Path Loss Modeling and Environment Features Powered Prediction for Sub-THz Communication

Sub-Terahertz communication has broad application prospects for realizing ultra-broadband sixth generation (6G) system. One fundamental challenge when moving to new spectrum is to understand the science of radio propagation and propose an accurate and effective channel prediction method. In this paper, we first conduct extensive vector network analyzer-based radio propagation measurements at 140 GHz and 220 GHz in indoor hallway and lobby environments and at 280 GHz in an indoor laboratory environment. Omnidirectional and best directional path loss are modeled by empirical single-band and multi-band path loss models. Numerical results demonstrate that large-scale close-in model in this paper is simpler and more physically-based compared to floating-intercept model. In particular, a path loss prediction method based on environment features is proposed, which can predict path loss directly by utilizing random forest method, and the propagation environment are defined and extracted by scatterer features and related features of the transmitter and receiver. The performance of the proposed method is better than that of empirical path loss models. The measured results not only enrich the datasets of indoor sub-THz channel propagation, also can guide communication systems, network planning and deployment for 6G and beyond.