Fast Reconstruction and Iterative Updating of Spatial Covariance Matrix for DOA Estimation in Hybrid Massive MIMO

Abstract

Due to the high resolution property, multiple signal classification (MUSIC) algorithm has been widely used for direction-of-arrival (DOA) estimation in wireless systems. To reduce the cost caused by radio frequency (RF) chains, hybrid structure has been adopted in massive MIMO systems operating at millimeter-wave bands. With hybrid structures, the received signals at the antennas are not fed directly to the receiver, and thus the spatial covariance matrix (SCM), which is essential to MUSIC algorithm, cannot be obtained using traditional sample average algorithm. Based on our previous works, we propose a fast beam sweeping algorithm in this article for SCM reconstruction in hybrid massive MIMO systems. Using multiple RF chains in hybrid structures, we find that beam sweeping can be conducted in a parallel manner so that the SCM can be reconstructed much faster than our previous works. To further accelerate the updating of the reconstructed SCM, an iterative updating procedure is also developed in this article. The iterative procedure can update the reconstructed SCM much faster than the batch-processing approach because it updates the SCM immediately as long as new beam sweeping results are available. Simulation results are also presented in this article to demonstrate the improvement of the proposed approach.