Performance Characterization of Random Proximity Sensor Networks

In this paper, we characterize the localization performance and connectivity of sensors networks consisting of binary proximity sensors using a random sensor management strategy. The sensors are deployed uniformly at random over an area, and to limit the energy dissipation, each sensor node switches between an active and idle state according to random mechanisms regulated by a birth-and-death stochastic process. We first develop an upper bound for the minimum transmitting range which guarantees connectivity of the active nodes in the network with a desired probability. Then, we derive an analytical formula for predicting the mean-squared localization error of the active nodes when assuming a centroid localization scheme. Simulations are used to verify the theoretical claims for various localization schemes that operate only over connected active nodes.