Hierarchical dirichlet process based gamma mixture modelling for terahertz band wireless communication channels and statistical modelling of 240 GHz - 300 GHz band

Abstract

Terahertz (THz) frequencies span from 0.1 THz to 10 THz, and because of their large bandwidth opportunity, they appear to be one of the strong candidates to satisfy future wireless communication standards and data rate requirements. Another advantage of the THz band is that it has not yet been standardized and is not allocated to any wireless communication application. THz frequencies enable the use of nano-sized antennas due to their short wavelengths. This particular feature is predicted to result in new communication application areas at the micro-scale over short distances. Furthermore, because these antenna sizes will also allow the deployment of large antenna arrays in a limited space, it is predicted to provide numerous new communication application areas at the macro-scale for long distances. The successful utilization of THz frequencies and the implementation of communication at these frequencies are also dependent on the design of efficient, cutting-edge hardware and antennas. To properly build this equipment, the channel structure of the THz band must be thoroughly researched. This can only be accomplished by collecting extensive measurements and analyzing those measurements. Unfortunately, in addition to their benefits, THz frequencies have a downside in the form of substantial propagation losses. These losses constrain the communication link and make channel characterization difficult. Molecular absorption is a major contributor to propagation loss at THz frequencies. When a radio frequency (RF) signal of a certain wavelength interacts with gaseous substances while propagating in the atmosphere, part of their energy is absorbed by these gas molecules, leading the molecule to move to a higher energy level. As a result, when THz waves interact with molecules, various power losses occur in different windows of the THz spectrum. Furthermore, the effect of molecular absorption is proportional to communication distance. Due to the unique channel characteristics of the THz band, extended measurement campaigns and comprehensive propagation channel modeling are essential to understand the spectrum and to develop reliable communication systems in these bands. These measurements and channel modeling should be performed in a variety of scenarios based on various macro and micro-scale communication applications. When the channel modeling approaches are considered, the most commonly utilized channel modeling techniques can be categorized into two, deterministic and statistical. Ray tracing and traditional statistical modeling are insufficient to construct a suitable channel model due to the wide bandwidth and rapid changes in the characteristics of THz channels. Statistical channel models can represent a wide range of settings and can adopt the variable structure of the THz channel. In contrast, deterministic models, such as ray tracing can only represent a single environment in a specific condition. A