Global node selection for localization in a distributed sensor network

This work considers the problem of selecting the best nodes for localizing (in the mean squared (MS) position error sense) a target in a distributed wireless sensor network. Each node consists of an array of sensors that are able to estimate the direction of arrival (DOA) to a target. Different computationally efficient node selection approaches that use global network knowledge are introduced. Performance bounds based on the node/target geometry are derived, and these bounds help to determine the necessary communication reach of the active nodes. The resulting geolocation performance and energy usage, based on communication distance, is evaluated for a decentralized extended Kalman filter (EKF) that is exploiting the different selection approaches.