Study on Multiobjective Path Optimization of Plant Protection Robots Based on Improved ACO‐DWA Algorithm

• Published in: Engineering reports (Hoboken, N.J.) Vol. 7; no. 2
• Main Authors: Niu, Jing, Shen, Chuanyan, Zhang, Lipeng, Gao, Guanghao, Zheng, Jiahao
• Format: Journal Article
• Language: English
• Published: Hoboken, USA John Wiley & Sons, Inc 01.02.2025; Wiley
• Subjects: access cost; Agricultural equipment; Algorithms; Ant colony optimization; Clustering; Genetic algorithms; Grid method; improved ACO‐DWA; Kinematics; Lasers; Leaves; Lidar; Matlab; Mountainous areas; Multiple objective analysis; Navigation; Navigation systems; Obstacle avoidance; Optimization; Path planning; Pheromones; plant protection robots; point clouds clustering; Robot dynamics; Robots; Spraying; Trajectory optimization; Velocity
• ISSN: 2577-8196; 2577-8196
• DOI: 10.1002/eng2.70009

ABSTRACT The use of large‐scale agricultural machinery for plant protection in nonstandard orchards of mountainous areas is very limited, while small‐wheeled robots have great application prospect. To address the false judgment of visual information caused by the shading of branches and leaves and the difficulty in avoiding obstacles in complex orchard terrain, an operation trajectory optimization approach based on the improved dynamic window algorithm (DWA) with ant colony optimization (ACO) was developed. First, the orchard environment information was acquired by laser radar, and the voxel method was used to reduce the point cloud density of orchard ground. The grid method was used to segment the point clouds. Using the K‐means clustering algorithm to extract navigation lines of the tree row. Second, combining with the kinematic model and the motion path constraints of robots, a series of candidate trace points were generated using the model‐based prediction algorithm (SBMPO). Then, using the improved ACO‐DWA algorithm, the robot access cost was integrated into the objective function of the search node, and the path planning was carried out online based on the environment grid map. Finally, in simulation and orchard validation scenes, the effects of this improved approach were checked through the simulation platform Matlab R2021 and ROS2 operating system. Simulation results on Matlab R2021 show that this improved algorithm has an average of 28%, 16% and 37% reduction in three indicators of the travel path, the number of detours, and the calculation time, respectively. In the orchard real‐time experiment, compared with other excellent algorithms, the navigation distance error is reduced obviously. These experimental results show that this method would obviously improve the robot access effect in the inner area between the tree row, significantly meet the requirements of high safety and speed level of robots in these scenes. Also, it could be applied in the field such as picking and spraying robots. To solve the problems of visual information misjudgment caused by closed orchard branches and leaves, as well as the delayed obstacle avoidance of robots caused by complex working terrain, a wheeled plant protection robot operation trajectory optimization method, which is based on the improved dynamic window algorithm integrating ant colony algorithm (ACO‐DWA) algorithm is proposed. Combined with the kinematic model and job specification constraints of the plant protection robot, a series of candidate trajectory sets are generated using the model based prediction algorithm (SBMPO). Using the improved ACO‐DWA algorithm, the robot's travel cost is integrated into the objective function of the search node, and the path planning is carried out online based on the environmental map.