Abstract
A new computationally efficient subspace-based method without eigendecomposition (SUMWE) is investigated for the direction estimation of coherent narrowband signals impinging on a uniform linear array by exploiting the array geometry and its shift invariance property. In the proposed method, the computationally cumbersome eigendecomposition and the evaluation of all correlations of the array data are not required, and the effect of additive noise is eliminated. And the SUMWE can be extended to the spatially correlated noise by choosing appropriate subarrays. Further an adaptive implementation of the SUMWE is presented for tracking the directions of slowly moving signals. The performance of the presented algorithms is verified through numerical examples.
