Scheduling for Scalable Energy-Efficient Localization in Mobile Ad Hoc Networks

Existing localization schemes in wireless ad hoc networks rely on redundant measurements from multiple devices with known positions in order to reduce error. However, when node density is high this can result in excessive localization messages with minimal improvement on position accuracy. In this work we present a scheduling algorithm to select a subset of active reference nodes to be used in localization, which has the effect of reducing message overhead, increasing network lifetime, and improving localization accuracy in dense mobile networks. We investigate the Cramer-Rao Lower Bound (CRLB) and existing single-hop localization techniques to determine the optimal average node density to ensure sufficient estimation accuracy. The correctness and effectiveness of the proposed scheme is evaluated through extensive simulation results, which show that in dense networks localization messages are greatly reduced and network lifetimes are more than doubled, while maintaining high estimation accuracy. Furthermore, computational time of localization algorithms is reduced, which effectively decreases accumulated error due to computation latency when locating a mobile device.