 2 This is a schematic diagram of an embodiment of the system for planning and controlling left-turn routes for vehicles at intersections provided by the present invention;

[0046] Figure 3 This is a schematic diagram of the conflict point between left-turning and oncoming straight-going vehicles in the method for planning and controlling left-turning paths at intersections provided by the present invention.

[0047] Figure 4a This is a schematic diagram of an obstacle avoidance scenario when a left-turning vehicle approaches the conflict point before a straight-going vehicle, as provided by the present invention, regarding the method for planning and controlling left-turning paths at intersections.

[0048] Figure 4b This is a schematic diagram of an obstacle avoidance scenario when a left-turning vehicle approaches the conflict point before a straight-going vehicle, as provided by the present invention, regarding the method for planning and controlling left-turning paths at intersections.

[0049] Figure 5a This is a schematic diagram of an obstacle avoidance scenario when a straight-going vehicle approaches the conflict point before a left-turning vehicle, as provided in the present invention, regarding the method for planning and controlling left-turning paths at intersections.

[0050] Figure 5b This is a schematic diagram of an obstacle avoidance scenario when a straight-going vehicle approaches the conflict point before a left-turning vehicle, as provided in the present invention, regarding the method for planning and controlling left-turning paths at intersections.

[0051] Figure 6 This is a schematic diagram of the left-turn comfort zone boundary of the method for planning and controlling left-turn paths of vehicles at intersections provided by the present invention;

[0052] Figure 7 This is a schematic diagram of the path planning method for the left-turn path planning and control of vehicles at intersections provided by the present invention. Detailed Implementation

[0053] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the invention.

[0054] This invention provides a method for planning and controlling left-turn routes for vehicles at intersections. Figure 1 This is a flowchart of the method for planning and controlling left-turn routes for vehicles at intersections provided in an embodiment of the present invention. Please refer to [link / reference]. Figure 1 The method for planning and controlling left-turn routes for vehicles at intersections includes the following steps:

[0055] S100. Establish an improved artificial potential field model, wherein the improved artificial potential field model includes at least a gravitational potential field model and an improved repulsive potential field model, and the improved repulsive potential field model includes at least a moving obstacle repulsive potential field model and a boundary repulsive potential field model.

[0056] S200. Establish a vehicle trajectory prediction model. Based on the vehicle trajectory prediction model, predict the vehicle parameters of surrounding vehicles in real time. The vehicle parameters include at least speed, steering angle and position coordinates.

[0057] S300. Based on vehicle parameters, an improved artificial potential field model is used to plan the path of the left-turning vehicle. When the path planned by the improved artificial potential field model gets stuck in a local optimum, the potential field coefficient of the gravitational potential field model is adjusted to determine the magnitude and direction of the net force on the left-turning vehicle and determine the next position of the left-turning vehicle. When the left-turning vehicle at the next position collides with the surrounding vehicles, the potential field coefficient of the repulsive potential field model is adjusted. The process is iterated until the left-turning vehicle reaches the destination point.

[0058] In this embodiment, firstly, an improved artificial potential field model is established, which includes at least a gravitational potential field model and an improved repulsive potential field model. The improved repulsive potential field model includes at least a moving obstacle repulsive potential field model and a boundary repulsive potential field model. Secondly, a vehicle trajectory prediction model is established. Based on the vehicle trajectory prediction model, the vehicle parameters of surrounding vehicles are predicted in real time, including at least speed, steering angle, and position coordinates. Finally, based on the vehicle parameters, the improved artificial potential field model is used to plan the path of left-turning vehicles. When the path planned by the improved artificial potential field model gets stuck in a local optimum, the potential field coefficient of the gravitational potential field model is adjusted to determine the magnitude and direction of the resultant force on the left-turning vehicle and determine the next position of the left-turning vehicle. When the left-turning vehicle at the next position collides with surrounding vehicles, the potential field coefficient of the repulsive potential field model is adjusted. This process is iterated until the left-turning vehicle reaches its destination. This increases the safety of vehicle travel and improves the overall traffic efficiency of the intersection.

