Expectation-Maximization Based Direction of Arrival Estimation Under a Mixture of Noise

In this paper, we propose a novel scheme for direction of arrival estimation in the presence of a noise which is a combination of white Gaussian distributed noise and spherically invariant random distributed noise. Such combination arises in practical scenario, in which, the Gaussian component represents the thermal noise (i.e., the internal noise), whereas, the spherically invariant random process represents the possible presence of outliers and/or non-homogeneities of the environment (i.e., interference, clutter, jammers). The classical direction of arrival estimation using the maximum likelihood is computationally intractable. In order to overcome this drawback while maintaining a fine accuracy and taking into account the presence of the two noise components, we design an expectation-maximization algorithm. Finally, numerical simulations show that the proposed algorithm outperforms the state-of-the-art.