Full Coverage Path Planning Methods of Harvesting Robot with Multi-Objective Constraints
Lihui Wang,Zhuoxuan Wang,Mingjie Liu,Zehua Ying,Ninghui Xu,Qian Meng
DOI: https://doi.org/10.1007/s10846-022-01722-0
2022-08-31
Journal of Intelligent and Robotic Systems: Theory and Applications
Abstract:Focusing on the low efficiency and weak universality in autonomous path planning of harvesting robots during the process of irregular farmland operations, a full coverage path planning method based on multi-objective constraints is proposed. Firstly, in sight of the influence of the shape of the farmland, the minimum turning radius of the harvesting robot, the distribution of the unloading position, and the manual operation experience on working path, a full coverage path planning model for the irregular quadrilateral farmland is established, which significantly improve the adaptability of path planning algorithm to working environment and performance of the harvesting robot. Secondly, according to the farmland boundary, selected operating direction, and estimated output of the operating row, an operating environment model is established. Then the corner turning method of the irregular quadrilateral farmland is researched, and the turning path by taking turning area as optimization target is designed. Finally, the shuttle method and the improved ant colony algorithm are used to design the full coverage path respectively, by considering the constrains of the harvesting robot's granary capacity, total travel distance, full-load travel distance and unloading position distribution as constrains. Experiment results show that the proposed algorithm can perform intelligent planning under the multi-objective constraint according to the principle of maximum load rate. When solving the traversal sequence of job rows, the algorithm is not affected by the traversal sequence, the full load rate is greater than 90%, and full coverage path planning is realized.