Zhang Wenyu, Zeng Ye, Wang Sifan, Wang Tao, Li Haomin, Fei Ke, Qiu Xinrui, Jiang Runpeng, Li Jun
College of Engineering, South China Agricultural University, Guangzhou, China.
Guangdong Laboratory for Lingnan Modern Agriculture, Guangzhou, China.
Front Plant Sci. 2023 Jul 21;14:1184352. doi: 10.3389/fpls.2023.1184352. eCollection 2023.
In orchard scenes, the complex terrain environment will affect the operational safety of mowing robots. For this reason, this paper proposes an improved local path planning algorithm for an artificial potential field, which introduces the scope of an elliptic repulsion potential field as the boundary potential field. The potential field function adopts an improved variable polynomial and adds a distance factor, which effectively solves the problems of unreachable targets and local minima. In addition, the scope of the repulsion potential field is changed to an ellipse, and a fruit tree boundary potential field is added, which effectively reduces the environmental potential field complexity, enables the robot to avoid obstacles in advance without crossing the fruit tree boundary, and improves the safety of the robot when working independently. The path length planned by the improved algorithm is 6.78% shorter than that of the traditional artificial potential method, The experimental results show that the path planned using the improved algorithm is shorter, smoother and has good obstacle avoidance ability.
在果园场景中，复杂的地形环境会影响割草机器人的作业安全。为此，本文提出一种改进的人工势场局部路径规划算法，引入椭圆斥力势场范围作为边界势场。势场函数采用改进的可变多项式并添加距离因子，有效解决了目标不可达和局部极小值问题。此外，将斥力势场范围改为椭圆，并添加果树边界势场，有效降低了环境势场复杂度，使机器人能够提前避开障碍物而不越过果树边界，提高了机器人自主作业时的安全性。改进算法规划的路径长度比传统人工势场法短6.78%，实验结果表明，采用改进算法规划的路径更短、更平滑，具有良好的避障能力。
