Research On Path Planning For Agricultural Robot In Complex Environment

With the development of modern agriculture,it has become a development trend to improve agricultural production efficiency with the help of intelligent agricultural equipment.Agricultural robot technology is the premise of intelligent agriculture,and the path planning problem is the core problem that needs to be solved in the robot field.During the walking process,the robot needs to plan a shortest path to avoid obstacles in the shortest time in a complicated working environment.This paper mainly studies the path planning of agricultural robots in different working environments and different working modes,including the existence of narrow aisle between points and the existence without narrow aisle in path planning,the detemination of the starting point and the ending point,the path planning between multiple points passing through multiple intermediate nodes and the robot collision problem that may occur when multiple robots work together.Aiming at the above problems,this paper proposes a corresponding solution strategy,and validates the effectiveness and superiority of the proposed method through experimental simulation.The main research contents are as follows:(1)An improved ant colony algorithm that can bypass narrow aisle and reduce computational complexity is proposed.At present,in the path planning simulation,due to the low accuracy of environmental modeling,the size of the robot and the rigid structure of the robot are neglected,causing the robot fail to pass the path planning in the actual environment In this paper,we first analyze the motion direction according to the position in the grid map to reduce unnecessary calculations,and then propose a method to judge the narrow aisle,bypass the narrow aisle and calculate a new weighted adjacency matrix.Finally,the optimal path is found by means of the Best-Worst Ant System(BWAS)algorithm.The improved ant colony algorithm can not only find an optimal path around the narrow aisle,but also has less computation and shorter time in the search process.(2)A parallel ant colony algorithm for solving multi-point path planning problems in facility greenhouses is proposed.In facility greenhouses,Agricultural robots often do not have a fixed work locations,and sometimes multiple locations are required between the start and end points,which can be attributed to TSP issues.Existing methods for solving TSP problems have trouble with approximate calculation,low precision,and neglect of security.In this paper,a new mathematical model on the combination of distance and security is set up,and at the same timea mathematical model for determining the number of ants is established by distance and complexity,Based on the above model,a parallel ant colony algorithm is proposed.When facing with multiple work places,the algorithm can accurately find the optimal walking sequence to avoid obstacles in a short time and complete the path planning.(3)A conflict resolution strategy based on multi-method fusion is proposed.In view of the current imperfect multi-robot conflict resolution strategy,in this paper,the possibleconflict types are summarized and analyzed,and then the corresponding strategies are proposed for different conflict types to achieve an efficient and comprehensive solution to the robot conflict problem.The most common rear-end problem between robots is studied in detail,and the relationship between safety distance and robot motion characteristics is proposed.(4)In order to realize the flexible use of the algorithm,the agricultural robot monitoring system software was written by means of MATLAB GUI.The user can select the environment,mode,and parameters as needed,and finally obtain the path planning result,the number of iterations,and the time spent.