Ensemble-Based Manifold Learning Methods for Localization in Wireless Sensor Networks

Abstract

Most of the present localization algorithms based on manifold learning in wireless sensor networks get the estimated sensor locations by using one neighborhood parameter. These algorithms are sensitive to the neighborhood parameter, and can not guarantee that the selected parameter of the neighborhood is optimal. To overcome this shortcoming, this paper proposes the robust localization method based on ensemble-based manifold learning in wireless sensor networks, and analyzes two ensemble-based methods. Experimental results show that this method not only improves the location accuracy, but also decreases the dependence on the neighborhood parameter.