Date of Award

Fall 2020

Document Type

Open Access Dissertation

Department

Electrical Engineering

First Advisor

David Matolak

Abstract

A candidate millimeter-wave (mmWave) frequency band and modulation scheme that could fit to many present and future applications has been presented in this work. As is being explored by industry, we also suggest the 30 GHz band as a candidate carrier frequency and non-coherent frequency shift keying (NC-FSK) as a potential modulation scheme for future communication applications. The primary applications are aimed at 5th generation (5G) cellular type systems. Propagation measurements were conducted for outdoor and indoor environments using directional horn antennas for both co-polarized and cross-polarized antenna configurations to model the path loss for our candidate band. The measurements were conducted in typical line-of-sight (LOS) and non-LOS (NLOS) environments in a large building on the University of South Carolina campus, specifically at Swearingen Engineering Center. Several propagation path loss (PL) models are presented based upon this collected data. We can use these PL models in link budgets for estimating transmit power, antenna gains, receiver characteristics (e.g., noise figure), and link distances. The measurements also contribute to the body of knowledge on wireless channel propagation path loss for bands near 30 GHz.

Another measurement campaign was also conducted at the USC campus to measure a unique and complicated vegetation attenuation that may be considered a large challenge to mmWave systems. Radio wave attenuation and depolarization effects through several broadleaf evergreen shrubs at 31 GHz are reported, based upon measurements. To obtain a comparative reference for this mmWave attenuation, another measurement was also conducted at 5 GHz. From these measurements, we analyzed the proportional relationships between the attenuation and the shrub density (related to species), depth, and measurement geometry. Three different shrub species with different densities and depths, and for different measurement geometries, were employed. Results are in terms of measured specific attenuations at 31 GHz—the attenuation in dB/m. These will also be useful for link budget design, and outdoor and outdoor-indoor models for future mmWave communication.

For our 5G modulation scheme candidate, we evaluate its performance at 31 GHz via an empirical 3-D mmWave channel simulator: the NYUSIM channel model. As with all digital communication systems, performance is measured in terms of error ratios, and we evaluate the bit error rate (BER) performance of NC-FSK for different symbol rates over a variety of wireless mmWave channels. The NC-FSK scheme is known to be energy efficient for large alphabet size, and this is one of its virtues. Another is that since it is a form of FM, nonlinear amplification (far less costly than linear amplification) can be used. The performance evaluations enable us to present enhancements and trade-offs that can be done to improve the system performance by adjustment of the design parameters, i.e., modulation alphabet size and symbol rate, which together determine bandwidth (BW).

Share

COinS