Improving the performance of cyclic ESPRIT via real-valued decomposition

By exploiting the real-valued eigendecomposition (EVD) of the centro-Hermitian matrix, a novel cyclic ESPRIT algorithm with signal selective property is presented by introducing a new forward backward smoothed covariance matrix. Compared with Cyclic ESPRIT algorithm, the proposed approach has a better performance in the presence of multipath propagation. In addition, this approach not only reduces the computational complexity, but also allows to select desired signals and to ignore interferences by exploiting the cyclostationarity property of signals of interest (SOIs). Simulation results that illustrate the performance of this approach in conjunction with cyclic ESPRIT algorithm are described.