An improved anchor selection strategy for wireless localization of WSN nodes

Indoor localization methods based on the received signal strength indicator are widely used in the literature since no additional hardware is required for data acquisition. In this paper, a novel localization algorithm which combines both classification and regression methods is proposed to enhance the localization accuracy of previous methods based on regression tree. The proposed approach is based on the selection of the three anchors nearest to the target for the generation of the training set and during the testing phase. The performances are evaluated using real measurements acquired in office rooms. The experimental results show that the anchor selection procedure provides an increased accuracy if compared to the standard regression tree localization algorithm.