Neural network promotes the transmission quality of remote health based on 5G technology

In this paper, to improve the wireless channel transmission quality in remote health using 5G, Bi-LSTM (Bi-directional Long Short Term Memory) neural network based channel estimation method is proposed to solve the problem of limited performance of traditional channel estimation methods in complex multipath channel environment. The wireless channel in this paper is processed in an end-to-end way, and compared with the traditional method which estimates CSI before recovering the signals we use deep learning tools to directly obtain the transmitted signal and hiding the CSI in the process. In order to solve the problem of channel distortion, the autoregressive process is used in the model of wireless communication channel. The WINNER II channel parameters are used to generate training data. Besides, we use the iterative training process of neural network to obtain the optimal solution of the channel autoregressive coefficient. From the simulation results, in the complex multipath channel environment, compared with the traditional channel estimation method (LS and MMSE) and the channel estimation method using DNN (Deep Neural Networks), the performance of BER in our network is better. And compared with the performance of NMSE with increasing number of pilots, our network has the best performance on 5G wireless signal transmission, which can consider to be used in remote health.