An efficient method for angular parameter estimation of incoherently distributed sources via beamspace shift invariance

In this paper, we propose a beamspace-based method for nominal direction-of-arrival (DOA) and angular spread estimation of incoherently distributed (ID) sources using a uniform linear array (ULA). Firstly, with generalized array manifold of the ULA, we obtain the beamspace array manifold by performing beamspace transformation on the received vector of two overlapping subarrays, and further derive the beamspace shift invariance structure via designing appropriate beamforming matrix. Next, the total least squares approach is used to estimate the nominal DOAs of ID sources. Finally, with the DOA estimates, the corresponding angular spreads are obtained by means of the central moments of the angular distribution. The proposed method does not involve any spectral search and reduces the dimension of matrix operations, thus it is obviously more efficient than the traditional algorithms. Simulation results indicate that our proposed algorithm is comparable to the existing algorithms when the number of sensors is large.