Direction estimation of coherent signals using spatial signature

A computationally efficient spatial signature-based (SS) method is proposed for estimating the directions of arrival of coherent narrowband signals impinging on a uniform linear array. The normalized SS of the coherent signals is blindly estimated from the principal eigenvector of array covariance matrix and then is used to estimate the directions with a modified Kumaresan-Prony method, where a linear prediction model is combined with "sub-array" averaging. The proposed method not only has the maximum permissible array aperture and computational simplicity, it also better resolves closely spaced coherent signals with a small length of data and at a lower signal-to-noise ratio.