Channel Estimation And Signal Detection For OFDM System Based On Deep Learning

The quality of channel estimation determines the performance of the entire communication system to a great extent.It has always been a hot issue in wireless communication system research.Channel estimation can obtain the state information of the channel,and recover the original data by compensating the data at the receiving end,thereby improving the transmission performance of the entire communication system.With the explosive growth of wireless communication services,the channel environment is becoming more and more complex,and people have higher and higher requirements for communication systems.Under complex channel conditions,how to make channel estimation more accurate has become a new research hotspot,especially in the future-oriented intelligent communication,the combination of new and old technologies is undoubtedly a hotspot at home and abroad.As one of the core technologies of 4G,OFDM technology is widely used in the new generation of wireless communication systems.Therefore,it is very necessary to improve and innovate the channel estimation algorithm in the OFDM system.In recent years,deep learning has developed rapidly,and has been widely used in various fields,such as computer vision,image processing,and language recognition.Deep learning has initially been applied in the field of wireless communication,and it has been proved that the effect of this combination is considerable.In this paper,based on the DNN channel estimation method,an improved RNN LSTM network that is good at processing text information is used.When the system has fewer pilots,a larger number of OFDM symbols,the removal of cyclic prefix,the impact of clipping noise,and various superposition of negative conditions,the channel estimation performance is compared with the traditional LS and MMSE methods,and the effect of experimental parameters on the results is studied.Various parameters are also tuned.The simulation results show that the deep learning method has a good performance in channel estimation.When the channel is severely interfered,the deep learning method is robust,and compared with the DNN,the channel estimation method based on the LSTM network performs better.