Experimental verification of an ann-based model for 2D DOA estimation of closely spaced coherent sources.

ABSTRACT This article presents development and validation of an artificial neural network (ANN)-based model for accurate two-dimensional direction of arrival estimation of two coherent sources. The ANN model is developed using simulated signals and its performance is evaluated by measurements inside an anechoic chamber. Simulations and experiments show good agreement in estimating both azimuth and elevation angles of coherent sources. The proposed model demonstrates ability to separate two closely spaced sources and outperforms MUSIC with spatial smoothing preprocessing in the case of small number of signal snapshots. © 2014 Wiley Periodicals, Inc. Microwave Opt Technol Lett 56:2558-2562, 2014 [ABSTRACT FROM AUTHOR]