Obstacle avoiding method for low-speed unmanned vehicle
By calculating the vehicle angle trajectory and grid map to optimize the steering angle, the generalization and real-time problems of unmanned vehicle obstacles are solved, the safety and flexibility of obstacles are improved, and the calculation redundancy and resource utilization are reduced.
Patent Information
- Application Number
- CN202510811055.9
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-06-17
- Publication Date
- 2025-08-08
AI Technical Summary
The existing obstacle-blocking methods for unmanned vehicles are insufficient in generalization capabilities, real-time and safety, especially in complex environments and moving obstacles, which are difficult to effectively avoid obstacles.
By calculating the isometric trajectories of all feasible angles of the vehicle's front and rear axles, a local grid map is established, an obstacle is imported and an expansion area is set, traversing the trajectory is found to find the nearest collision point, combining the multi-dimensional cost function to optimize the steering angle, balance the target orientation, manipulation continuity and obstacle avoidance intensity, and filtering unoccupied trajectory segments to reduce redundant calculations.
It significantly improves the safety, flexibility and real-time nature of unmanned vehicle obstacles, reduces processor load, adapts to complex and changeable environments, and achieves efficient computing and low resource utilization.
Smart Images
Figure CN120440075A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the technical field of low-speed unmanned driving, and in particular to an obstacle avoidance method for a low-speed unmanned driving vehicle. Background Art
[0002] The driving safety and passability of autonomous vehicles have always been a hot topic and a challenging issue in research. The related technologies and methods are also constantly being upgraded. First, traditional PID control: A PID controller adjusts the steering angle or speed, using obstacle distance as feedback to calculate the steering angle. PID parameters require manual tuning for specific vehicle models and scenarios, resulting in poor generalization and requiring recalibration when the vehicle platform is changed or the environment changes. Next are the improved A* and RRT path planning algorithms: A* uses heuristic search to generate the global shortest path, while RRT (Rapid Random Tree) expands path branches in complex environments. Some research combines the two to create a hybrid A* algorithm, balancing efficiency and feasibility. The A* algorithm requires replanning the path and cannot respond to moving obstacles in real time.
[0003] A further approach is the model prediction method: predicting the future positions of the vehicle and obstacles and optimizing the cost function to avoid moving obstacles. This method has many parameters, is complex to implement, and may have no solution.
[0004] Therefore, a method for avoiding obstacles for low-speed unmanned vehicles is proposed. Summary of the Invention
[0005] In response to the shortcomings of the existing technology, the present invention provides a method for low-speed unmanned vehicle obstacle avoidance, which has the advantages of calculating the optimal feasible steering angle, including determining the feasible trajectory, establishing a grid map, placing obstacles into the map, finding the nearest collision point, updating J, and determining the steering angle, thus solving the limitations of existing unmanned vehicle obstacle avoidance technology. Technical Solution
[0006] To achieve the above object, the present invention provides the following technical solutions: A method for avoiding obstacles for a low-speed unmanned vehicle comprises the following steps: Step 1: Calculate the equal length trajectories of all possible turning angles of the front and rear axles: First, the length of each trajectory and the number of trajectory points are set, and the front and rear axle trajectories are calculated; then the front and rear axle trajectories are generated respectively; finally, the front and rear axle trajectory points are stored; Step 2: Create a local grid map; Step 3: Import obstacles into the grid map and set: First, the obstacle coordinates are converted to a Cartesian coordinate system and imported into a grid map. Then, the obstacle expansion is set. Finally, the occupied points within the vehicle in the expanded map are processed. Step 4: Traverse all trajectories on the grid map and find the closest point on the trajectory that collides with the obstacle: First, obtain all occupied points of the trajectory on the grid map; then calculate the distance of the nearest occupied point in the trajectory from the car along the trajectory; Step 5: For all trajectories that match the actual vehicle steering, calculate and update the performance index: First, design the performance indicator function; Then, the trajectory with the minimum distance between the obstacle and the front and rear axes greater than the specified threshold is selected; Then take the current trajectory and the 5 adjacent trajectories on its left and right, and calculate the performance index J; Finally, update the minimum index value J and record the corresponding angle; Step 6: After traversing all corners, output the angle corresponding to the minimum index to complete the work steps Preferably, in step 1, the turning radius of different turning angles is first calculated according to the Ackerman steering principle: The formula for calculating the turning radius is: ; Where wheelbase is the wheelbase and θ is the turning angle.
