Real-Valued DOA Estimation Utilizing Enhanced Covariance Matrix With Unknown Mutual Coupling
Ye Tian, Ran Wang, Hua Chen, Yunbai Qin, Ming Jin
Abstract
Due to the space limitations, the array in a massive multiple-input multiple-output (MIMO) system often suffers from unknown mutual coupling. Meanwhile, small records of data observations may coexist. Such two limitations bring a challenge for accurate direction-of-arrival (DOA) estimation. To conquer this challenge, a real-valued DOA estimation method is proposed in this letter, whose core is to eliminate the influence of unknown mutual coupling by the inherent mechanism, as well as enhance the sampled covariance matrix estimation with the linear shrinkage technique combined with Rao-Blackwell Ledoit-Wolf (RBLW) estimator under the case of small sample size. Considering the result that the direct usage of the shrinkage target of RBLW estimator can yield an improved DOA estimation under low SNRs, a modified method depends on the eigenvalue comparison is also addressed. Simulation results show that the proposed method can provided an increased accuracy with reduced complexity.