Team-triggered coordination of robotic networks for optimal deployment

This paper introduces a novel team-triggered algorithmic solution for a distributed optimal deployment problem involving a group of mobile sensors. Distributed self-triggered algorithms relieve the requirement of synchronous periodic communication among agents by providing opportunistic criteria for when communication should occur. However, these criteria are often conservative since worst-case scenarios must always be considered to ensure the monotonic evolution of a relevant objective function. Here we introduce a team-triggered algorithm that builds on the idea of `promises' among agents, allowing them to operate with better information about their neighbors when they are not communicating, over a dynamically changing graph. We analyze the correctness of the proposed strategy and establish the same convergence guarantees as a coordination algorithm that assumes perfect information at all times. The technical approach relies on tools from set-valued stability analysis, computational geometry, and event-based systems. Simulations illustrate our results.