Abstract

Wireless communication has undergone a significant growth to meet the unexpected demand of wireless data traffic over the past two decades. As manifested by the revolution of the third and fourth generations and long-term evolution advanced (LTE-A), engineers and researchers have been devoted to the development of the next-generation (5G) wireless solutions to meet the anticipated demand of 2020. To this end, cooperative relay communication has been introduced as an enabling technology to increase the throughput and extend the coverage of the broadband wireless networks. Decode-and-forward (DF) has been known as an effective cooperative relaying strategy for its outstanding features. On the other hand, merging massive multi-input-multi-output (MIMO) with cooperative DF relay is considered as a key technology for 5G wireless networks to improve the quality-of-service (QoS) in a cost-effective manner. The objective of this thesis is to establish and solve a power allocation optimization problem for energy efficient multi-pair DF relay systems integrated with massive MIMO.

The first part of the thesis is focused on a constrained optimization problem to minimize the total transmit power for each transmission phase of the DF relay. Due to the non-convexity characteristic, the objective function is approximated as a convex function by means of complementary geometric programming (CGP) which is then solved by a sequence of geometric programming (GP). A lower bound of average SINR is also introduced by adopting the MMSE channel state information (CSI) to relax the constraint functions in the standard GP form. Finally, we proposed a homotopy or continuation method based algorithm to solve the optimization problem via popular CVX optimization toolbox. MATLAB simulations are conducted to validate the proposed algorithm.

In the second part, another optimization problem is presented for the entire two-hop transmission of the DF relay to improve the global energy efficiency (GEE) under different channel conditions. Here, we estimate the channel by maximum likelihood (ML) criterion and investigate a closed-form expression of GEE. Further, GEE is approximated in a convex form by applying CGP due to the difficulty arising from the non-convexity and a lower bound of the average SINR expression is also derived to relax the constraint functions in the GP problem. Numerical results showing a detailed comparison of GEE under ML and MMSE channel estimation conditions and the performance improvement from the proposed algorithm are provided.