Terahertz communications: Physical layer enablers and analysis

Abstract

Undoubtedly, spectrum scarcity constitutes the main bottleneck of current wireless networks. It is therefore imperative to move beyond the sub-6 GHz band in order to overcome this limitation. Toward this direction, terahertz (THz) communication is deemed a promising solution for future wireless systems owing to the abundant spectrum resources at these frequencies. Despite the prospect of terabit- per-second wireless links, THz signals suffer from severe propagation losses, which can undermine the communication range and performance of THz systems. In this dissertation, we tackle this challenge by putting forward two key physical layer technologies, namely massive multiple-input multiple-output (MIMO) and intelligent reflecting surfaces (IRSs). More particularly, this dissertation consists of two parts. In the first part, we thoroughly study the spatialwideband effect in THz massive MIMO. We commence by demonstrating that conventional narrowband beamforming/combining leads to substantial performance degradation for large antenna arrays and high transmission bandwidths. With this in mind, we propose a wideband array architecture based on true-timedelay and virtual subarrays. For the channel estimation problem, we introduce a wideband dictionary along with a novel variant of the orthogonal matching pursuit algorithm. Numerical simulations are provided showcasing that the proposed design enables: i) nearly squint-free beamforming/combining with a small number of true-time-delay elements; and ii) accurate channel acquisition with reduced pilot overhead even in the low signal-to-noise-ratio regime. In the second part, we focus on the fundamentals of IRSs at THz frequencies. Specifically, we show that an IRS has the potential to improve the energy efficiency of THz MIMO, when it is placed close to one of the link ends. As a result, electrically large IRSs are expected to operate in the radiating near-field zone, where the spherical wavefront of the emitted electromagnetic (EM) waves cannot be neglected. To this end, we introduce a spherical wave channel model by leveraging EM theory, which includes far-field as special case. Furthermore, we discuss the importance of using EM principles to characterize the path loss of IRS-aided links, as simplistic models may wrongly estimate the link budget and actual system performance. Our analysis reveals that: i) conventional far-field beamforming is highly suboptimal in terms of power gain, and hence beamfocusing is the optimal mode of operation for THz IRSs; and ii) frequencydependent beamfocusing is required in wideband THz transmissions, as beam squint can substantially reduce the achievable data rate.