Autonomy Algorithms for Stable Dynamic Linear Convoying of Autonomous Marine Vehicles

Linear convoying of multiple vehicles involves typically several vehicles in a follow-the-leader formation with prescribed separation distance between vehicles. Following vehicles are primarily concerned with following the exact path of the leader rather than maintaining a constant 180 degree relative bearing. This paper describes algorithms for the three primary phases of this type of mission, the initial order selection phase, and the initial convoying aggregation phase, and the stable state transiting phase. The algorithms described are built upon the open source MOOS-IvP marine autonomy project. They include algorithms for decentralized task allocation to address the initial ordering phase, and a new convoying behavior of the IvP Helm for implementing the initial aggregation and stable state convoying phases. The convoying behavior uses the multi-objective optimization structure of the IvP Helm to allow existing obstacle avoidance and COLREGS collision avoidance behaviors.