DISTANCE-BASED COORDINATION FOR A DECENTRALIZED UNMANNED SURFACE VEHICLE SWARM CONTROLLER WITH VARYING SWARM SIZE

Abstract

This thesis seeks to explore decentralized, vision-based control algorithms for a team of unmanned, autonomous vehicles equipped with limited field-of-view cameras when a change in team size is required. In the battlefield, there is potential for the communication links between unmanned vehicles in a team to be disrupted by enemy capabilities. Furthermore, vehicles are bound to receive battle damage or malfunction, and additional vehicles may be needed for replacement or to augment a team's capabilities. A change in team size may cause a number of operational issues for the remaining vehicles, such as the inability to track targets, to maintain formation, and to avoid collision. This thesis offers a modification to a previously developed decentralized, vision-based collision avoidance control algorithm by transitioning the algorithm to one in which vehicles maintain a separation distance from one another rather than keeping a desired phase shift to avoid collision. The performance of both the baseline and the modified control algorithm is then analyzed in simulations where the group size changes with respect to time.