Towards Articulating Failures and Fault Trees and Verification for Building Resilience in Robot Swarms

Robot Swarms are starting to enter real-life scenarios. These scenarios pose more complex events than the swarms face in lab settings. Chances of failure are higher in practical situations due to the increased number of potential failure cases. However, current swarm design methodologies do not explicitly consider fault analysis during the design phase of robot swarms. In this paper, we approach the problem of fault analysis in robot swarm mission design from a systems engineering perspective and provide a framework to identify a swarm’s potential failure modes while designing the robot swarms.