Jiamin Chu

Jiamin Chu

Auto-generated profile, based on publicly available information

Jiamin Chu

Jiangsu Engineering Research Center of Novel Optical Fiber Technology and Communication Network, Suzhou Key Laboratory of Advanced Optical Communication Network Technology, the School of Electronic and Information Engineering, Soochow University, Suzhou, Jiangsu, China

Research Focus & Metrics

Academic Career Length

6Years (2021-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
17 November 2021 in Journal of Lightwave Technology
Verified
Journal Article
Closed access
JCJiamin Chu
Jiamin Chu
MGMingyi Gao
Mingyi Gao
XLXiaoli Liu
Xiaoli Liu
MBMeihua Bi
Meihua Bi
HHHe Huang
He Huang
GSGangxiang Shen
Gangxiang Shen
Filter bank multicarrier with offset quadrature amplitude modulation (FBMC/OQAM) is a promising candidate for 5G mobile fronthaul system. Due to the inherent imaginary interference of FBMC/OQAM signals, an accurate channel estimation is particularly indispensable. In this paper, an efficient method is proposed based on complex-valued neural networks (CVNN) for an accurate channel estimation of FBMC/OQAM signals with low computational complexity and pilot overhead. Here, the channel frequency response (CFR) is first calculated for training the CVNN. Next, the CFR estimated from the extracted pilots is exploited as the input of the trained CVNN to yield the CFR of data symbols. We experimentally demonstrate a 12.5-GBd intensity modulation direct detection (IM/DD) FBMC/OQAM transmission system over 30-km and 50-km standard single mode fibers (SSMF). The experimental results show that the proposed method achieves 3-dB and 1-dB receiver sensitivity improvement at the bit error ratio (BER) of 3.8×10 −3 with only 5% pilot overhead respectively, compared to the conventional least square (LS) and linear minimum mean error (LMMSE) methods. When the pilot overhead decreases from 10% to 1%, its BER performance is always better than LS and LMMSE. For the computational complexity, its complexity is the same order of magnitude as LS and lower than LMMSE. On the other hand, the CVNN requires less hidden neurons to keep the similar BER performance as real-valued neural networks (RVNN). Meanwhile, it has excellent resilience to fiber chromatic dispersion over 30-km and 50-km SSMF.