Analysis of the combined effects of finite samples and model errors on array processing performance

The authors consider the performance of the class of signal subspace fitting algorithms for signal parameter estimation using narrowband sensor array data. The principle sources of estimation error in such applications are the finite sample effects of additive noise and imprecise models for the antenna array and spatial noise statistics. The covariance matrix of the estimation error when all of these error sources are present is found to be the sum of the individual contributions from each component separately. This simplifying fact allows for the derivation of an overall optimal subspace weighting for a particular array and noise covariance error model. In fact, the resulting algorithm yields the lowest possible asymptotic estimation error variance of any method for the model in question.<>