Sequential estimation of the range and the bearing using the Zero-Forcing Music approach

We consider the range and bearing estimation of near-field narrow-band sources from noisy data observed across a passive sensor array. For some difficult scenarios as for correlated and largely spaced sources at low SNRs, or correlated and closely spaced sources, the Near FieLd (NFL) version of the MUltiple SIgnal Classification (MUSIC) algorithm is no longer reliable. In this paper, we adapt and extend the sequential Zero-Forcing MUSIC (ZF-MUSIC) algorithm, which avoids the delicate search of multiple maxima, to the NFL context. In order to compare the NFL ZF-MUSIC with the Second-Order Statistics Weighted Prediction (SOS-WP) algorithm, we assumed an uniform sampling in the spatial domain. However, the proposed algorithm can be exploited for general array geometries. Finally numerical simulations show that the variance of the proposed algorithm achieves the Cram é r-Rao Bound (CRB) in difficult scenarios and for sufficient Signal to Noise Ratio (SNR).