A framework for motion coordination of small teams of agricultural robots

Abstract

This paper presents a control framework for coordinating teams of tractor-robots operating in the same field. The framework supports master-slave and peer-to-peer modes of operation. It is based on a distributed control approach, where each robot is equipped with its own nonlinear model predictive tracking controller. Each controller minimises the tracking error along a finite horizon and thus provides accurate tracking. When necessary, it also avoids collisions with nearby vehicles by altering the path velocity profile or the path’s geometry. To do this, it uses information about the motion trajectories of all other vehicles which may interfere with its own projected motion. The communication requirements of this approach along with the computational complexity of model predictive control limit the application of this framework to small robot teams consisting of a few vehicles. Simulation experiments verified that the proposed distributed control scheme can be used for accurate, coordinated path tracking. Further research is required concerning its robustness in the presence of disturbances and its stability in the presence of communication delays.