An analytical approach to the optimal deployment of wireless sensor networks

In this work we propose and investigate a novel research topic in Wireless Sensor Networks (WSNs): sensor deployment in order to maximize the interest of the gathered information. The target areas are characterized by zones with different interest levels. We model the interest variable as an "importance function" that assigns a quantitative reference to each point. Due to the nature of WSNs, the sensor deployment must guarantee that the information of all nodes can reach certain control nodes or sinks. Otherwise, the events captured by the nodes are lost. This condition is equivalent to ensuring the existence of a communication path from every node to at least one of the sinks. We propose a global optimization model that fulfills all the conditions and we solve it with a simulated annealing algorithm that determines optimal node placement. Moreover, we also characterize the effects of an aerial deployment, that is, the effect of node placement inaccuracy in network behavior and performance. We have found that small placement variations may lead to strong fluctuations in the quality of the connectivity properties of the network topology, and thus to a significant performance degradation in terms of captured interest. © 2008 Springer-Verlag Berlin Heidelberg.