Robust Localization in Wireless Sensor Networks

Localization is essential for wireless sensor networks (WSNs). It is to determine the positions of sensor nodes based on incomplete mutual distance measurements. In this paper, to measure the accuracy of localization algorithms, a ranging error model for time of arrival (TOA) estimation is given, and the Cramer-Rao bound (CRB) for the model is derived. Then an algorithm is proposed to deal with the case where 1) ranging error accumulation exists, and 2) some anchor nodes broadcast inaccurate/wrong location information. Specifically, we first present a ranging error-tolerable topology reconstruction method without knowledge of anchor node locations. Then we propose a method to detect anchor nodes that have inaccurate/wrong position information. Simulations demonstrate the improvement of our algorithm compared to others.