DOA estimation of harmonic acoustic signals

In this paper, we propose a novel method to estimate the direction of arrival (DOA) of the harmonic acoustic signals received by a uniform linear sensor array. The wideband Direction of Arrival (DOA) estimation problem is formulated under Bayesian framework. The received signals are firstly divided into narrow sub-bands, and the sparsity in the angular domain in each sub-band is exploited to give a coarse estimation of the DOA. The unknown information on the sub-bands occupied by the received signals is extracted by imposing a Dirichlet process prior on the latent parametric space. Then a 1-D searching procedure based the reconstructed results and the obtained information of sub-band frequency occupation is used to give a refined DOA estimation. Results of numerical simulations show that the proposed method outperforms the state-of-the-art algorithms for DOA estimation of harmonic acoustic signals.