The resilience evaluation of unmanned autonomous swarm with informed agents under partial failure

Abstract

The resilience analysis of unmanned autonomous swarms under various disturbances provides valuable insights for the design of unmanned swarms and mission planning. Current studies usually focus on the complete failure of the individual in the swarm where the individual and/or its connections are removed and/or rewired. In this paper, we study the issue of swarm resilience with some agents suffering partial failure. Three motion models and their corresponding resilience processes are proposed respectively given three partial failure modes. Furthermore, an improved resilience evaluation metric considering the absolute time scale and minimum mission performance requirement is proposed. The collective behavior and resilience process under different partial failures and strategies are studied through simulations. The results reveal the swarm resilience decreases as the swarm size increases, and the swarm resilience is sensitive to degree and location failure strategy.