RRT vehicle path planning method based on vehicle trajectory prediction in V2X environment

Abstract

The invention discloses a vehicle path planning method in a V2X environment, and particularly relates to an RRT vehicle path planning method based on vehicle trajectory prediction in the V2X environment. The method mainly comprises the following steps of initializing a starting point and an ending point of an algorithm, acquiring driving information of surrounding vehicles by using a V2X technology, generating an initial sampling area, predicting tracks of the surrounding vehicles in a period of time in the future according to a vehicle track prediction algorithm, and taking the predicted tracks as predicted virtual obstacles. Adjusting the initially generated sampling area, adopting an optimized sampling area strategy to obtain an optimized sampling area, taking a starting point as a root node of the tree, randomly generating a new node in a fan-shaped area with a specified direction angle at a fixed step length, and checking whether the new node intersects with an obstacle or not. And if no collision exists, adding the node and the path into the RRT tree. And if the new node is close to the terminal point, stopping searching. Otherwise, continuing iteration until the path is found or the maximum number of iterations is reached.

Description

Technical Field

[0001] The present invention patent discloses a vehicle path planning method in a V2X environment, and in particular, relates to an RRT vehicle path planning method based on vehicle trajectory prediction in a V2X environment. Background Art

[0002] Due to the large changes in traffic volume on highways, the high speed of vehicles, the complex and changeable environment, and the need for fast response and high reliability in lane change path planning to ensure safety during high-speed driving, traditional path planning algorithms currently have problems such as low planning efficiency and limited application scenarios, and cannot adapt to changes in different traffic environments and vehicle states. Redundant calculations and low efficiency in the planning process. At the same time, traditional algorithms have certain limitations in application scenarios of complex road conditions and high-density traffic, and it is difficult to meet the needs of efficient and safe lane changes.

[0003] For the problem of highway path planning, there are existing methods based on grid search, random sampling and geometric paths. The path planning algorithm based on grid search is to rasterize the known surrounding environment map, generate a raster map containing obstacle information and feasible domain information, and then use the search algorithm to find a drivable path with the lowest cost from the starting point to the target point. The random sampling method randomly generates sample points in the configuration space, connects the starting point, sample points, and target points for collision detection, and finally determines a collision-free path as the final trajectory. The geometric path planning algorithm refers to generating a smooth curve cluster through a certain function or interpolation fitting method, and then selecting an optimal path from the curve cluster based on a series of specific evaluation indicators.

[0004] In the existing random sampling RRT algorithm, the position and shape of obstacles will change over time in a dynamic environment, which requires the method to dynamically adjust the shape and position of obstacles in the sampling area according to the vehicle driving conditions. In a complex environment, random sampling is performed within a fixed sampling area to expand the tree structure, resulting in low safety of the planned path. Therefore, an efficient and safe highway lane change path planning method is urgently needed. Summary of the invention

[0005] To solve the above problems, the present invention provides an RRT vehicle path planning method based on vehicle trajectory prediction in a V2X environment, which aims to reduce the safety risk of vehicle lane changing on highways, improve the efficiency of vehicle path planning, and better cope with complex traffic environments and dynamically changing road conditions.

[0006] In order to achieve the above object, the present invention adopts the following technical scheme:

[0007] An embodiment of the present invention provides an RRT vehicle path planning method based on vehicle trajectory prediction in a V2X environment, comprising the following steps:

[0008] Step 1: Initialize the starting point and target of the algorithm;

[0009] Step 2: Use V2X technology to obtain the driving conditions of the vehicle and surrounding vehicles, generate an initial sampling area, use the trajectory prediction algorithm to predict the trajectory of the surrounding vehicles of the vehicle, and adjust the planned sampling area range based on the predicted virtual obstacles and road conditions according to the optimized sampling area strategy to obtain the optimized sampling area;

[0010] Step 3: The current node extends to the limited direction angle fan-shaped area to generate a new node;

[0011] Step 4: Check whether the newly generated node collides with obstacles in the map. If a collision occurs, the node is discarded. If the new node is valid, it is added to the RRT generation tree and the path from the current node to the new node is recorded.

[0012] Step 5: Evaluate the distance between the new node and the target point. If the new node is close enough to the target point, terminate the search process and complete the path planning.

