Authors

Abstract

Currently, drones have been gradually applied in the field of agriculture, and have been widely used in various types of agricultural aerial operations such as precision sowing, pesticide spraying, and vegetation detection. The use of agricultural UAVs for pesticide spraying has become an important task in the agricultural plant protection process. However, in the crop spraying process of agricultural UAVs, it is necessary to traverse multiple spray points and plan obstacle avoidance paths, which greatly affects the efficiency of agricultural UAV crop spraying operations. To address the above issues, traditional particle swarm optimization (PSO) algorithms have strong solving capabilities, but they are prone to falling into local optima. Therefore, this study proposes an improved PSO algorithm combined with the A* algorithm, which introduces a nonlinear convergence factor balancing algorithm for global search and local development capabilities in the traditional PSO algorithm, and adopts population initialization to enhance population diversity, so that the improved PSO algorithm has stronger model solving capabilities. This study designs two scenarios for agricultural UAV crop spraying path planning: one without obstacles and one with obstacles. Experimental simulation results show that using the PSO algorithm to solve the obstacle-free problem and then using the A* algorithm to correct the path obstructed by obstacles in the obstacle scenario, the agricultural UAV crop spraying trajectory planning based on the PSO-A* algorithm is real and effective. This research can provide theoretical basis for agricultural plant protection and solve the premise of autonomous operation of UAVs.

Abstract in Chinese

目前，无人机已逐步应用于农业领域，并广泛应用于精密播种、农药喷洒、植被探测等各类农业空中作业。使用农业无人机喷洒农药已成为农业植保过程中的一项重要任务。然而，在农业无人机的作物喷洒过程中，需要穿越多个喷洒点并规划避障路径，这大大影响了农业无人机作物喷洒作业的效率。针对上述问题，传统的粒子群优化算法具有较强的求解能力，但容易陷入局部最优。因此，本研究提出了一种与A*算法相结合的改进PSO算法，在传统PSO算法中引入了一种具有全局搜索和局部开发能力的非线性收敛因子平衡算法，并采用种群初始化来增强种群多样性，使改进的PSO算法具有更强的模型求解能力。本研究设计了两种农业无人机作物喷洒路径规划场景：一种是无障碍场景，另一种是有障碍场景。实验仿真结果表明，在障碍场景中，使用PSO算法解决无障碍问题，然后使用A*算法校正障碍物遮挡的路径，基于PSO-A*算法的农业无人机作物喷洒轨迹规划是真实有效的。该研究可以为农业植物保护提供理论依据，解决无人机自主运行的前提。