For Full-Text PDF, please login, if you are a member of IEICE,|
or go to Pay Per View on menu list, if you are a nonmember of IEICE.
A Taylor Based Localization Algorithm for Wireless Sensor Network Using Extreme Learning Machine
Xiong LUO Xiaohui CHANG Hong LIU
IEICE TRANSACTIONS on Information and Systems
Publication Date: 2014/10/01
Online ISSN: 1745-1361
Type of Manuscript: Special Section PAPER (Special Section on Frontiers of Internet of Things)
wireless sensor network, time difference of arrival, extreme learning machine, non-line-of-sight,
Full Text: PDF(524KB)>>
More recently, there has been a growing interest in the study of wireless sensor network (WSN) technologies for Interest of Things (IoT). To improve the positioning accuracy of mobile station under the non-line-of-sight (NLOS) environment, a localization algorithm based on the single-hidden layer feedforward network (SLFN) using extreme learning machine (ELM) for WSN is proposed in this paper. Optimal reduction in the time difference of arrival (TDOA) measurement error is achieved using SLFN optimized by ELM. Compared with those traditional learning algorithms, ELM has its unique feature of a higher generalization capability at a much faster learning speed. After utilizing the ELM by randomly assigning the parameters of hidden nodes in the SLFN, the competitive performance can be obtained on the optimization task for TDOA measurement error. Then, based on that result, Taylor algorithm is implemented to deal with the position problem of mobile station. Experimental results show that the effect of NLOS propagation is reduced based on our proposed algorithm by introducing the ELM into Taylor algorithm. Moreover, in the simulation, the proposed approach, called Taylor-ELM, provides better performance compared with some traditional algorithms, such as least squares, Taylor, backpropagation neural network based Taylor, and Chan positioning methods.