Point to-Point and Path Following Navigation of Mobile Robot in Farm Settings

A fully integrated navigation strategy of a wheeled mobile robot in farm settings and off-road terrains is described here. The proposed strategy is composed of four main actions which are: sensor data analysis, obstacle detection, obstacle avoidance, and goal seeking. Using these actions, the navigation approach is capable of autonomous row-detection, row-following and path planning motion in outdoor settings such as farms. In order to drive the robot in off-road terrain, it must detect holes or ground depressions (negative obstacles), that are inherent parts of these environments, in real-time at a safe distance from the robot. Key originalities of the proposed approach are its capability to accurately detect both positive (over ground) and negative obstacles, and accurately identify the end of the rows of trees/bushes in farm/orchard and enter the next row. Experimental evaluations were carried out using a differential wheeled mobile robot in different farm settings. The mobile robot, used for experiments, utilizes a tilting unit which carries a laser range finder to detect objects in the environment, and a RTK-DGPS unit for localization. The experiments demonstrate that the proposed technique is capable of successfully detecting and following rows (path following) as well as robust navigation of the robot for point-to-point motion.