[0013] Step 6: If no path is found or the maximum number of iterations is not reached, repeat the above steps until a feasible path from the starting point to the target point is found or the preset iteration limit is reached.

[0014] On the basis of the above technical solution, preferably, the predicted virtual obstacle is obtained by analyzing the historical motion data and current state of the vehicle, using a trajectory prediction algorithm to predict the future position and speed of the vehicle, and mapping these predictions onto the coordinate axes to transform them into virtual obstacles.

[0015] On the basis of the above technical solutions, preferably, the trajectory prediction algorithm includes a physical model-based method, a statistical learning algorithm, a deep learning method and a multimodal fusion method.

[0016] On the basis of the above technical solution, preferably, the optimization sampling area strategy is based on the current position and steering angle of the obstacle vehicle combined with the initial sampling area and the sampling area drawn by the vehicle prediction algorithm, and the calculation steps are:

[0017] The first step is to define the initial sampling area.

[0018]

[0019] Among them, X max ,Y max is the maximum horizontal and vertical coordinate value of the coordinate system, Represents two-dimensional real number space.

[0020] The second step is to calculate the minimum turning radius of each obstacle vehicle.

[0021]

[0022] Among them, φ i The current turning angle of the front wheel of the i-th obstacle vehicle, L i is the length of the i-th obstacle vehicle.

[0023] In the third step, the sampling area is defined based on the current position and steering angle of the obstacle vehicle.

[0024]

[0025] Among them, T is the set of predicted trajectories of obstacle vehicles, and each trajectory T i A series of points P (i,j) =(x i,j ,y i,j )composition.

[0026] The fourth step is to expand each predicted trajectory into the predicted obstacle area.

[0027]

[0028] Among them, W i is the width of the i-th obstacle car.

[0029] The fifth step is to calculate the final optimized sampling area.

[0030]

[0031] Compared with the prior art, the present invention has the following beneficial technical effects:

[0032] 1. Combining V2X technology and trajectory prediction algorithm to improve the real-time data processing capability of this method, the entire path planning system can quickly respond to environmental changes, significantly improving the real-time response capability of path planning, thereby more effectively responding to traffic changes and vehicle dynamics;

[0033] 2. Combined with V2X technology, this method can obtain the status information of surrounding vehicles in real time, avoid potential collision risks, effectively improve the safety of path planning, and ensure traffic safety during driving;

[0034] 3. Optimize the sampling area strategy, which can effectively handle large-scale vehicles and complex road networks, and has good system scalability and adaptability, suitable for the development needs of future intelligent transportation systems. BRIEF DESCRIPTION OF THE DRAWINGS

[0035] In order to more clearly illustrate the technical solution of the embodiment of the present invention, the specific implementation of the present invention is further described in detail below with reference to the accompanying drawings, wherein:

[0036] Figure 1 It is a flow chart of the RRT vehicle path planning method based on vehicle trajectory prediction in the V2X environment of the present invention;

[0037] Figure 2 It is a schematic diagram of the optimization sampling area strategy of the present invention;

[0038] Figure 3 is a schematic diagram of a virtual obstacle predicted by the present invention;

[0039] Figure 4 This is a schematic diagram of generating sampling points within the angle fan-shaped area specified in the patent of the present invention. DETAILED DESCRIPTION

[0040] In order to make the purpose, technical scheme and advantages of the present invention more clear, the present invention is further described in detail in conjunction with the illustrations. The described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0041] See also Figure 1 The embodiment of the present invention discloses a RRT vehicle path planning method based on vehicle trajectory prediction in a V2X environment, comprising the following steps:

[0042] Step 1: Set the starting point and target point of the algorithm;

[0043] Step 2: First, use V2X technology to obtain the speed, position, lateral and longitudinal acceleration, front wheel steering angle, and vehicle length and width of the vehicle and surrounding vehicles in real time to generate an initial sampling area;

[0044] Step 3: Based on the initial sampling area and the vehicle prediction algorithm, the trajectory prediction results of the vehicles around the vehicle are obtained. The predicted virtual obstacles are obtained by optimizing the sampling area strategy and the planned sampling area range is adjusted according to the road conditions to obtain the final planned sampling area. The trajectory prediction results of the vehicles around the vehicle are: S i (t)={t,x t ,y t ,L,W,v t ,a t}, the sampling area optimization strategy is as follows Figure 2 shown.

