Performance analysis of propagator‐based ESPRIT for direction of arrival estimation

Compared with the standard estimation of signal parameters via rotational invariance technique (ESPRIT) for direction of arrival estimation, the propagator‐based ESPRIT is a more computationally efficient method since it requires no eigenvalue decomposition when calculating the signal subspace. In this study, the propagator‐based ESPRIT method is analysed by deriving its theoretical asymptotic error. With first‐order perturbation analysis, it is shown that both propagator‐based ESPRIT and standard ESPRIT have the same asymptotic error with high signal‐to‐noise ratio. Simulation results of both the propagator‐based ESPRIT and the standard ESPRIT are presented, and they are shown to be consistent with the theoretical derivations.