Distributed mobility control for fault-tolerant mobile networks

Large-scale networks of mobile agents autonomously performing long-term sensing and communication tasks will be foreseeable in the near future. Since nodes in such a system communicate in a multi-hop fashion and the environment may be constantly changing due to the mobility, it is important for such a system to be able to tolerate temporary node failures. Moreover it would be ideal if the communications between nodes are not interrupted while nodes are moving around. In this paper we present a provably correct distributed rendezvous algorithm to achieve both goals by making the underlying communication graph biconnected. The proposed algorithm only requires relative position information of each node and works both synchronously and asynchronously. We further propose for each node a sleep/awake power control schedule in conjunction with the node's mobility strategy to compensate the energy consumption due to the node's movement.