Secure SDM–OFDM Signal Transmission Based on Memristive Neural Network

Qing Zhong,Bo Liu,Jianxin Ren,Xiangyu Wu,Yaya Mao,Yu Bai,Shuaidong Chen,Rong Tang,Jianye Zhao,Zhirui Guo,Jiajia Shen,Feng Tian
DOI: https://doi.org/10.1016/j.optcom.2022.128030
IF: 2.4
2022-01-01
Optics Communications
Abstract:This paper proposes a physical layer security strategy of space division multiplex orthogonal frequency division multiplexing system based on memristive neural network (MNN). Among them, the two-level chaotic control, the application of the memristive chaotic model with conditional symmetry and the encryption of the MNN are all proposed for the first time in the access network. The memristive chaotic model with conditional symmetry generates 8 multiple masking factors and achieves high-security encryption. In the encryption process, the key space is selected by Logistic mapping, and the MNN encryption mechanism disturbances and compresses bits. The proposed encryption strategy has been proved on the experimental platform. The generated OFDM signals are modulated, which transmit data rate of 124.4 Gb/s over the 2 km multi-core fiber (MCF). Bit error rate (BER) and different anti-interference are analyzed in experiments. The key space of this proposed scheme is 2×10399, which is large enough number to effectively resist any brute force or illegal optical network units (ONUs) intrusion in passive optical network (PON). The results show that the proposed SDM–OFDM​ scheme has high sensitivity and security. This scheme has very reliable safety performance and benefit improvement, and is very promising in the future SDM–OFDM system.
What problem does this paper attempt to address?