Dynamic noise-masked 3D constellation encryption scheme based on gated recurrent unit networks.

Area of Science:

• Optical Communications
• Cryptography
• Machine Learning

Background:

• Secure data transmission is crucial for high-capacity optical networks.
• Existing encryption methods face challenges in dynamic key generation and security against sophisticated attacks.

Purpose of the Study:

• To propose a dynamic noise-masked 3D constellation encryption scheme for enhanced security and performance in optical communications.
• To leverage gated recurrent unit (GRU) neural networks and chaotic systems for secure, data-synchronized key generation.

Main Methods:

• Utilized a three-dimensional six-row stacking (3D-SRS) structure for constellation shaping.
• Employed a GRU neural network to dynamically generate initial keys for a chaotic model, synchronized with transmitted data.
• Implemented noise-masking technology for secure key embedding and recovery.

Main Results:

• The 3D-SRS constellation achieved a 0.31 dB performance gain over the traditional cubic (TC) structure at the forward error correction (FEC) threshold.
• The chaotic system provided a key space of 10^134, significantly enhancing resistance to brute-force attacks.
• Dynamic key generation showed variations of at least 10^-2, preventing key reuse.

Conclusions:

• The proposed scheme effectively improves transmission security in optical communications.
• The dynamic noise-masked 3D constellation encryption supports high-reliability and high-capacity encrypted optical networks.