Spectral Efficiency and Energy-efficiency Trade-off in Spectrum Sharing and RIS-based Wireless Systems

Abstract

Spectral efficiency (SE) and energy-efficiency (EE) are the two key performance indicators for next-generation wireless systems. The SE specifies how much information can be communicated by the system per unit time and bandwidth. The EE measures the amount of information transmitted per unit energy consumption. For the efficient usage of frequency spectrum and energy, the SE and EE of a system need to be improved. However, there exists a fundamental trade-off between the SE and EE. Improving one key performance indicator may degrade the other. We focus on analyzing this tradeoff for an underlay spectrum sharing system and a reconfigurable intelligent surface (RIS)-aided system in this thesis. We first propose a novel EE-aware joint antenna selection and power adaptation (EE-ASPA) rule for the secondary user (SU) of the underlay spectrum sharing system, when it is subject to the stochastic interference-outage constraint and the peak transmit power constraint. We present a geometric representation for the optimal power in terms of the channel gains within the secondary system and between the secondary and primary systems, and analyze its properties. We study several insightful special cases of the rule and its behavior in various regimes of the system parameters. We present an iterative subgradient-based algorithm and a computationally simpler bound-based algorithm to determine the constants of the rule. We also propose the SE-optimal ASPA rule. While the SE increases as the transmit power increases, its EE decreases. On the other hand, the EE of EE-ASPA does not decrease as the transmit power increases, while its SE saturates. It achieves a markedly higher EE compared to the SE-optimal ASPA and other conventional ASPA rules from the literature.

Next, we study the SE-EE trade-off in a RIS-aided system with a base station (BS) and user equipment (UE) that work in the time-division duplex mode. We propose the SE-optimal training durations and transmit powers for pilot and data when BS and UE are energy-constrained. The proposed scheme accounts for the channel estimation errors. We first derive a novel lower bound for the SE, which captures the cumulative impact of the channel estimation errors in the RIS phase-shift configuration and then in the coherent data demodulation. We present two innovations to make the derivation tractable and insightful. First, we present a novel approximation for the effective downlink channel gain in the presence of estimation errors. Second, we prove that the central limit theorem applies to the effective downlink channel gain even in the presence of estimation errors and correlated cascaded channels due to closely-spaced RIS elements. We extend our study to the cascaded channel grouping scenario that uses fewer pilots. The proposed scheme achieves a larger SE compared to the equal power allocation and the on/off-based scheme. Cascaded channel grouping improves the SE further because of its lower training overhead. It also achieves a larger EE. As the energy budgets of the UE and the BS increase, the SE increases. However, the EE increases and then decreases.