Distributed nonlinear optimization-based control for multi-agent systems navigation in a coastal environment

This paper presents a distributed optimization-based control algorithm for the safe navigation of multi-surface vehicles in a complex environment with fixed and moving obstacles. We first generate off-line within the a priori known environment a collision-free path using RRT* (optimal Rapidly-exploring
Random Tree) algorithm. Next, potential field constructions
are developed through the use of on-off barrier functions
to ensure vehicle connectivity and fixed/moving obstacles
avoidance. These ingredients are integrated in a distributed
NMPC (Nonlinear Model Predictive Control) framework which activates the constraints only in the view range of the agents. The algorithm shows good results in simulations for both connectivity maintenance and validation of the COLREGS rules.