In this paper, a flexible physical security coding scheme integrating chaotic neural network (CNN) and non-linear encryption is proposed for orthogonal frequency division multiplexing wavelength division multiplexing passive optical network (OFDM-WDM-PON). The scheme improved the flexibility, adjustability and the key space of chaotic encryption system by introducing chaos into neural networks. The system will encrypt the bit series, probability shaping points, and subcarriers position of the OFDM signal through linear encryption and non-linear encryption concurrently. Results show that a key sensitivity of 10⁻¹⁵ and a key space of more than 10²⁷⁹ can be achieved. The encrypted system's Lyapunov is 5.2631, along with 12 parameters can be dynamically changed in the range of 0∼5. Furthermore, when the bit error rate (BER) is less than 3.8×10⁻³, probabilistic shaping (PS) technology decreases power loss by around 0.5 dB. A 20.454 Gb/s data transmission experiment was successfully verified for a span of 25 Km single-mode fiber. According to the experimental results, the proposed encryption scheme is likely to be used in future OFDM-WDM-PON transmission systems.

中文翻译：

---

OFDM-WDM-PON结合混沌神经网络的灵活非线性物理安全编码方案

本文针对正交频分复用波分复用无源光网络(OFDM-WDM-PON)提出了一种结合混沌神经网络(CNN)和非线性加密的灵活物理安全编码方案。该方案通过在神经网络中引入混沌，提高了混沌加密系统的灵活性、可调整性和密钥空间。该系统将同时通过线性加密和非线性加密对OFDM信号的比特序列、概率整形点和子载波位置进行加密。结果表明，10 ^-15的键灵敏度和大于 10 ²⁷⁹的键空间可以实现。加密系统的李雅普诺夫为5.2631，有12个参数可以在0∼5范围内动态变化。此外，当误码率 (BER) 小于 3.8×10 ^{-3 时}，概率整形 (PS) 技术将功率损耗降低约 0.5 dB。20.454 Gb/s 数据传输实验在 25 公里单模光纤的跨度上得到了成功验证。根据实验结果，所提出的加密方案很可能用于未来的OFDM-WDM-PON传输系统。