Hybrid Precoding via Successive Refinement for Millimeter Wave MIMO Communication Systems

We propose a hybrid precoding design for spatial multiplexing in millimeter wave massive multiple-input multiple-output systems. The proposed design relies on the popular orthogonal matching pursuit (OMP) algorithm. The OMP algorithm uses necessary columns of the right/left (for receiver/transmitter) unitary matrix of the channel (obtained via a singular value decomposition) to obtain the initial baseband (digital) precoder using principles of basis pursuit. The resulting baseband precoder is then used to refine the RF (analog) precoder. This iterative refinement improves the quality of hybrid preocders. Quantization of the RF and baseband precoding matrices is also discussed for limited feedback systems, where the high-dimensional RF precoder is quantized using codebooks generated using the $K-$ means clustering algorithm. On the other hand, the low-dimensional baseband precoder is quantized using well-known Grassmannian codebooks. Simulation results show that the proposed scheme performs close to the fully digital precoding scheme.