Detecting topological change using a wireless sensor network

Abstract

Dynamic geographic phenomena, such as forest fires and oil spills, can have dire environmental, sociopolitical, and economic consequences. Mitigating, if not preventing such events requires the use of advanced spatio-temporal information systems. One such system that has gained widespread interest is the wireless sensor network (WSN), a deployment of sensor nodes - tiny untethered computing devices, which run on batteries and are equipped with one or more commercial off-the-shelf or custom-made sensors and a radio transceiver. This research deals with initial attempts to detect topological changes to geographic phenomena by an environmentally deployed wireless sensor network (WSN). After providing the mathematical and technical preliminaries, we define topological change and present in-network algorithms to detect such changes and also, to manage the WSN's resources efficiently. The algorithms are compared against a resource-heavy continuous monitoring approach via simulation. The results show that two topological changes, hole loss and hole formation, can be correctly detected in-network and that energy is greatly saved by our event-driven approach. In future work, we hope to test the algorithms over a broader range of topological changes and to relax some of the network assumptions. © 2008 Springer-Verlag Berlin Heidelberg.