Research Article
Deep Learning Network for Frequency Offset Cancellation in OFDM Communication System
@INPROCEEDINGS{10.1007/978-3-031-04409-0_1, author={Qingyang Guan and Shuang Wu}, title={Deep Learning Network for Frequency Offset Cancellation in OFDM Communication System}, proceedings={Machine Learning and Intelligent Communications. 6th EAI International Conference, MLICOM 2021, Virtual Event, November 2021, Proceedings}, proceedings_a={MLICOM}, year={2022}, month={5}, keywords={Deep learning network Carrier frequency offset BER}, doi={10.1007/978-3-031-04409-0_1} }- Qingyang Guan
Shuang Wu
Year: 2022
Deep Learning Network for Frequency Offset Cancellation in OFDM Communication System
MLICOM
Springer
DOI: 10.1007/978-3-031-04409-0_1
Abstract
A deep learning network for OFDM system is proposed to eliminate the CFO (carrier frequency offset) interference in OFDM system. The CFO greatly reduces the BER performance for the communication system. The frequency offset interference introduced needs to be eliminated before signal demodulation. Therefore, we propose the method to eliminate weights by establishing a deep learning network, and then form the optimization elimination weight matrix through iteration. Among them, the hidden layer and weights are trained and fine-tuned in the forward direction to cancel the interference introduced by CFO. Compared with MMSE and LS algorithm, the proposed deep learning network greatly improves the bit error rate performance. The simulation has proved that the proposed deep learning network algorithm has BER performance in OFDM systems.
