Authors

Abstract

Mechanical devices operating in vineyards are subject to interference from obstacles such as vine branches and leaves, reducing fruit-picking efficiency. To address stable obstacle avoidance, an improved RRT algorithm based on global adaptive step size with target bias sampling was developed. First, the kinematic equations for the grape-picking robotic arm were established using the D-H method, and forward and inverse kinematic calculations were performed. Matlab software was used to verify the accuracy of the kinematic analysis. To overcome the traditional RRT algorithm's limitations in planning collision-free paths, dynamic update and global adaptive step-size strategies were proposed, and MATLAB software was used to perform simulation experiments. The results demonstrated that the improved RRT algorithm, compared to the traditional RRT, RRT_informed, and RRT_star algorithms in both 2D and 3D map scenarios, had advantages such as reduced planning time, fewer sampling points, and shorter path length. Finally, grape-picking tests conducted in both laboratory and orchard environments showed that our improved RRT algorithm significantly reduced planning time for the fruit-picking path, verifying its effectiveness.

Abstract in Chinese

机械设备在葡萄果园环境中作业会受到藤枝叶等障碍物的干扰，导致果实采摘效率低。为实现稳定避障，研究出一种基于全局自适应步长与目标偏置采样的改进型RRT算法。首先，通过D-H法建立了葡萄采摘机械臂运动学方程，并进行了正、逆运动学计算，且应用Matlab软件验证机械臂运动学分析的正确性。然后，针对传统RRT算法在规划无碰撞路径时缺乏目标导向性等问题，提出了动态更新和全局自适应步长策略，应用MATLAB软件进行了仿真实验，验证了我们改进后的RRT算法相对于传统RRT算法、RRT_informed算法和RRT_star算法在二维和三维地图场景中，具有规划耗时低、采样点个数少以及路径长度短的优点。最后，分别在实验室和真实果园进行了葡萄采摘试验，结果表明，我们研究的改进型RRT算法在应用于葡萄机械臂中，其规划采摘果实路径时间较短，验证了算法的有效性。