Premium
Impact of the Propagation Model on the Capacity in Small‐Cell Networks: Comparison Between the UHF/SHF and the Millimeter Wavebands
Author(s) -
Teixeira Emanuel,
Sousa Sofia,
Velez Fernando J.,
Peha Jon M.
Publication year - 2021
Publication title -
radio science
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.371
H-Index - 84
eISSN - 1944-799X
pISSN - 0048-6604
DOI - 10.1029/2020rs007150
Subject(s) - ultra high frequency , path loss , non line of sight propagation , extremely high frequency , radio spectrum , radio propagation , physics , millimeter , ranging , log distance path loss model , transmitter , radio propagation model , interference (communication) , radius , acoustics , telecommunications , computational physics , optics , computer science , wireless , computer network , channel (broadcasting)
This work shows how both frequency and the election of path loss model affect estimated spectral efficiency. Six different frequency bands are considered, ranging from 2.6 GHz in the ultra high frequency (UHF) band to 73 GHz in the millimeter wave bands (mmWaves), using both single‐slope and two‐slope path‐loss models. We start by comparing four urban path loss models for UHF: the urban/vehicular and pedestrian test environment from the ITU‐R M. 1255 Report, which includes the two‐slope urban micro line‐of‐sight (LoS) and NLoS, from the ITU‐R 2135 Report. Then, we consider mmWaves taking into consideration the modified Friis propagation model, followed by an analysis of the throughput for the 2.6, 3.5, 28, 38, 60, and 73 GHz frequency bands. We have found that the signal‐to‐interference‐plus‐noise ratio, as estimated with the more realistic two‐slope model, is lower for devices that are within the break‐point of the transmitter, which is a small distance in the UHF/SHF band. As a result, spectral efficiency is higher with mmWaves than with UHF/SHF spectrum when cell radius is under 40 m but not when cells are larger. Consequently, mmWaves spectrum will be more valuable as cells get small. We also find that capacity as estimated with the two‐slope model is considerably smaller than one would obtain with the one‐slope model when cells are small but there is little difference in the models when cells are larger. Thus, as cells get smaller, the use of one‐slope models may underestimate the number of cells that must be deployed.