DETECTION AND ESTIMATION IN SENSOR ARRAYS USING WEIGHTED SUBSPACE FITTING

This paper addresses the problem of signal parameter estimation of narrow-band emitter signals impinging on an array of sensors. A multidimensional estimation procedure is proposed, which applies to arbitrary array structures and signal correlation. The method is based on the recently introduced weighted subspace fitting (WSF) criterion, and includes schemes for both detecting the number of sources and estimating the signal parameters. A Gauss-Newton type algorithm is suggested for minimizing the WSF criterion. A new detection scheme is also formulated based on the asymptotic distribution of the WSF cost function. Strong consistency of the detection algorithm is proved for arbitrary signal correlation, including coherence. The WSF detection method is compared to a recently proposed information theoretic approach, and found to provide a significant improvement for high signal correlation scenarios. Simulations are carried out comparing the proposed WSF technique to the deterministic maximum likelihood (ML) method. The WSF scheme is found to be limited only by the estimation accuracy and not by the initialization or detection. This does not appear to be true for the ML method.