[0045] The sampling area optimization strategy is:

[0046] The first step is to define the initial sampling area.

[0047]

[0048] Among them, X max ,Y max is the maximum horizontal and vertical coordinate value of the coordinate system, Represents two-dimensional real number space.

[0049] The second step is to calculate the minimum turning radius of each obstacle vehicle.

[0050]

[0051] Among them, φ i The current turning angle of the front wheel of the i-th obstacle vehicle, L i is the length of the i-th obstacle vehicle.

[0052] In the third step, the obstacle zone is defined based on the vehicle’s current position and steering angle.

[0053]

[0054] Among them, T is the set of predicted trajectories of obstacle vehicles, and each trajectory T i A series of points P (i,j) =(x i,j ,y i,j )composition.

[0055] The fourth step is to expand each predicted trajectory into a predicted virtual obstacle area such as Figure 3 shown.

[0056]

[0057] Among them, W i is the width of the i-th obstacle car.

[0058] The fifth step is to calculate the optimized sampling area.

[0059]

[0060] Step 4: Generate sampling points within the limited angle fan-shaped area. The generated schematic diagram is as follows: Figure 4 As shown, check whether the newly generated node collides with the obstacles in the map. If a collision occurs, the node is discarded. If the new node is valid, it is added to the RRT generation tree and the path from the current node to the new node is recorded.

[0061] Step 5: Evaluate the distance between the new node and the target point. If the new node is close enough to the target point, terminate the search process and complete the path planning.

[0062] Step 6: If no path is found or the maximum number of iterations is not reached, repeat the above steps until a feasible path from the starting point to the target point is found or the preset iteration limit is reached.

Claims

1. A RRT vehicle path planning method based on vehicle trajectory prediction in a V2X environment, characterized in that: include: Step 1: Initialize the starting point and target of the algorithm; Step 2: Use V2X technology to obtain the driving conditions of the vehicle and surrounding vehicles, generate an initial sampling area, use the trajectory prediction algorithm to predict the trajectory of the surrounding vehicles of the vehicle, and adjust the planned sampling area range based on the predicted virtual obstacles and road conditions according to the optimized sampling area strategy to obtain the optimized sampling area; Step 3: The current node extends to the limited direction angle fan-shaped area to generate a new node; Step 4: Check whether the newly generated node collides with obstacles in the map. If a collision occurs, the node is discarded. If the new node is valid, it is added to the RRT generation tree and the path from the current node to the new node is recorded. Step 5: Evaluate the distance between the new node and the target point. If the new node is close enough to the target point, terminate the search process and complete the path planning. Step 6: If no path is found or the maximum number of iterations is not reached, repeat the above steps until a feasible path from the starting point to the target point is found or the preset iteration limit is reached.

2. The RRT vehicle path planning method based on vehicle trajectory prediction in a V2X environment as claimed in claim 1, characterized in that: In step 1, the initial sampling area is a sampling area drawn by radar, lidar, and camera based on data collected from the vehicle's own sensors using V2X communication technology.

3. The RRT vehicle path planning method based on vehicle trajectory prediction in a V2X environment as claimed in claim 1, characterized in that: In step 2, the optimized sampling area strategy is a sampling area drawn based on the current position and steering angle of the obstacle vehicle combined with the initial sampling area and the vehicle prediction algorithm.

4. The RRT vehicle path planning method based on vehicle trajectory prediction in a V2X environment as claimed in claim 1, characterized in that: In step 2, the trajectory prediction algorithm includes a physical model-based method, a statistical learning algorithm, a deep learning method, and a multimodal fusion method.

5. The RRT vehicle path planning method based on vehicle trajectory prediction in a V2X environment as claimed in claim 1, characterized in that: In step 2, the predicted virtual obstacle is achieved by analyzing the historical motion data and current state of the vehicle, using a trajectory prediction algorithm to predict the future position and speed of the vehicle, and mapping these predictions onto the coordinate axis to transform them into virtual obstacles.

Images