Blind closed-form DOA estimation for distributed sources

To distributed sources, the usage of conventional point-source model will cause inaccuracy in the direction-of-arrival (DOA) estimation. Conventional methods for distributed sources usually estimate the DOAs and the distribution parameters with multi-dimensional peak search operations, which will lead to much higher computational complexity than that in the point source model. In this paper, a blind parameter estimation method is proposed for distributed sources, where the DOAs are estimated in a closed form. As a result, the computational burden is decreased a lot. Furthermore, the proposed method can work in the scenario with mixture sources of many kinds of distributions. The selection method and its spatial-time domain version are also discussed in this paper. Simulation results show the effectiveness of our method.