Optical networks are evolving toward ultrawide bandwidth and autonomous operation. In this scenario, it is crucial to accurately model and control optical power evolutions (OPEs) through optical amplifiers (OAs), as they directly affect the signal-to-noise ratio and fiber nonlinearities. However, a fundamental contradiction arises between the complex physical phenomena in optical transmission and the required precision in network control. Traditional theoretical methods underperform due to ideal assumptions, while data-driven approaches entail exorbitant costs associated with acquiring massive amounts of data to achieve the desired level of accuracy. In this work, we propose a Bayesian inference framework (BIF) to construct the digital twin of OAs and control OPE in a data-efficient manner. Only the informative data are collected to balance the exploration and exploitation of the data space, thus enabling efficient autonomous-driving optical networks (ADONs). Simulations and experiments demonstrate that the BIF can reduce the data size for modeling erbium-doped fiber amplifiers by 80% and Raman amplifiers by 60%. Within 30 iterations, the optimal controlling performance can be achieved to realize target signal/gain profiles in links with different types of OAs. The results show that the BIF paves the way to accurately model and control OPE for future ADONs.

Chaotic optical communication has shown large potential as a hardware encryption method in the physical layer. As an important figure of merit, the bit rate–distance product of chaotic optical communication has been continually improved to 30 Gb/s × 340 km, but it is still far from the requirement for a deployed optical fiber communication system, which is beyond 100 Gb/s × 1000 km. A chaotic carrier can be considered as an analog signal and suffers from fiber channel impairments, limiting the transmission distance of high-speed chaotic optical communications. To break the limit, we propose and experimentally demonstrate a pilot-based digital signal processing scheme for coherent chaotic optical communication combined with deep-learning-based chaotic synchronization. Both transmission impairment recovery and chaotic synchronization are realized in the digital domain. The frequency offset of the lasers is accurately estimated and compensated by determining the location of the pilot tone in the frequency domain, and the equalization and phase noise compensation are jointly performed by the least mean square algorithm through the time domain pilot symbols. Using the proposed method, 100 Gb / s chaotically encrypted quadrature phase-shift keying (QPSK) signal over 800 km single-mode fiber (SMF) transmission is experimentally demonstrated. In order to enhance security, 40 Gb / s real-time chaotically encrypted QPSK signal over 800 km SMF transmission is realized by inserting pilot symbols and tone in a field-programmable gate array. This method provides a feasible approach to promote the practical application of chaotic optical communications and guarantees the high security of chaotic encryption.

We propose a general guideline on the design of a stimulated Brillouin scattering (SBS)-based microwave photonic filter (MPF) using a directly modulated pump. Filter gain profiles and passband ripples with waveform repetition periods of the driving current ranging from 2 to 100 ns are measured after the transmission of different fiber lengths. The results show that the filter performance has nothing to do with the fiber length, and the digital-to-analog converter bandwidth requirement for the driving current is no more than 500 MHz. Therefore, the low cost, flexible reconfiguration, and miniaturization characteristics make an SBS filter using a directly modulated pump a promising choice as an MPF.

We extensively discuss 25 Gb/s per wavelength capacity in both IEEE and ITU-T standardization to support the increasing bandwidth requirement. In this Letter, we propose to use the optical dispersion compensation technique in an optical line terminal (OLT) combined with a bandwidth-limited electro-absorption modulated laser in an optical network unit to achieve 25 Gb/s capacity for the upstream link. We evaluate the positive and negative dispersion tolerances of 25 Gb/s electrical duo-binary (EDB) and pulse-amplitude modulation (PAM-4) signals. We achieve 39.5 and 31 dB upstream loss budgets for the 25 Gb/s EDB and PAM-4 signals by using 600 and 500 ps/nm optical dispersion compensation in OLT, respectively, both supporting 0–40 km differential reach.

We experimentally demonstrate a real-time quasi-full-duplex 400G/300G optical interconnection over 20 km multicore fibers (MCFs), using 10G-class transponders operated in the C-band. Optical delay interferometer (ODI)-based optical frequency equalization is applied to mitigate chirp and dispersion induced impairments, so that the tolerance to inter-symbol interference (ISI) can be enhanced, thus enabling 4×25 Gb/s on-off keying (OOK) transmission per core over severely limited bandwidth channels. Real-time bit error ratio (BER) performances of the bidirectional 400 Gb/s transmission are measured without using extra digital signal processing (DSP) or electrical equalization, which ensures low complexity and less power consumption.

By transferring 100 nm gold-coated CVD monolayer graphene onto the well-polished surface of D-shaped fiber, we achieve a graphene in-line polarizer with a high polarization extinction ratio of ～27 dB and low insertion loss of 5 dB at 1550 nm, meanwhile achieving a strong saturable absorption effect of 14%. The manufacture of thisgraphene in-line polarizer also simplifies the graphene transfer process. To explore the potential applications of the new device, we also demonstrate noise-like pulse generation and supercontinuum spectrum generation. By launching the designed graphene device into a fiber ring laser cavity, 51 nm bandwidth noise-like pulse is obtained. Then, launching the high-power noise-like pulse into high nonlinear fiber, a 1000 nm wide supercontinuum spectrum is obtained, which is favorable for sensing and nonlinearities scientific fields.