Sporadic pairwise synchronization for decentralized decision-making in mobile agent teams

This paper focuses on the problem of cooperative search and task response in a heterogeneous team of autonomous mobile agents with limited communication. The agents are engaged in a mission to search for targets, to verify their existence, and to take appropriate action for each verified target in an uncertain environment. Each agent makes its decisions based on a subjective information base (SIB), constructed through observation and communication with other agents. Communication between agents is of two types: Type I) Each agent periodically broadcasts recently obtained or generated information over a limited radius; and Type II) When two agents come sufficiently close, they exchange and merge their entire SIBs. Type I communication results in the diffusion of information among all the agents, and is, in principle, sufficient for decision-making. However, for a fixed population of agents with specified dynamics, this places a lower limit on the range and frequency of broadcast. We hypothesize that including a small amount of Type II communication allows for significant decrease in these limits, and leads to a more efficient system. This hybrid approach represents a general communication strategy for multi-agent systems.