Research Round-up: November - December 2020

It was a productive end of year for the TRANSNET Programme, especially in the publications department. Here’s our research round-up from November and December, featuring papers on machine learning in ultrafast photonics, dual frequency combs and digital back propagation. Plus, our papers from ECOC!

Journal Paper
Fast mode decomposition in few-mode fibers (
Nature Communications

This paper presents a high-performance mode decomposition mathematical algorithm with a processing time of tens of microseconds, several orders of magnitude faster than alternative machine learning techniques. The results are an important step towards the development of future cost-efficient phase retrieval receivers for various applications of few-mode fibers, ranging from imaging to telecommunications.

Journal Paper 
Machine learning and applications in ultrafast photonics (
Nature Photonics

TRANSNET Deputy Director, Sergei Turitsyn from Aston University, is a co-author on this Review paper which examines ‘machine-learning ultrafast photonics’ as an established and active field, and also considers the challenges and future directions as research continues to progress in the development of machine-learning algorithms and photonics technologies. The Review is also authored by colleagues from Tampere University, University Bourgogne Franche-Comte and Novosibirsk State University.

Journal Paper
Noise and distortion analysis of dual frequency comb photonic RF channelizers (
Optics Express

This paper assesses the fundamental performance limitations of dual frequency comb channelisers for radio frequency signal processing. Theoretical modelling is verified using numerical stimulations which demonstrate that dual comb-based processors are ultimately sensitive to the relative phase noise between the two frequency combs in high bandwidth scenarios.

Journal Paper
Complex-Valued Neural Network Design for Mitigation of Signal Distortions in Optical Links (
Journal of Lightwave Technology 

In this paper, a new neural network for nonlinear compensation – placed at the receiver end of optical transmission systems – is presented and its performance examined numerically and experimentally in several fibre systems. Compared to previously proposed neural networks and standard digital back-propagation, the new design resulted in significant performance improvements. It was also able to mitigate impairments arising in transmitter and receiver components as well as nonlinear transmission distortions. The research is a collaboration across Aston University, Infinera, the Technical University of Munich and Telecom Italia.

Journal Paper
Digital Back Propagation via Sub-band Processing in Spatial Multiplexing Systems (
Journal of Lightwave Technology

This work shows, for the first time, that in spatial multiplexed systems transmission performance can be improved by sub-band processing of back propagated channels. The method presented in the paper uses a separate channels approach to digital back propagation and is an important step towards the development of digital nonlinear compensation for spatial multiplexed systems.

Journal Paper
Modelling the delayed nonlinear fiber response in coherent optical communications (
Journal of Lightwave Technology

In this paper, the effect of the real part of the Raman spectrum on nonlinear interference is explored. An extension to the inter-channel stimulated Raman scattering Gaussian Noise model, to include the real Raman spectrum, is proposed and validated by numerical simulations. The results show that the impact of the real Raman spectrum can be considered as a multiplicative factor to the instantaneous response. The new version of the model, which now accounts for the complete Raman response, may be used in the design and operation of optical communication transmission links and networks.

Conference Papers at ECOC
The European Conference on Optical Communication (ECOC) was held virtually in early December, and several TRANSNET members presented papers, including our PI Polina Bayvel. Full details in the link below. 
TRANSNET speakers at ECOC

 For a full list of published TRANSNET research, including journal and conference papers, please take a look at our publications page.