Neural Network Based Analysis of Terahertz Frequency Signal Propagation for B5G/6G Wireless Networks

Abstract

Pico-cell based very high-speed wireless technologies require new medium access control mechanisms to provide top efficiency in the control plane. Beyond 5G and 6G wireless services are studied currently with synthetic data generated with special simulators. In this paper we used NS3 TeraSim tool to evaluate upload communication cases from mobile terminals to unique base station in different population and topology scenarios during 10 ms simulated time interval. Fractal based wavelet analysis is used to extract features of channel access in different simulation cases and classify them with recurrent neural network. The methodology utilized performs stable and unstable phases of the new IEEE pre-standard mechanism called Adaptive Directional Antenna Protocol for THz.