Unitary cyclic ESPRIT-like direction finding

Two unitary formulations of cyclic direction-of-arrival (DOA) algorithms, which are appropriate for both cyclic ESPRIT algorithm and conjugate cyclic ESPRIT algorithm, are proposed. Firstly, two approaches achieve a substantial reduction of the computational complexity via real-valued decomposition. Secondly, unitary method I cannot separate highly correlated or coherent signals, but unitary method II can separate them. Finally, these unitary methods allow to select desired signals and to ignore interferences for uncorrelated source scenarios, but only unitary method II has signal selective property in multipath propagation. Compared with cyclic ESPRIT algorithm, simulation results show the effectiveness of the proposed methods.