[0059] In some embodiments, in step S100, an artificial potential field model is constructed by introducing other constraints such as boundary constraints and vehicle dynamics constraints inside the intersection. Based on the left-turning situation of vehicles at the signalized intersection, the artificial potential field function is improved to establish an improved artificial potential field model. The improved artificial potential field model includes at least a gravitational potential field model and an improved repulsive potential field model. The improved repulsive potential field model includes at least a moving obstacle repulsive potential field model and a boundary repulsive potential field model. The constructed improved artificial potential field model is used to control the autonomous vehicle with collision risk. A repulsive potential field is established for the straight-going obstacle vehicle, and the autonomous vehicle turning left dynamically performs left-turning motion in the synthetic potential field.

[0060] In some embodiments, a gravitational potential field model is established, and during vehicle operation, the target point generates a gravitational potential field U on the vehicle. att In order to reach the target point, the vehicle will be subject to the gravitational pull F generated by the target point. att The gravitational potential field strength at the target point is related to the distance between the vehicle turning left and the target point. Generally, the greater the distance, the stronger the gravitational potential field and the greater the gravitational force; the closer the distance, the weaker the gravitational potential field and the smaller the gravitational force. The gravitational potential field at the target point is 0. The calculation formula for its gravitational potential field model is:

[0061]

[0062]

[0063] Among them, U att Let k be the gravitational potential field of the target point with respect to the left-turning vehicle. att r is a constant of the gravitational potential field. tvLet F be the distance between the target vehicle and the obstacle, a be the gravitational potential factor, and F be the distance between the target vehicle and the obstacle. att The gravitational force exerted by the target point on the vehicle is the negative gradient of the gravitational potential energy.

[0064] In some embodiments, traditional artificial potential fields typically only consider the repulsive potential field of stationary obstacles, neglecting the influence of moving obstacles and road boundaries. However, in the scenario of vehicles turning left at an intersection, in addition to stationary obstacles, moving obstacles such as oncoming straight-ahead vehicles that conflict with left-turning vehicles pose a greater threat to intersection safety and a greater risk to left-turning vehicles due to their own speed, acceleration, and other characteristics. Traditional artificial potential field models cannot well describe the potential field state of moving obstacles. Therefore, it is necessary to improve the traditional artificial potential field model to consider both dynamic and static obstacles, enabling it to better describe the states of both types of obstacles and to include road boundaries. For a schematic diagram of the conflict point between left-turning and oncoming straight-ahead vehicles, please refer to [link / reference needed]. Figure 3 During the process of vehicles arriving at and passing through an intersection, generally speaking, depending on the order in which left-turning vehicles and oncoming straight-going vehicles arrive at the point of conflict, there are two obstacle avoidance processes at the point of convergence. Please refer to [link / reference needed]. Figure 4a , Figure 4b , Figure 5a , Figure 5b ,in, Figure 4a , Figure 4b This scenario involves a left-turning vehicle approaching the point of conflict before a straight-going vehicle. In this case, to pass through the intersection more quickly, the left-turning vehicle will turn ahead of the straight-going vehicle and pass through the intersection in front of it. Figure 5a , Figure 5b This scenario involves an obstacle avoidance situation where a vehicle going straight approaches the point of conflict before a vehicle turning left. The left-turning vehicle will delay its turn and pass behind the straight-going vehicle. During obstacle avoidance, the relative distance and angle between the vehicle and the obstacle directly reflect the vehicle's safety status. Driving experience tells us that as long as either of these two factors is zero (i.e., the distance is far enough or the deviation angle is large enough), there is no risk of collision. The closer the vehicle is to the obstacle, the larger the potential field function generated by the obstacle should be. Furthermore, the potential field function should be related to the relative position and velocity between the vehicle and the obstacle. Therefore, the left-turning vehicle is simplified as a point mass, and the relative position, angle, and velocity between the left-turning vehicle and the obstacle are determined. Based on the relative position, angle, and velocity, a repulsive potential field model of the moving obstacle is established. The calculation formula for the repulsive potential field model of the moving obstacle is as follows:

[0065]

[0066] Among them, U obs k represents the repulsive potential field generated by the obstacle vehicle. obs Let v0 and v be the repulsive potential field constants of the obstacle vehicle.i The speeds of the vehicle turning left and the vehicle blocking the obstacle are θ0 and θ, respectively. i These are the angles between the left-turning vehicle and the obstacle vehicle, respectively, and the line connecting them to their centers of mass; r0 is the furthest distance the obstacle vehicle can travel. ov This refers to the distance between a vehicle turning left and an obstacle vehicle.

