Hyperlocal weather parameter sensing with mmWave signals

Abstract

The evolution of mobile communication technologies to achieve higher throughputs has led to the use of higher frequency bands. 5G technologies are working on the mmWave spectrum, which are frequencies between 30 GHz and 300 GHz, and it is expected that 6G would use even higher frequencies. The wavelength of the signals in these bands are like those used in radars, giving the possibility to use the wave for other things be-sides transmitting information. Network sensing is one of the use cases that can be exploited from the mmWave. Signal loss under different weather conditions has been studied and modeled for over 20 years. Based on these models, this thesis develops a deep learning LSTM model that accurately detects precipitation from a mmWave backhaul link.