Complexity-reduced direction-of-arrival estimation method for highly correlated sources

The paper presents an approach based on a preprocessing technique that allows previously known methods of reducing the complexity of subspace-based methods to be adapted to the case of highly correlated signals. Snapshot vectors are first preprocessed, and then used to compute the signal subspace of the spatially smoothed correlation matrix. The proposed method is a complexity-reduced version of the DEESE algorithm (Grenier et al., 1993). It gives better resolution in noisy environments than the common spatial-smoothing method with a relatively small increase in computational cost