Optical chaos shift keying communication system via neural network-based signal reconstruction
摘要
Chaotic optical secure communication provides reliable physical layer protection for information transmission. In this work, we propose a neural network-based high-order optical chaos shift keying communication scheme. Digital information is modulated by dynamically switching the mapping relationships between the historical and future states of the emitter system. At the receiver, a pre-trained neural network-based discriminator identifies the underlying mapping relationships of this system for each symbol period, enabling demodulation without requiring physical-layer chaos synchronization. The simulation results indicate that the proposed scheme supports a data transmission rate of 2.67 Gbps at a signal-to-noise ratio (SNR) of 10 dB and maintains a reliable transmission rate of 0.66 Gbps even at a low SNR of 5 dB. Furthermore, the chaotic signal generated by the scheme exhibits strong resilience against return map attacks and time-delay signature (TDS) extraction. These results demonstrate the capability of the proposed scheme to maintain reliable communication in high-noise environments.