[0067] In some embodiments, in addition to the repulsive field generated by the obstacle, please refer to Figure 6 A boundary repulsive potential field model is constructed on lane 2 where the vehicle travels to constrain its movement on the road. To make the left-turning process more realistic and establish a safe and standardized driving area, boundary constraints need to be introduced into the distance potential field model to ensure that the vehicle moves within a suitable area without affecting the driving of vehicles in other lanes. When turning at an intersection, to complete the turning process safely and comfortably, the turning vehicle must travel within an appropriate area, which is called the comfort zone. Therefore, boundary constraints need to be established in the trajectory planning of the left-turning vehicle. The boundary line of the left-turning comfort zone is established based on the original lane and the target lane. Among them, line segment AB is the inner boundary of the vehicle's turning driving. The outer lane line of the target lane and the right lane line of the original lane are extended to establish boundary lines CD and DE, respectively. CD is the upper boundary and DE is the right boundary. The boundary line of the left-turning comfort zone is determined. Based on the boundary line of the left-turning comfort zone, the distance of the left-turning vehicle to the boundary line is determined. Based on the distance of the left-turning vehicle to the boundary line, a boundary repulsive potential field model is established. The calculation formula of the boundary repulsive potential field model is as follows:

[0068]

[0069]

[0070]

[0071] Among them, U DE The repulsive potential field generated by the boundary DE, U CD The repulsive potential field generated by the boundary CD, U AB Let be the repulsive potential field generated by the boundary line AB, k1, k2, and k3 be the constants of the boundary potential fields of the boundaries DE, CD, and AB, respectively, x, y, and z be the distances of the left-turning vehicle to the boundaries DE, CD, and AB, and x0, y0, and z0 be the distances of the left-turning vehicle from the boundary repulsive potential field set.

[0072] In some embodiments, in step S200, a vehicle trajectory prediction model is established, and based on the vehicle trajectory prediction model, the surrounding environment information and vehicle parameters of surrounding vehicles are predicted in real time. The vehicle parameters of surrounding vehicles include at least speed, steering angle, and position coordinates.

[0073] In some embodiments, in step S300, based on the vehicle parameters of surrounding vehicles, when the autonomous vehicle makes a left turn inside the intersection, an improved artificial potential field model is used to plan the left-turning vehicle path and perform real-time left-turn trajectory control. To consider the vehicle's kinematic constraints, the improved artificial potential field method simplifies the car to a point mass, retaining the center of mass position, velocity vector, and heading angle parameters of the left-turning vehicle. The artificial potential field is placed in a global coordinate system, and the coordinates of obstacles in the artificial potential field are consistent with the actual environment. The forces in the artificial potential field act on the center of mass position, guiding the vehicle's movement. The vehicle's steering action is taken into account in the path planning to ensure the rationality of the planned path. Please refer to [link to relevant documentation]. Figure 7 , Figure 7 This is a schematic diagram of the vehicle's trajectory during each time interval ΔT, where α(T) and α(T+ΔT) are the vehicle's heading angles at time T and T+ΔT, respectively. In the simplified vehicle kinematic model, the vehicle's velocity direction and heading angle are consistent. The resultant force provides the vehicle with acceleration changes, determining the velocity direction of the next step, i.e., the heading angle. The formula for calculating the speed of a vehicle turning left is:

[0074]

[0075] v n+1 =v n +a n t,

[0076] Among them, a n F is the acceleration vector of the vehicle turning left. SUM Let m be the resultant force of the left-turning vehicle, t be the mass of the left-turning vehicle, and v be the set time step for trajectory updates. n v is the speed of a vehicle turning left. n+1 This indicates the next speed for vehicles turning left.

