Location Reconciliation in Distributed Sensor Localization

Unattended ground sensor networks have attracted considerable attention and research recently. Sensor self-localization is an important initial step in setting up a network for its intended operations. However, most of the research to date on sensor localization is based on centralized processing, requiring pair-wise communications of all sensors not only within the entire sensor network, but also with a central processing unit. In this paper, a recently proposed decentralized, or distributed, sensor self-localization scheme is revisited. This scheme requires only local communications among a few neighboring sensors, and requires no central processing unit. Such a scheme reduces power consumption and enhances the robustness of the sensor network. However, the sensor localization information among different sensor clusters need to be reconciled, a problem that is specific to this distributed localization approach. We present further solutions to the reconciliation problem with supporting computer simulation results