Authors

Abstract

Aiming at the suboptimal local path, slow convergence speed, and many inflection points in the path planning of fruit and vegetable picking robots in complex environments, a global planning method combining particle swarm optimization (PSO) algorithm and A* algorithm was proposed. Firstly, Manhattan distance was taken as a heuristic function of global programming based on the A* algorithm. Secondly, the step size of PSO was adjusted to optimize the path, shorten the path length, and reduce the number of inflection points. Finally, the planned path of the fruit and vegetable picking robot was smoothed so that it could steadily move along a smoother driving path in real scenarios. The experimental results show that compared with the traditional PSO algorithm, the hybrid algorithm based on the improved A* algorithm and PSO algorithm achieves a smoother path and fewer folding points. In comparison with the PSO algorithm, moreover, this algorithm can guarantee the path generation efficiency and the global optimum. In the end, the effectiveness of the proposed method was verified by shortening the path length and reducing the accumulative number of inflection points.

Abstract in Chinese

针对复杂环境下果蔬采摘机器人路径规划中存在的局部路径次优、收敛速度慢、拐点多等问题，提出了一种基于粒子群优化算法和A*算法相结合的全局规划方法。首先，基于A*算法，将曼哈顿距离作为全局规划的启发式函数。其次，通过调整粒子群算法的步长来优化路径，缩短路径长度，减少拐点数量。最后，对果蔬采摘机器人规划路径进行平滑处理，使机器人能够在真实场景中平稳移动，使采摘机器人行驶路径更加平滑。实验结果表明，与传统的粒子群优化算法相比，基于改进A*算法和粒子群算法的混合算法具有更平滑的路径和更少的折叠点。与粒子群优化算法相比，该算法能够保证生成路径的效率和全局最优性。通过缩短路径长度和减少累积拐点数量，验证了该方法的有效性。