[0077] In some embodiments, the specific steps for planning the path of left-turning vehicles using the improved artificial potential field model are as follows:

[0078] Step 1: Set the coordinates of the target point for vehicles turning left;

[0079] Step 2: Predict vehicle parameters of surrounding vehicles in real time;

[0080] Step 3: Determine the path planning for left-turning vehicles, i.e., check whether the left-turning vehicles have reached the target point. When the left-turning vehicles have reached the target point, the path planning is completed. When the left-turning vehicles have not reached the target point, determine the magnitude and direction of the attractive force, repulsive force, and resultant force of the left-turning vehicles.

[0081] Step 4: When the net force of the left-turning vehicle is 0, that is, when the planned path is trapped in a local optimum, adjust the potential field coefficient K value of the gravitational potential field model, and recalculate the magnitude and direction of the net force on the left-turning vehicle in the artificial potential field until the net force is not 0.

[0082] Step 5: Calculate the acceleration and steering angle of the left-turning vehicle. When the steering angle of the left-turning vehicle exceeds the kinematic constraints, use the maximum steering angle to calculate the next heading angle of the left-turning vehicle. Use the maximum acceleration to calculate the speed of the left-turning vehicle. Based on the speed of the left-turning vehicle and the next heading angle, calculate the next position of the left-turning vehicle and update the vehicle position coordinates of the left-turning vehicle.

[0083] Step 6: Based on the vehicle position coordinates of the left-turning vehicle and the vehicle parameters of the surrounding vehicles, determine the collision risk between the left-turning vehicle and the surrounding vehicles. When the left-turning vehicle at the next position collides with the surrounding vehicles, increase the potential field coefficient K value of the repulsive potential field model and repeat steps 1 to 6 until the left-turning vehicle reaches the target point.

[0084] To ensure the smoothness of the planned path, a series of path points are formed by dense coordinates of the vehicle's centroid. Different calculation step sizes need to be set for different vehicle speeds; the faster the vehicle speed, the smaller the step size.

[0085] Based on the above-described method for planning and controlling left-turn routes for vehicles at intersections, this invention also provides a corresponding system for planning and controlling left-turn routes for vehicles at intersections. Please refer to [link / reference needed]. Figure 2 The system 400 for planning and controlling the left-turn path of vehicles at the intersection includes an artificial potential field model establishment module 410, a surrounding vehicle parameter prediction module 420, and a left-turn vehicle path planning module 430.

[0086] The artificial potential field model establishment module 410 is used to establish an improved artificial potential field model, wherein the improved artificial potential field model includes at least a gravitational potential field model and an improved repulsive potential field model, and the improved repulsive potential field model includes at least a moving obstacle repulsive potential field model and a boundary repulsive potential field model.

[0087] The surrounding vehicle parameter prediction module 420 is used to establish a vehicle trajectory prediction model and predict the vehicle parameters of surrounding vehicles in real time based on the vehicle trajectory prediction model. The vehicle parameters include at least speed, steering angle and position coordinates.

[0088] The left-turning vehicle path planning module 430 is used to plan the path of left-turning vehicles based on vehicle parameters using an improved artificial potential field model. When the path planned by the improved artificial potential field model gets stuck in a local optimum, the potential field coefficient of the gravitational potential field model is adjusted to determine the magnitude and direction of the net force on the left-turning vehicle and determine the next position of the left-turning vehicle. When the left-turning vehicle at the next position collides with the surrounding vehicles, the potential field coefficient of the repulsive potential field model is adjusted. The process is iterated until the left-turning vehicle reaches the destination point.

[0089] In summary, the method and system for planning and controlling left-turning paths at intersections provided by this invention first establishes an improved artificial potential field model, which includes at least an gravitational potential field model and an improved repulsive potential field model. The improved repulsive potential field model includes at least a repulsive potential field model for moving obstacles and a boundary repulsive potential field model. Second, a vehicle trajectory prediction model is established. Based on this model, the vehicle parameters of surrounding vehicles are predicted in real time, including at least speed, steering angle, and position coordinates. Finally, based on the vehicle parameters of surrounding vehicles, the improved artificial potential field model is used to plan the path for left-turning vehicles. When the path planned by the improved artificial potential field model gets stuck in a local optimum, the potential field coefficient of the gravitational potential field model is adjusted to determine the magnitude and direction of the net force on the left-turning vehicle, thus determining the next position of the left-turning vehicle. When the left-turning vehicle at the next position collides with surrounding vehicles, the potential field coefficient of the repulsive potential field model is adjusted. This process is iterated until the left-turning vehicle reaches its destination, improving the overall traffic efficiency of the intersection.

