A fast direct source localization approach for acoustic sensor array
Kai Yu, Kung Yao, Ralph E. Hudson, Charles E. Taylor, Zhi Wang
Abstract
We present a novel Fast Direct Source Localization (FDSL) approach for acoustic isotropic sensor array applications. Unlike previous approximate maximum likelihood (AML) approach, the proposed FDSL focuses on the phase shift caused by time delays among sensors and obtains an analytical result without exhaustive search for all the possible locations. Using this phase shift model, conventional phase uncertainty problem can be solved by a light-weight linear search in the limited phase space, which greatly reduces the computational complexity of array processing. Theoretical analysis of the FDSL has been proposed and compared with the Cramér Rao Bound (CRB) of the AML approach. Simulations using real bird call data validate the advantage of proposed method.
BibTeX
@inproceedings{icassp2016_afastdirectsourc,
title = {A fast direct source localization approach for acoustic sensor array},
author = {Kai Yu and Kung Yao and Ralph E. Hudson and Charles E. Taylor and Zhi Wang},
booktitle = {ICASSP 2016},
year = {2016}
}