RESEARCH PHYSICS-BASED RAINFALL ENVIRONMENT EMULATION FOR RADIO WAVE ATTENUATION STUDIES USING MULTI-PARAMETER TIME SERIES DATA
Keywords:
Rain attenuation; radio propagation; rainfall emulation; time-series data; physics-informed modelling; millimetre-wave communication.Abstract
Background: At microwave and millimeter-wave frequencies, rain affects radio signals with the effects of absorption, scattering and depolarization. The traditional models of attenuation typically assume that the rain rate is the dominant factor, and fail to fully capture the effect of drop size, wind, the atmosphere and surface wetting on the attenuation phenomenon.
Objective: This study aimed to create and validate an emulator that simulates the rain conditions from the physics-based rainfall data in the form of multiple time series components to study radio-wave attenuation.
Experimental Design: Quantitative experimental design. The controlled rainfall rates were created ranging from 5 to 150 mm h−1^{-1} for radio links ranging from 10 to 60 GHz. The following data was synchronously recorded: Rain rate, drop size distribution, wind speed, temperature, humidity, received power, signal to noise ratio, bit error rate and attenuation. Various models, including statistical models, time-series models, conventional physical models, machine-learning models and a hybrid model combining conventional physics and machine learning were tested.
Results: The uniformity of rainfall and the ability to replicate the programmed rainfall was produced from the emulator. The relationship between the rainfall intensity, frequency, path length and the diameter of the rain drops with the attenuation was found to be non-linear. When the rainfall was heavy it caused the RPR to decrease, SNR to decrease and BER to increase and packet loss to increase. The hybrid model performed best in predicting and most closely simulating attenuation peaks, fade durations and the time it takes for the recovery after rain compared to comparison models.
Conclusions: Combining the physics of rain with multi-parameter temporal modelling was a repeatable way to assess the performance of rain-affected satellite, 5G and 6G links.
