Zibar Darko

Visit duration time (from – to): 31 May 2021 - 31 July 2021
Affiliation: Technical University of Denmark
Host Department: DET - Department of Electronics and Telecommunications
Coordinator at Polito: ANDREA CARENA

To satisfy the demands for future data rates, the next generation of optical communication systems will need to employ systems with an ultra-wideband transmission window (encompassing the O+E+S+C+L bands), coupled with the use of space-division multiplexing (SDM) via multicore and multimode fibers. Moreover, as the world is becoming increasingly focused on reducing CO2 footprint, power consumption of optical communication systems will also need to be take into account.

Until now, a figure of merit has been spectral-efficiency-transmission-distance product. However, revising this figure of merit into spectral-efficiency-transmission-distance-energy-efficiency product may be more appropriate to address the future challenges. 

Given the high-level of complexity of future generation optical communication systems and networks, machine learning will be imperative in maximizing the spectral-efficiency-transmission-distance-energy-efficiency product. This will require equipping photonic components such as lasers, frequency combs, modulators, optical amplifiers and receivers with some sort of intelligent control and interconnectivity. The objective of my resreach activity is therefore to explore how to  introduce an “artificial network brain (ANB)” to manage communication between components, make appropriate decisions and give instructions.  

More specifically, I would also like to perform research that focuses on machine learning schemes for nonlinear distortion-free communication over the fiber-optic channel, machine learning enabled programmable optical amplifiers and frequency combs.