Directions of Arrival estimation of correlated sources in presence of Mutual coupling

Abstract

This paper proposes a self-calibration algorithm, that requires only real valued operations, to estimate the Directions of Arrival (DoAs) of correlated sources in the presence of unknown Mutual coupling. Based on the standard MUSIC algorithm and with respect to the same imposed constraints, we develop the Iterative Unitary-MUSIC algorithm that provides superior high resolution localisation capabilities even for correlated sources scenarios with a reduced computations cost as well as a low processing time compared with previous works. In our analysis, the induced Electro-Motive Force (EMF) method is used to model the mutual coupling matrix. Through computer simulations, we demonstrate that our approach has more interesting performances even if all coupling parameters are included and especially when the incoming signals are highly correlated. Moreover, we drive more flexible identifiability conditions for the uniqueness of the estimation solution that required a reduced number of sensors to estimate the same number of parameters compared with previous works.