Maze mapping based modified depth first search algorithm simulator for agricultural environment

Abstract

Along with the rapid growth of technology, the increasing usage of Autonomous mobile robots (AMRs) in agriculture aims to improve productivity and work efficiency. However, navigating an AMR in an agriculture landscape is very challenging since it requires an AMR to identify its surrounding area and finds the path for maneuvering around the field. In this work, we considered an agricultural landscape as an unknown maze. The walls in a maze represent the surrounding plants. We proposed a modified depth-first search (DFS) algorithm for mapping the environment based on predefined moving rules and developed a mobile robot simulator using the Python turtle library to simulate the proposed approach. In the simulator, the robot scans the surrounding environment to determine the surrounding walls and empty track. Then it moves one at a time to a nearby cell based on the given rules. In case of an intersection, the robot saves the junction information before moving to the next cell. This way, when the robot arrives at a dead-end, it can backtrack to the nearest junction with an unvisited path. The simulation results show that the robot can map the maze and then back to its initial location.