| Rainfall is an important process in the water cycle system,which is of great significance for water resource management and flood warning.Since the existing traditional rainfall observation equipment has problems such as high cost and insufficient sampling,the use of rainfall attenuation in radio communication technology for rainfall rate calculation and analysis becomes a relatively cost-effective method.In view of this,this study first analyzes the signal attenuation on the link,and separately studies the effects of two variables,wet antenna attenuation as well as rainfall attenuation,on the link signal.In the study of wet antenna attenuation,communication links under four different environments are designed,and the interference of wet antenna attenuation on the link is studied,and it is proposed that wrapping treatment of the antenna and then rainfall attenuation study can improve the recognition accuracy,and according to the different rainfall rate environment to collect the spectrum signal data within a certain time range,combined with the corresponding rainfall data,a rainfall rate calculation model based on rainfall attenuation is established,and the accuracy of the model is verified by experimental data.The research process is as follows.(1)This paper first studied and analyzed the overall influence of rainfall attenuation and wet antenna attenuation on the total amount of link attenuation on the link.It is found that the attenuation on the link affected by rainfall is mainly wet antenna attenuation and rainfall attenuation.In this paper,firstly,the relevant experiments were conducted for wet antenna attenuation,and the experimental platform based on GNU Radio transmitting software,Hack RF One transmitter and 3900 A monitoring receiver was built in this paper,and the signal spectrum signals under different antenna conditions were collected.(2)By analyzing the spectrum signals under four antenna conditions,the experiments were conducted using the deep learning based wet antenna attenuation identification analysis method for verifying the impact of wet antenna attenuation on the link signal and how to solve the wet antenna attenuation,the experimental results show that the deep learning based wet antenna attenuation identification analysis method can more accurately learn the attenuation brought by the wet antenna And in the subsequent rainfall attenuation research,the antenna can be wrapped so that the antenna can not directly contact the water vapor to avoid the inaccuracy of the rainfall attenuation research caused by the dynamic change of the wet antenna attenuation.(3)Data acquisition using a communication base station as the transmitting source,and relevant experiments using a deep learning-based rainfall attenuation identification and analysis method.The traditional classification neural network is used for feature extraction to learn the rainfall attenuation features under different rainfall rate environments and to distinguish different rainfall environments according to different attenuation values for the purpose of rainfall monitoring.This study conducts experiments for several rainfall events,and through the analysis of the experimental data,it is found that the traditional classification neural network is more sensitive to the extraction and learning of rainfall attenuation features,and the recognition accuracy of the collected spectrum signals under different rainfall environments can reach 89%.(4)In order to further improve the recognition accuracy,this paper establishes a rainfall rate calculation model based on improved classification neural network by using fully connected network,channel attention and spatial attention mechanisms on the basis of traditional classification neural network,and by comparing and analyzing the experimental data,it is found that the use of improved classification neural network is more sensitive to the extraction and learning of rainfall attenuation features,and the recognition accuracy of the spectrum signals collected under different rainfall The accuracy of the model was verified by comparing and analyzing the experimental data.The study shows that the rainfall rate calculation model based on rainfall attenuation can accurately estimate the rainfall intensity,and this method can be used to supplement the actual rainfall monitoring. |
