Sensor Deployment and Relocation: A Unified Scheme

Sensor networks are envisioned to revolutionize our daily life byubiquitously monitoring our environment and/or adjusting it to suit our needs.Recent progress in robotics and low-power embedded systems has made itpossible to add mobility to small, light, low-cost sensors to be used inteams or swarms. Augmenting static sensor networks with mobile nodesaddresses many design challenges that exist in traditional staticsensor networks. This paper addresses the problem of topology control inmobile wireless networks. Limitations in communication, computation and energycapabilities push towards the adoption of distributed, energy-efficientsolutions to perform self-deployment and relocation of the nodes.We develop a unified, distributed algorithm that hasthe following features. During deployment, our algorithm yields aregular tessellation of the geographical area with a given nodedensity, called \em monitoring configuration. Upon theoccurrence of a physical phenomenon,network nodes relocate themselves so as to properly sample andcontrol the event, while maintaining the network connectivity.Then, as soon as the event ends, all nodes return to the monitoringconfiguration. To achieve these goals, we use a virtualforce-based strategy which proves to be very effective even whencompared to an optimal centralized solution. We assess the performanceof our approach in the presence of events with different shapes, and weinvestigate the transient behavior of our algorithm. This allows us toevaluate the effectiveness and the response time of the proposedsolution under various environmental conditions.