基于贝塞尔曲线的动态识别区农机避障路径实时规划

To solve the problem of real-time path planning for agricultural machinery to avoid obstacles during field operations, this research proposes an algorithm for real-time planning of obstacle avoidance paths for agricultural machinery using cubic Bezier curves in a dynamic recognition area. Firstly, a dynamic recognition area for the walking of agricultural machinery operations is constructed, and obstacles are perceived by a laser radar in the dynamic recognition area. Then, the selection range of control points for the obstacle avoidance path is calculated using the obstacle information, and a path cluster satisfying multiple constraints such as the minimum turning radius of the agricultural machinery is generated. At the same time, the optimal obstacle avoidance path is selected from the path cluster with the goal of minimizing curvature. Finally, a real-time obstacle avoidance path planning experiment is carried out. The experimental results show that the maximum curvature and average curvature of the obstacle avoidance path planned by the algorithm in this paper are 0.126 and 0.054 m⁻¹ respectively; the maximum lateral error and average lateral error generated during the path tracking process are 0.12 and 0.057 m respectively; and the distance from the tractor to the outer contour of the obstacle is greater than 0.375 m. Compared with existing algorithms, the maximum curvature and average curvature of the obstacle avoidance path planned by the algorithm in this paper are reduced by 25.9% and 42.6% respectively, and the maximum lateral error and average lateral error generated during the path tracking process are reduced by 36.8% and 28.8% respectively. The research results can provide technical support for unmanned tractor operations.

为解决农机田间作业避障路径实时规划问题,该研究提出一种在动态识别区内利用三阶贝塞尔曲线实时规划农机避障路径算法。首先构建农机作业行走动态识别区,在动态识别区内利用激光雷达感知障碍物。然后利用障碍物信息计算避障路径控制点选取范围,生成满足农机最小转弯半径等多约束条件下的路径簇,同时以曲率最小为目标从路径簇中选取最优避障路径。最后进行避障路径实时规划试验。试验结果表明,本文算法规划的避障路径最大曲率和平均曲率分别为0.126和0.054 m~(-1);路径跟踪过程中产生的最大横向误差和平均横向误差分别为0.12和0.057 m;拖拉机到障碍物外轮廓的距离大于0.375 m。和现有算法比较,本文算法规划的避障路径最大曲率和平均曲率分别减少25.9%和42.6%,路径跟踪过程中产生的最大横向误差和平均横向误差分别减少36.8%和28.8%。研究结果可为拖拉机无人驾驶作业提供技术支撑。