An Error Propagation Aware Algorithm for Precise Cooperative Indoor Localization

Recently, node localization for sensor networks has attracted considerable attention for military application. Despite the recent proposals, the relationship between the channel behavior and the performance analysis of cooperative algorithms has not been addressed. The assumptions about the statistics of the ranging error used in the literature are either too general or overly optimistic. Additionally when sensors collaborate to localize each other, there is no attempt to characterize the error in the position of the nodes. This lack of error propagation-awareness can degrade the performance of an algorithm and create divergence problems. In this paper we first introduce detailed modeling of channel propagation in indoor environments in the form of novel empirical path loss (PL) and distance measurement error (DME) models developed from the results of UWB channel measurements. Then we integrate these models in developing error propagation aware (EPA) precise cooperative localization algorithm that tracks the extent of the position error in each sensor node and its overall effect on subsequent multi-laterations. Finally we compare the algorithm against the Cramer-Rao lower bound (CRLB)