Near Field Source Separation Improvement Through Unconditional Maximum Likelihood Estimator

In this work, we are focused in the improvement of the near field source separation through the approach of the Unconditional Maximum Likelihood (UML) estimator. Four aspects are considered: separation sources, SNR variation, snapshots and multiple sources, in order to evaluate their influence in the capacity separation for the case of closely spaced sources. In this way, we can establish the minimum conditions for the sources separation. In addition, we investigated the effects of snapshots and the increasing number of sources in their spatial position estimation. Using Monte Carlo simulation, we obtained the Root Mean Square (RMS) error of the source's direction of arrival. For evaluation purposes we include also MUSIC simulations. Our results show that the UML estimator improves source's separation performance under low SNR and snapshot values as well as increasing number of sources.