Alternating Minimization Algorithm for Hybrid Precoding in Millimeter Wave MIMO Systems

Millimeter-Wave (mm-Wave) communication has been considered as the leading enabling technology for 5G networks as it provides greater spectrum other than existing cellular bands. It holds a promise to provide an unparalleled 5G cellular network capability increase. Because of the short wavelength of mm-Wave signals, Multiple-Input-Multiple-Output (MIMO) systems through precoding can exploit the use of large-scale antennas to battle rain attenuation and path loss. In contrast to the conventional MIMO systems, mm-Wave MIMO cannot perform precoding entirely at the baseband using digital precoders, as only a limited number of signal mixers and analog-to-digital converters can be supported considering their power consumption. As a cost-effective alternative, a significant attention has recently been given to a hybrid precoding transceiver architecture that combines a digital precoder with an analog precoder. This thesis discusses an Alternating Minimization (AltMin) algorithm focused on manifold optimization for the design of a hybrid precoder, consequently comparing its results to the fully digital precoder design, that nonetheless, has a high complexity. Algorithms of AltMin are further applied to the broadband settings with orthogonal frequency division multiplexing (OFDM) modulation. Simulation results demonstrate that in terms of spectral efficiency, AltMin algorithm can significantly outperform existing algorithms and, more importantly, in some cases it can achieve an optimum performance. Keywords: 5G, alternating minimization, hybrid precoding, low-complexity, manifold optimization, and millimeter-wave