Exploiting spatial sparsity for estimating channels of hybrid MIMO systems in millimeter wave communications

Abstract

Hybrid multiple input multiple output (MIMO) systems consist of an analog beamformer with large antenna arrays followed by a digital MIMO processor. Channel estimation for hybrid MIMO systems in millimeter wave (mm-wave) communications is challenging because of the large antenna array and the low signal-to-noise ratio (SNR) before beamforming. In this paper, we propose an open-loop channel estimator for mm-wave hybrid MIMO systems exploiting the sparse nature of mm-wave channels. A sparse signal recovery problem is formulated for channel estimation and solved by the orthogonal matching pursuit (OMP) based methods. A modification of the OMP algorithm, called the multi-grid (MG) OMP, is proposed. It is shown that the MG-OMP can significantly reduce the computational load of the OMP method. A process for designing the training beams is also developed. Specifically, given the analog training beams the baseband processor for beam training is designed. Simulation results demonstrate the advantage of the OMP based methods over the conventional least squares (LS) method and the efficiency of the MG-OMP over the original OMP.