Estimation of signal steering vectors in uncalibrated arrays

This paper considers the problem of extracting the signal powers and steering vectors from the signal covariance matrix without the knowledge of array manifold. Under the assumption of narrowband and uncorrelated signals, it is shown that it is necessary and sufficient for the signal powers and steering vectors to satisfy two conditions, termed the signal subspace and orthogonality conditions, in order for them to match the signal covariance matrix asymptotically. Based upon these two conditions, an algorithm is derived to iteratively search for the signal powers and steering vectors which closely match the signal covariance matrix estimated from the observed data. A real-data example is presented to illustrate the robustness of the proposed algorithm.<<ETX>>