Channel Estimation Based On FSRCNN Under Agricultural Communication Environments

The development of modern agriculture has gradually become large-scale and diversified in production forms.In order to accelerate the transformation from traditional agriculture to modern agriculture,intelligent and networked methods have become important means of agricultural production.Modern communication technology is one of the key technologies in intelligent agriculture.Transmission of agricultural data in agricultural environment by means of wireless communication can effectively improve production efficiency and convenience.Channel estimation technology is an important process in wireless communication,which estimates the characteristics of the channel through the state of the received signal to receive information more accurately.Because the agricultural environment has complex and diverse influence factors on wireless communication,the amplitude,phase and frequency of the received signal will change greatly,and the change rule of the received signal is quite different from the process of mobile communication.Therefore,it is more and more important to study and improve wireless communication in agricultural environment.In order to improve the communication quality in agricultural environment,this paper proposes a channel estimation scheme based on image super-resolution reconstruction technology for static and dynamic communication scenarios in agricultural environments.This scheme uses the deep learning method Fast Super Resolution Convolutional Neural Network(FSRCNN)algorithm to improve interpolation in traditional channel estimation.Channel State Information(CSI)is inputted in the form of a low resolution matrix,which replaces the interpolation operation and undergoes super-resolution reconstruction to output a high-resolution matrix,to achieve the goal of improving channel estimation accuracy.This paper built a wireless communication system based on USRP and Lab VIEW,collected real channel data in different agricultural environment,and established a agricultural environment channel information database.Finally,experiments and verifications were conducted using simulated and real datasets under different conditions.