UC Santa Cruz

Massive Multiple-Input Multiple-Output (MaMIMO) antenna framework is one of the disruptive technologies that is shaping the current and future generations of wireless communication standards. The requirements of 5G and 6G wireless standards constitute significant improvements in spectral efficiency, throughput, and network densification compared to the previous generations of wireless standards. It would be impossible to attain such aggressive goals without leveraging the advantages of the MaMIMO architectures. However, the ramifications associated with MaMIMO architectures that comprise of a large number of antennas and other components in its radio frequency (RF) chains are decreased network energy efficiency (NEE) and increased hardware cost. At the same time, the 5G and the 6G standards also mandate improvements in the overall NEE by many orders of magnitude. As an example, the 5G standard necessitates a 100x improvement in the overall NEE compared to the 4G standard like LTE. Hence the design of the MaMIMO framework along with baseband algorithms for optimal resource utilization to optimize performance and power consumption is of paramount importance.

In this thesis, we focus on circumventing the challenges of power consumption (or energy efficiency) of the MaMIMO transceiver systems by (a) allocation of resources like ADC bit-resolution in each of the RF chains for varying channel conditions and (b) by identifying phase shifts of the reflecting elements associated with the reconfigurable intelligent surfaces (RIS) in the RIS-assisted MaMIMO systems to enable non-line-of-sight (NLOS) communication between the transmitter and receiver of interest under interference. Such MaMIMO frameworks are envisioned to be at the heart of the next-generation wireless backhaul links in both vehicular and cellular networks. The proposed resource allocation algorithms ensure optimal performance (energy efficiency, throughput, and MSE) of the system under power constraints. The MaMIMO components like hybrid precoder and combiner are also designed jointly with resource allocation. The resource allocation algorithms are designed to ensure reduced computational complexity! In addition, this thesis poses the problem of constrained resource allocation in MaMIMO as a class of constrained combinatorial problems and develops two information-theoretic algorithms, namely Information-assisted dynamic programming (IADP) and Information-directed branch-and-prune algorithm (IADP) to solve them. This thesis expounds on the mathematical framework developed that forms the basis of these algorithms and shows that the proposed algorithms guarantee near-optimal performance with huge computational savings. The proposed algorithms are used to solve resource allocation problems (a) and (b). Using simulations it is shown that the proposed algorithms outperform the state-of-the-art algorithms with significant computational savings! The proposed algorithms also find applications in solving large-sized problems in other domains like DNA sequencing, which is also examined briefly in this thesis.