Tunable All-Optical Pattern Recognition System Based on Nonlinear Optical Loop Mirror for Bit-Flip BPSK Targets

Abstract

As the basic physical infrastructure of various networks, optical networks are crucial to the advancement of information technology. Meanwhile, as new technologies emerge, the security of optical networks is facing serious threats. To improve the security of optical networks, optoelectronic firewalls primarily leverage all-optical pattern recognition to perform direct detection and analysis of data transmitted through the optical network at the optical layer. However, the current all-optical pattern recognition system still faces some problems when deployed in optical networks, including phase-lockingand relatively low recognition efficiency and scalability. In this paper, we propose a tunable all-optical pattern recognition system based on a nonlinear optical loop mirror (NOLM) for bit-flip BPSK targets. The operational principles and simulation setup of the proposed system are comprehensively described. Numerical simulations demonstrate that the system can accurately recognize and determine the position of 4-bit and 8-bit bit-flip BPSK targets in 16-bit input data with tunable frequencies of 192.8 THz and 193.4 THz at a data rate of 100 Gbps. Finally, the impact of input noise is evaluated by extinction ratio (ER), contrast ratio (CR), Q factor, bit error rate (BER), amplitude modulation (AM), and signal-to-noise ratio (SNR) under both frequencies.