Nonparametric belief propagation based cooperative localization: A minimum spanning tree approach

Nonparametric belief propagation (NBP) algorithm can result in approximately optimal performance for probabilistic localization in wireless sensor networks without loops theoretically. However, in loopy networks the accuracy of NBP is doubtful and the computational complexity is high. In this paper, a novel approach running NBP on a minimum spanning tree (MST) is proposed, which mitigates the influence of loops and significantly reduces the computational cost as compared with the conventional NBP schemes. In addition, different from other spanning trees, the MST can confine more NBP particles into the bounding circle. Therefore, it shows better resistance to measurement errors. Numerical results show that the proposed method achieves better performance in terms of accuracy in highly connected networks, and the computational cost is much lower than the conventional NBP methods.