[0007] Preferably, in step 1, the length s of each trajectory and the number of trajectory points are set, and the front and rear axle trajectories are calculated as: Calculate the coordinates of the rear axle trajectory points: , where s is the arc length along the trajectory; Calculate the coordinates of the front axle trajectory points: , the front axle position is the wheelbase extended along the tangent direction of the rear axle position.
[0008] Preferably, the step 2 is specifically to establish an occupancy grid map with the center of the rear axle of the vehicle as the origin, 5m in front of the vehicle, 3m behind the vehicle, and 4m on each side.
[0009] Preferably, in step 4, if the entire trajectory is not occupied, the last point is set to be occupied.
[0010] Preferably, the performance indicator function in step 5 is: ; φ, θ, θ * They are the input desired turning angle, the output turning angle corresponding to the current trajectory, and the last output turning angle.
[0011] Preferably, the X f and X rare the distances from the front and rear axes along the trajectory to the obstacle, respectively; Q is the coefficient for calculating the distance from the adjacent trajectory to the obstacle; j, q, and k are all weight coefficients.
[0012] Compared with the prior art, the present invention provides a method for avoiding obstacles for a low-speed unmanned vehicle, which has the following beneficial effects: 1. This low-speed unmanned vehicle obstacle avoidance method significantly improves the safety, flexibility, and real-time performance of unmanned vehicle obstacle avoidance through comprehensive environmental perception, target direction constraints, and trajectory feasibility analysis.
[0013] 2. This low-speed unmanned vehicle obstacle avoidance method traverses 131 trajectories, filters unoccupied trajectory segments to record the nearest available points, and combines rapid evaluation of local trajectories to reduce redundant calculations and lower processor load, thereby achieving efficient computing and low resource usage. BRIEF DESCRIPTION OF THE DRAWINGS
[0014] Figure 1 A schematic diagram of a grid map with obstacle coordinates according to the present invention; Figure 2 Schematic diagram of equal-length trajectories of all feasible turning angles of the front and rear axles of the vehicle of the present invention; Figure 3 This is a schematic diagram of a grid map after the obstacle expansion coefficient is determined in the present invention; Figure 4 Schematic diagram of the optimal turning angle of the front and rear axles of the vehicle of the present invention. DETAILED DESCRIPTION
[0015] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, 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 making creative efforts are within the scope of protection of the present invention.
[0016] See also Figures 1-4 ; Example
[0017] A method for avoiding obstacles for a low-speed unmanned vehicle comprises the following steps: Step 1: Calculate the equal length trajectories of all possible turning angles of the front and rear axles: The length of each track is set to 3m, with track points spaced 0.1m apart; Generate front and rear axle trajectories for angles from -60° to 60°, one for each degree, for a total of 121 trajectories.
[0018] The inner side is the rear axle track, and the outer side is the front axle track. Figure 2 As shown; Then store the front and rear axle trajectory points; Create a local grid map, import obstacles into the grid map and set them.
[0019] Convert the obstacle coordinates from polar coordinates to Cartesian coordinates and import them into the grid map; Set the obstacle expansion coefficient to 0.6, such as Figure 3 : Set the occupied points inside the vehicle in the expanded map to 0.
[0020] Traverse all trajectories on the grid map and find the closest point on the trajectory that collides with the obstacle: Get the occupied points of the trajectory points on the grid map; Calculate the distance of the nearest occupied point in the trajectory from the car along the trajectory; If the entire trajectory is unoccupied, the last point is set to be occupied; For all trajectories that match the actual vehicle steering, calculate and update the performance index: Design performance indicator function:
[0021] φ, θ, θ * are the input desired turning angle, the output turning angle corresponding to the current trajectory, and the last output turning angle; X f and X r are the distances from the front and rear axes along the trajectory to the obstacle, respectively, and Q is the coefficient for calculating the distance from the adjacent trajectory to the obstacle; The weight coefficients of j, q, and k are set to 5, 2, and 2 respectively.
[0022] Based on parameters such as maximum vehicle speed and turning radius, only the trajectories where the minimum distance between the obstacle and the front and rear axles is greater than 0.5 are taken for calculation.
[0023] Take the current trajectory and its five adjacent trajectories on the left and right, and calculate the performance index J.
[0024] Update the minimum index value J and record the corresponding angle.
[0025] After traversing all corners, output the angle corresponding to the minimum index.
[0026] Figure 4 is the optimal turning angle calculated under the current conditions.
