Task Assignment and Trajectory Planning Algorithm for a Class of Cooperative Agricultural Robots

Abstract

Agricultural field operations, such as harvesting for fruits and scouting for disease, are labor intensive and time consuming. With the recent push toward autonomous farming, a method to rapidly generate trajectories for a group of cooperative agricultural robots becomes necessary. The challenging aspect of solving this problem is to satisfy realistic constraints such as changing environments, actuation limitations, nonlinear heterogeneous dynamics, conflict resolution, and formation reconfigurations. In this paper, a hierarchical decision making and trajectory planning method is studied for a group of agricultural robots cooperatively conducting certain farming task such as citrus harvesting. Within the algorithm framework, there are two main parts (cooperative level and individual level): (1) in the cooperative level, once a discrete reconfiguration event is confirmed and replanning is triggered, all the possible formation configurations and associated robot locations for specific farming tasks will be evaluated and ranked according to the feasibility condition and the cooperative level performance index; and (2) in the individual level, a local pursuit (LP) strategy based cooperative trajectory planning algorithm is designed to generate local optimal cooperative trajectories for agricultural robots to achieve and maintain their desired operation formation in a decentralized manner. The capabilities of the proposed method are demonstrated in a citrus harvesting problem.