An accurate ensemble-based wireless localization strategy for wireless sensor networks

In wireless systems, localization is still an important challenge to ensure innovative based services solutions. In this paper, a novel localization algorithm which intends to improve robustness and accuracy of previous work based on regression tree is proposed. The suggested approach is a learning based ensemble technique which combines several regression trees. Anchor selection procedure is associated to the proposed algorithm to ensure better performance. We take into consideration two performance keys : the localization error and the computation complexity. Experimental results show that the ensemble method is simple and accurate compared to localization algorithms currently available in the literature.