Efficient Connectivity Maintenance For Distributed Robotic Systems

Abstract

Teams of multiple autonomous robots have the potential to improve upon many robotic tasks performed by individuals. As these robots move about, they must maintain connectivity through direct links or multi-hop paths in order to exchange useful information for the completion of collaborative tasks and avoid duplicating work. In this thesis, we survey the field to identify an optimally versatile connectivity maintenance algorithm supporting arbitrary tasks. To this end, we implement, modify, and optimize several connectivity maintenance algorithms from the literature and evaluate their performance on tasks such as trajectory following and multi-agent search. Finally we leverage these results to discuss the trade-offs between these algorithms for use cases with different priorities.