A Distributed Localization Algorithm for Wireless Sensor Networks Based on Robust Statistic

This research presents a distributed and low-complexity algorithm for range-based localization in wireless sensor networks (WSNs), which is an extension of the Distributed Spatially Constrained Local (DSCL) algorithm. The proposed method uses robust functions suitable to avoid outliers derived from corrupted estimated distances among sensor nodes, where the estimated distances are generated under two well known range-estimation techniques. The accuracy of the proposed algorithm was tested using a set of 10 distributed WSNs with noisy distance measurements between sensor nodes, deployed arbitrarily on a 2-D area of 100 m x 100 m. Regarding to the DSCL algorithm, the proposed approach DSCL/Smooth-l_1 reduces, in average, around 57% the error on position estimates under the RSSI model. On the other hand, the scheme DSCL/Cauchy-Lorentzian greatly outperforms the original one in the presence of outliers in distance estimates.