Model-driven Time-varying Channel Estimation Method for OFDM System Based on Neural Network

Channel estimation plays an important role in orthogonal frequency division multiplexing (OFDM) systems, which directly affects the performance of wireless communication systems in high-speed mobile scenarios. In this paper, we propose a model-driven channel estimation method for OFDM systems with doubly selective channels. Instead of completely abandoning traditional communication modules, we first obtain rough channel estimation using the least square (LS) method and then use a neural network called RefineNet for accuracy improvement. The results show that our proposed method significantly outperforms the linear minimum mean squared error (LMMSE) method.