Short External Cavity Laser-based Reservoir Computing for Fiber Transmission Data Recovery
Apostolos Argyris, Ingo Fischer · 2025
Reservoir computing (RC) has emerged as a powerful framework for tackling complex computational tasks and efficient signal processing by harnessing the dynamical properties of physical systems. In this work, we explore the use of time-delay RC with semiconductor lasers (SLs) and ultra-short external feedback cavities for data recovery in direct-detection optical transmission systems. We focus on a 200 km transmission link employing 28 GBaud PAM-4 modulation, where signal integrity is heavily degraded by chromatic dispersion (CD) and self-phase modulation (SPM). The RC is applied as a post-processing tool after optical detection to mitigate these impairments. We evaluate the advantages of incorporating both amplitude and phase information from the received signal, while using short reservoir lengths and strong feedback coupling which collectively enhance the system’s memory capacity.