[0027] This method balances the following key indicators by introducing a multi-dimensional cost function (including target direction deviation, steering continuity constraint and trajectory availability weighting): Goal-oriented: By Prioritize approaching the target angle, reduce path deviation, and solve the problem of target direction deviation; Control Continuity: Utilize Restrain sudden changes in steering angle and improve driving stability; Obstacle avoidance strength: pass item Dynamically weight the average collision distance of the five adjacent trajectories (Q is an array of weight factors) to enhance decision robustness; This method traverses 131 trajectories, filters unoccupied segments, and records the nearest available point. This method, combined with rapid evaluation of local trajectories (five adjacent left and right trajectories), reduces redundant computation and processor load, resulting in efficient computation and low resource usage. Furthermore, this method offers real-time environment modeling and dynamic obstacle avoidance capabilities. It utilizes a binary map to update environmental information in real time, and dynamically adjusts obstacle zones through a series of functions. Compared to traditional static maps, this method can adapt to complex and changing scenarios, significantly improving the real-time and robustness of obstacle avoidance.
[0028] The beneficial effects of the present invention are as follows: the obstacle avoidance method for low-speed unmanned vehicles significantly improves the safety, flexibility and real-time performance of unmanned vehicles by integrating environmental perception, target direction constraints and trajectory feasibility analysis. By traversing 131 trajectories and screening unoccupied trajectory segments to record the nearest available points, combined with rapid evaluation of local trajectories, it reduces redundant calculations and reduces processor load, thereby achieving efficient computing and low resource usage.
[0029] Although the embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the invention, and that the scope of the invention is defined by the appended claims and their equivalents.
Claims
1. A method for avoiding obstacles for a low-speed unmanned vehicle, characterized in that: The following steps are involved: Step 1: Calculate the equal length trajectories of all possible turning angles of the front and rear axles: First, the length of each trajectory and the number of trajectory points are set, and the front and rear axle trajectories are calculated; then the front and rear axle trajectories are generated respectively; finally, the front and rear axle trajectory points are stored; Step 2: Create a local grid map; Step 3: Import obstacles into the grid map and set: First, the obstacle coordinates are converted to a Cartesian coordinate system and imported into a grid map. Then, the obstacle expansion is set. Finally, the occupied points within the vehicle in the expanded map are processed. Step 4: Traverse all trajectories on the grid map and find the closest point on the trajectory that collides with the obstacle: First, obtain all occupied points of the trajectory on the grid map; then calculate the distance of the nearest occupied point in the trajectory from the car along the trajectory; Step 5: For all trajectories that match the actual vehicle steering, calculate and update the performance index: First, design the performance indicator function; Then, the trajectory with the minimum distance between the obstacle and the front and rear axes greater than the specified threshold is selected; Then take the current trajectory and the 5 adjacent trajectories on its left and right, and calculate the performance index J; Finally, update the minimum index value J and record the corresponding angle; Step 6: After traversing all corners, output the angle corresponding to the minimum index to complete the work steps.
2. The obstacle avoidance method for a low-speed unmanned vehicle according to claim 1, characterized in that: In step 1, the turning radius at different angles is calculated according to the Ackermann steering principle: The formula for calculating the turning radius is: ; Where wheelbase is the wheelbase and θ is the turning angle.
3. The obstacle avoidance method for a low-speed unmanned vehicle according to claim 1, characterized in that: In step 1, the length s of each trajectory and the number of trajectory points are set, and the front and rear axle trajectories are calculated as: Calculate the coordinates of the rear axle trajectory points: , where s is the arc length along the trajectory; Calculate the coordinates of the front axle trajectory points: , the front axle position is the wheelbase extended along the tangent direction of the rear axle position.
4. The obstacle avoidance method for a low-speed unmanned vehicle according to claim 1, characterized in that: The second step is to establish an occupancy grid map with the center of the rear axle of the vehicle as the origin, 5m in front of the vehicle, 3m behind the vehicle, and 4m on the left and right.
5. The obstacle avoidance method for a low-speed unmanned vehicle according to claim 1, characterized in that: If the entire trajectory is not occupied in step 4, the last point is set to be occupied.
6. The obstacle avoidance method for a low-speed unmanned vehicle according to claim 1, characterized in that: The performance indicator function in step 5 is: ； φ, θ, θ * They are the input desired turning angle, the output turning angle corresponding to the current trajectory, and the last output turning angle.
7. The obstacle avoidance method for a low-speed unmanned vehicle according to claim 6, characterized in that: The X f and X r are the distances from the front and rear axes along the trajectory to the obstacle, respectively; Q is the coefficient for calculating the distance from the adjacent trajectory to the obstacle; j, q, and k are all weight coefficients.