Sparse hybrid precoding and combining in millimeter wave MIMO systems

Abstract

Millimeter wave (mmWave) communication allows us to exploit a new spectrum band between 30 GHz to 300 GHz to meet the growing demands of capacity for fifth generation (5G) wireless communication systems. Multiple-input multiple-output (MIMO) antennas can be used to tackle higher path loss and attenuation at mmWave frequencies compared to microwave bands. Beamforming, called precoding at the transmitter, is performed digitally in conventional microwave frequency MIMO systems, but at mmWave frequencies the higher cost and power consumption of system components means that the system cannot implement one radio frequency (RF) chain per antenna. To enable spatial multiplexing, hybrid precoders using fewer RF chains than antennas emerge as cost-effective and power saving alternative for the transceiver architecture of mmWave MIMO systems. This paper demonstrates the hybrid precoder design with its spectral efficiency and energy efficiency characteristics, and we compare the performance with that of optimal digital precoding (with one RF chain per antenna) and simplified beam steering systems. It also includes two different algorithmic solutions to meet the optimization objective. The orthogonal matching pursuit (OMP) algorithm appears to provide high performance solution to the problem, whereas the gradient pursuit (GP) algorithm is proposed as a cost-effective and fast approximation solution that can still provide equally high performance.