Direction of Arrival Estimation by Using Artificial Neural Networks

Abstract

In the literature there are many algorithms for direction of arrival estimation like MUSIC, ESPRIT, First order forward prediction, Capon etc. These algorithms have heavy calculation operations. This situation could cause lags in response time of the algorithm, and may pose an important disadvantage in real time applications. To overcome this problem, artificial neural network (ANN) could be used. The training stage of an ANN needs significant time and sources but after training, the estimation by using ANN is very fast. In this study, an ANN approach has been proposed for direction of arrival estimation in uniform linear array antenna. In training, the whole pseudo spectrum is scanned by 10 degree steps. In the simulation process, it is accepted that a uniform linear array consists of 5 isotropic antenna elements and there are 1 to 4 arrival signals. Tests of the trained ANN have been done for various directions of arrival angles, and satisfactory results have been obtained.