UCL launches £6.1m research programme in the future of intelligent optical network

Today, UCL researchers, along with collaborators at Aston and Cambridge Universities, have launched a 6-year research programme to develop an adaptive and intelligent optical network, transforming the way networks are designed today to cope with the yet unknown needs of tomorrow's society.

The goal of the new research programme, entitled 'Transforming networks - building an intelligent optical infrastructure (TRANSNET)', is to create inteligent optical fibre networks, automatically assessing need and dynamically providing capacity.

Optical networks underpin the entire digital communications and data infrastructure, yet are a finite resource, that is to say, there is a maximum amount of data that can travel down each fibre at any one time. TRANSNET will enable the use of this finite resource with maximum efficiency.

The research will investigate the use of state-of-the-art artificial intelligence (AI) and machine learning(ML) techniques to control elements of the network in response to different requirements and priorities of the users.  

TRANSNET is funded by a £ 6.1m research grant from the Engineering and Physical Science Research Council (EPSRC). In addition, the research is undertaken in collaboration with 28 industrial and academic partners contributing a further £5.6m to the work. 

Professor Polina Bayvel CBE, Co-Director of the Institute of Communications and Connected Systems and head of the UCL Optical Networks research group, will lead the programme. Explaining the work Polina said:

“We envisage that machine learning (ML) will become ubiquitous in future optical networks, at all levels of design and operation, from digital coding, equalisation and impairment mitigation, through to monitoring, fault prediction and identification, and signal restoration, traffic pattern prediction and resource planning.We envisage that machine learning (ML) will become ubiquitous in future optical networks, at all levels of design and operation, from digital coding, equalisation and impairment mitigation, through to monitoring, fault prediction and identification, and signal restoration, traffic pattern prediction and resource planning. We envisage that machine learning will be integrated throughout future optical networks, at all levels of design and operation, from digital coding, equalisation and impairment mitigation, through to monitoring, fault prediction and identification, and signal restoration, traffic pattern prediction and resource planning.

Expressing the importance of the work Professor Bayvel went on to say:

With the vast demands placed on data networks we are soon going to meet a crunch point where the yet un-known needs of our future society will not be met by current infrastructure. The next-generation of digital infrastructure needs more than raw capacity - it requires flexible resource and capacity provision in combination with low latency, simplified and modular network architectures to maximise data throughput when and where it is needed. This needs to be achieved with network resilience combined with overall security.