A game theoretic approach to sensor relocation in mobile sensor networks

A mobile sensor network (MSN) is composed of a distributed collection of nodes, each of which has sensing, computation, communication and locomotion capabilities. The locomotion capability of nodes makes mobile sensor networks able to complete various tasks, such as surveillancing over some mission space. In this paper we consider a class of MSN applications which have a limited number of sensors randomly deployed, in a task area, with different coverage values at different positions. The sensors' task is to relocate themselves from initial positions to optimal positions so as to maximize the total values. Towards this task, a distributed relocation protocol, based on a potential game approach, is proposed. Each sensor moves to a better sensor position constrained by the locomotion energy consumption in response to other sensors locomotion. The dynamics will finally converge to an equilibrium point, where no better position can be obtained by deviating from this point, within finite steps. Simulation results show that a good coverage performance can be obtained using our proposed distributed protocol.