An improved algorithm based on ESPRIT-Like for coherent signals

Combining the ESPRIT-Like algorithm with the propagator method, this paper presents an improved algorithm, which reconstructs the spatial covariance matrix into two Toeplitz patterns, which have translational features. Compared with the ESPRIT-Like algorithm, the direction of arrival (DOA) estimation using the improved algorithm cannot be affected by the coherency, and more importantly it does not require any eigendecomposition of the cross-spectral matrix (CSM), and hence the computation can be more reduced. Simulations are given to support our theoretical results.