A scalable, robust, ultra-low complexity agent swarm for area coverage and interception tasks

Simulations of biologically inspired swarms where agents jointly achieve tasks using local rules rather than global centralized or distributed control have demonstrated the high performance of agent swarms on a variety of tasks (such as surveillance, plume tracking, or target interception). However, most swarm systems rely on the information exchange of agents with their neighbors, which in practical instantiations would involve digital communication. Moreover, many systems would require global positioning methods (e.g., GPS) to determine the exact location of agents in their environment. We propose a beacon-based principle for target-oriented navigation of large numbers of autonomous agents, which is radically different from previous methods in that it neither requires digital communication nor any kind of global position information for coordination of movements and interactions and, moreover, has only minimal "computing" requirements. Results from extensive simulations of the system in an area coverage and agent interception task show that (1) the system achieves perfect task performance (i.e., all hostile agents are intercepted), (2) scales (works with an arbitrary number of agents), and (3) is robust (adapts to changes in agent position and configuration)