Real-Valued Sparse DOA Estimation for MIMO Array System Under Unknown Nonuniform Noise

In this paper, the problem of the direction of arrival (DOA) estimation for the multiple input multiple output (MIMO) array system is considered as a real-valued sparse signal recover procedure under the condition of unknown nonuniform noise. Then, a real-valued covariance vector-based sparse Bayesian learning framework is proposed, in which the reduced dimensional (RD) transformation is utilized to remove the redundant elements of MIMO array system, and a linear transformation is applied to eliminate the influence of unknown non-uniform noise. Then by supposing that the source powers follow an independent prior Gaussian distribution with zero-mean, a real-valued covariance vector-based sparse Bayesian model is formulated. And considering its unknown variance as hyper-parameters, they can be estimated by adopting the expectation-maximization algorithm. Finally, the DOA can be achieved according to the spatial spectrum of hyper-parameters. Simulation results have demonstrated that our proposed method not only achieves more superior performance but also provides robustness against nonuniform noise, compared with other recently reported sparse signal representation based methods.