A Kernel Extreme Learning Machines Algorithm for Node Localization in Wireless Sensor Networks
Li Wang, Meng Joo Er, Zhang Shi
Abstract
Node localization is one of the promising research issues in Wireless Sensor Networks (WSNs). A novel node localization algorithm termed Kernel Extreme Learning Machines based on Hop-count Quantization (KELM-HQ) is proposed. The proposed algorithm employs the real number hop-counts between anchors and unknown nodes as the training inputs and the locations of the anchors as the training targets for KELM training. The proposed method also employs the real number hop-counts between unknown nodes as the test samples to compute the locations of unknown nodes by the trained KELM. Simulation results demonstrate that the proposed KELM-HQ algorithm improves the accuracy of node localization and it outperforms state-of-the-arts localization methods.