论文信息 - Generalization Performance of Artificial Neural Network Based-Demodulator for Eigenvalue Modulated Signal

Generalization Performance of Artificial Neural Network Based-Demodulator for Eigenvalue Modulated Signal

The generalization performance of an artificial neural network-based demodulator for an eigenvalue modulated signal is investigated. Successful demodulation for transmission distances from zero to 3,000 km and carrier frequency offset within 1 GHz is demonstrated.

Ken Mishina | Akihiro Maruta | Yuki Yoshida | Daisuke Hisano

[1] Sergei K. Turitsyn,et al. Nonlinear Fourier Transform for Optical Data Processing and Transmission: Advances and Perspectives , 2017, 2018 European Conference on Optical Communication (ECOC).

[2] Jimmy Ba,et al. Adam: A Method for Stochastic Optimization , 2014, ICLR.

[3] Ken Mishina,et al. Experimental demonstration of neural network based demodulation for on-off encoded eigenvalue modulation , 2019 .

[4] Chao Lu,et al. Correlated Eigenvalues of Multi-Soliton Optical Communications , 2019, Scientific Reports.

[5] Shingo Sato,et al. Demodulation of Eigenvalue Modulated Signal Based on Eigenvalue-Domain Neural Network , 2020, 2020 Optical Fiber Communications Conference and Exhibition (OFC).

[6] Ken Mishina,et al. Hyper Multi-level Modulation Based on Optical Eigenvalue Multiplexing , 2018, 2018 Photonics in Switching and Computing (PSC).

[7] A. Hasegawa,et al. Eigenvalue communication , 1993 .

[8] Metodi P. Yankov,et al. Time-Domain Neural Network Receiver for Nonlinear Frequency Division Multiplexed Systems , 2018, IEEE Photonics Technology Letters.