An Improved One-bit OFDM Receiver Based on Model-Driven Deep Learning
Zan Lu, Wei Li, Youyun Xu · 2019
Accurate channel estimation and signal detection are very difficult for orthogonal frequency division multiplexing (OFDM) receiver under the limit of one-bit complex quantization which can greatly reduce power loss and systematic complexity. In this paper, we propose an improved one-bit OFDM receiver based on model-driven deep learning (DL). Different from the conventional one-bit receiver based on autoencoder architecture of DL, our proposed one-bit receiver consists of channel estimation module and signal detection module, and each of which is constructed by a deep neural network after using traditional communication method as initialization. Simulation results show that our scheme is superior to AE-OFDM based on autoencoder in view of bit error rate (BER) performance.