[0090] The specific embodiments of the present invention described above do not constitute a limitation on the scope of protection of the present invention. Any other corresponding changes and modifications made in accordance with the technical concept of the present invention should be included within the scope of protection of the claims of the present invention.

Claims

1. A method for planning and controlling left-turn routes for vehicles at intersections, characterized in that, Includes the following steps: An improved artificial potential field model is established, wherein the improved artificial potential field model includes at least a gravitational potential field model and an improved repulsive potential field model, and the improved repulsive potential field model includes at least a moving obstacle repulsive potential field model and a boundary repulsive potential field model. A vehicle trajectory prediction model is established, and based on the vehicle trajectory prediction model, the vehicle parameters of surrounding vehicles are predicted in real time, wherein the vehicle parameters include at least speed, steering angle and position coordinates. Based on the vehicle parameters, the improved artificial potential field model is used to plan the path for the left-turning vehicle. When the path planned by the improved artificial potential field model gets stuck in a local optimum, the potential field coefficient of the gravitational potential field model is adjusted to determine the magnitude and direction of the net force on the left-turning vehicle and to determine the next position of the left-turning vehicle. When the left-turning vehicle at the next position collides with surrounding vehicles, the potential field coefficient of the repulsive potential field model is adjusted. This process is iterated until the left-turning vehicle reaches its destination. The planning of the left-turning vehicle path using the improved artificial potential field model based on the vehicle parameters includes: Step 1: Set the coordinates of the target point for vehicles turning left; Step 2: Predict vehicle parameters of surrounding vehicles in real time; Step 3: Determine the path planning for left-turning vehicles. When the left-turning vehicle has not completed the path planning, determine the magnitude and direction of the attractive force, repulsive force, and resultant force acting on the left-turning vehicle. Step 4: When the net force on the left-turning vehicle is 0, adjust the potential field coefficient of the gravitational potential field model and recalculate the magnitude and direction of the net force on the left-turning vehicle in the artificial potential field until the net force is not 0. Step 5: Determine the acceleration and steering angle of the left-turning vehicle. When the steering angle of the left-turning vehicle exceeds the kinematic constraints, use the maximum acceleration and maximum steering angle to calculate the next position of the left-turning vehicle and update the vehicle position coordinates. Step 6: Based on the vehicle position coordinates of the left-turning vehicle and the vehicle parameters of the surrounding vehicles, determine the collision risk between the left-turning vehicle and the surrounding vehicles. When the left-turning vehicle collides with the surrounding vehicles at the next position, adjust the potential field coefficient of the repulsive potential field model and repeat steps 1 to 6 until the left-turning vehicle reaches the target point.

2. The method for planning and controlling left-turn routes for vehicles at intersections according to claim 1, characterized in that, The gravitational potential field model includes: Determine the distance from the left-turning vehicle to the target point, and establish a gravitational potential field model based on the distance.

3. The method for planning and controlling left-turn routes for vehicles at intersections according to claim 2, characterized in that, The calculation formula for the gravitational potential field model is as follows: ， ， in, Let be the gravitational potential field of the target point with respect to the left-turning vehicle. Let be a constant of the gravitational potential field. The distance between the target vehicle and the obstacle. The gravitational potential field factor, The gravitational force exerted by the target point on the vehicle is the negative gradient of the gravitational potential energy.

4. The method for planning and controlling left-turn routes for vehicles at intersections according to claim 1, characterized in that, The repulsive potential field model of the moving obstacle includes: The relative position, angle, and speed between the left-turning vehicle and the obstacle are determined, and a repulsive potential field model of the moving obstacle is established based on the relative position, angle, and speed.

