Delay-based reservoir computing for signal recovery in optical communication systems
Yunfan Zhang, Tiegen Liu, Jian Zhao, Tianhua Xu · Optical Fiber Technology · 2025
Over the past few years, reservoir computing (RC) has garnered significant attention, with delay-based RC being proposed as a promising paradigm for realizing RC networks in hardware. In this paper, we have implemented and evaluated two major RC methods: one based on a Mach–Zehnder modulator and the other on a semiconductor laser with external injection and optical feedback. We conducted investigations to evaluate the capabilities of these methods in performing signal recovery tasks in an optical transmission system, which was also numerically simulated for various transmission scenarios. Specifically, the neighboring symbol pattern technique was introduced to enhance the chromatic dispersion compensation ability of these two RC methods. The results indicate that, despite their differing advantages, both methods can perform as well as classic digital signal processing (DSP) techniques, even for high modulation formats such as 16QAM, 64QAM, and 256QAM. This demonstrates their great potential as alternatives to traditional DSP approaches.