In this follow-up to an ANTS 2004 paper we continue to investigate the problem of gathering a swarm of multiple robotic agents on the plane using very limited local sensing capabilities. In our previous work, we assumed that the agents cannot measure their distance to neighboring agents at all. In this paper, we consider a crude range-limited sensing capability that can only tell if neighboring agents are either near or far. We introduce two new variants of our previously proposed algorithm that utilize this capability. We prove the correctness of our algorithms, and show that the newly added capability can improve the performance of the algorithm significantly.
展开▼
