Robust Maximum Likelihood Acoustic Energy Based Source Localization in Correlated Noisy Sensing Environments

Abstract

Acoustic energy based localization with wireless sensor networks is an interesting solution to locate sources and targets. For simplicity, localization formulation based on the maximum likelihood (ML) approach considers that the source and noise samples are uncorrelated and represented by a Gaussian distribution. However, the acoustic background noise can severely affect the accuracy of the location estimation. This paper proposes an accurate error estimate in which the correlation of the received signals at each wireless sensor is represented by a Hurst exponent and modeled by a fractional Gaussian noise (fGn). The experimental results show that the proposed solution is more appropriate for the source localization estimation under real acoustic noises and even for highly non-stationary sources.