LAUSR

In recent years, interest in the use of mobile robots in the agricultural industry has increased, both to address labor shortages in rural areas and to increase food production in a more sustainable way. In order to have an efficient navigation system to cover long crop row fields, a path planner algorithm must consider maneuvering restrictions of the targeted robot. Most state-of-the-art works in agricultural navigation systems are intended for robots with a high degree of maneuverability that can typically make in-place turnings. This work aims to fill the gap in terms of the development of an efficient navigation system for car-like robots with limited turning radius in crop row fields. For this, we combine the global path planner A* and the local trajectory planner Timed Elastic Band (TEB). Additionally, we state the problem of finding an optimal path that covers the entire field as a Travelling Salesman Problem (TSP) that is based on the different turning maneuvers the robot can perform at field headlands. The solution of the TSP results in a time efficient coverage strategy that aligns with the robot's kinematics. Experiments performed in the Gazebo simulation environment show a reduction in field completion times of up to 20%, compared to trivial coverage paths. On the other hand, deviation of the robot with respect to the center of the field furrows was in all cases less than 10cm, which proves that the entire system operates with sufficient accuracy to avoid damaging the crops.