Identification analysis of radio propagation links for rainfall environment based on deep learning

Precipitation is a common natural phenomenon that plays an important role in climate regulation. At the same time, due to the spatial and temporal heterogeneity of precipitation, natural disasters such as floods and droughts often occur, which can seriously damage the economy and livelihood of humans. We use AI technology to study the attenuation caused by precipitation during the propagation of radio wave, trying to find the precipitation information from different attenuation, and then achieve the purpose of precipitation monitoring. The experimental results show that the neural network successfully learns the attenuation characteristics using the communication base station as the signal source and the standard meteorological precipitation information at the time of data collection as the label, and the experimental accuracy is around 95%. This suggests that the use of radio links combined with AI for rainfall monitoring is a novel approach to rainfall measurement and has some research implications.