A Fast Direct Source Localization Approach for Acoustic Sensor Array

Kai Yu,Kung Yao,Ralph E. Hudson,Charles Taylor,Zhi Wang
DOI: https://doi.org/10.1109/icassp.2016.7472270
2016-01-01
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.
What problem does this paper attempt to address?