5. The method for planning and controlling left-turn routes for vehicles at intersections according to claim 4, characterized in that, The calculation formula for the repulsive potential field model of the moving obstacle is: ， in, The repulsive potential field generated by the obstacle vehicle. Let be the constant of the repulsive potential field of the obstacle vehicle. and The speeds of vehicles turning left and vehicles obstructing traffic are respectively. and These are the angles between the left-turning vehicle and the obstructing vehicle, respectively, and the line connecting them to their centers of mass. The furthest distance at which the obstacle vehicle acts. This refers to the distance between a vehicle turning left and an obstacle vehicle.

6. The method for planning and controlling left-turn routes for vehicles at intersections according to claim 1, characterized in that, The boundary repulsive potential field model includes: Determine the boundary line of the left-turn comfort zone, determine the distance of the left-turning vehicle to the boundary line based on the boundary line, and establish a boundary repulsive potential field model based on the distance of the left-turning vehicle to the boundary line.

7. The method for planning and controlling left-turn routes for vehicles at intersections according to claim 6, characterized in that, The calculation formula for the boundary repulsive potential field model is as follows: ， ， ， in, The repulsive potential field generated at the boundary DE The repulsive potential field generated by the boundary CD The repulsive potential field generated by the boundary line AB , , Let be the constants of the boundary potential fields at boundaries DE, CD, and AB, respectively, and let x, y, and z be the distances of the left-turning vehicle to boundaries DE, CD, and AB. , , The distance between the left-turning vehicle and the boundary repulsive potential field.

8. The method for planning and controlling left-turn routes for vehicles at intersections according to claim 1, characterized in that, The formula for calculating the speed of the vehicle turning left is: ， ， in, For left-turning vehicles, For the resultant force of vehicles turning left, For the mass of vehicles turning left, The set time step for trajectory updates. The speed of vehicles turning left. This indicates the next speed for vehicles turning left.

9. A system for planning and controlling left-turn routes for vehicles at intersections, characterized in that, include: An artificial potential field model establishment module is used to establish an improved artificial potential field model, wherein the improved artificial potential field model includes at least a gravitational potential field model and an improved repulsive potential field model, and the improved repulsive potential field model includes at least a moving obstacle repulsive potential field model and a boundary repulsive potential field model. The surrounding vehicle parameter prediction module is used to establish a vehicle trajectory prediction model and, based on the vehicle trajectory prediction model, predict the vehicle parameters of surrounding vehicles in real time. The vehicle parameters include at least speed, steering angle, and position coordinates. A left-turning vehicle path planning module is used to plan the path of a left-turning vehicle based on the vehicle parameters and using the improved artificial potential field model. When the path planned by the improved artificial potential field model gets stuck in a local optimum, the potential field coefficient of the gravitational potential field model is adjusted to determine the magnitude and direction of the net force on the left-turning vehicle and determine the next position of the left-turning vehicle. When the left-turning vehicle at the next position collides with surrounding vehicles, the potential field coefficient of the repulsive potential field model is adjusted. This process is iterated until the left-turning vehicle reaches the destination point. The planning of the left-turning vehicle path based on the vehicle parameters and using the improved artificial potential field model includes: Step 1: Set the coordinates of the target point for vehicles turning left; Step 2: Predict vehicle parameters of surrounding vehicles in real time; Step 3: Determine the path planning for left-turning vehicles. When the left-turning vehicle has not completed the path planning, determine the magnitude and direction of the attractive force, repulsive force, and resultant force acting on the left-turning vehicle. Step 4: When the net force on the left-turning vehicle is 0, adjust the potential field coefficient of the gravitational potential field model and recalculate the magnitude and direction of the net force on the left-turning vehicle in the artificial potential field until the net force is not 0. Step 5: Determine the acceleration and steering angle of the left-turning vehicle. When the steering angle of the left-turning vehicle exceeds the kinematic constraints, use the maximum acceleration and maximum steering angle to calculate the next position of the left-turning vehicle and update the vehicle position coordinates. Step 6: Based on the vehicle position coordinates of the left-turning vehicle and the vehicle parameters of the surrounding vehicles, determine the collision risk between the left-turning vehicle and the surrounding vehicles. When the left-turning vehicle collides with the surrounding vehicles at the next position, adjust the potential field coefficient of the repulsive potential field model and repeat steps 1 to 6 until the left-turning vehicle reaches the target point.

Images