Intelligent Optical Transport Networks

This research intends to explore the future dynamic optical networks with flexible network functions and fast network reconfigurations.  The envisioned dynamic optical networks could deliver network services quickly in a low-latency manner.  The flexible network resource can accommodate network requests by providing just-enough network hardware to maximize the utilisation of network resources.  The margin-reduced dynamic optical network will be one of the key research themes for intelligent optical transport.  To achieve this, two key research directions are explored.

AI-driven dynamic network automation

Dynamic network automation will explore network abstraction, performance prediction, on-line device diagnostic by analysing network operation and performance monitoring data.  The implemented network-scale cloud network operation and monitoring database stores all the information about the operation of the whole network.  With advanced machine learning technologies, network resource can be dynamic, abstracted, planned and deployed according to the real network resources.  Machine-Learning algorithms can also provide aging information about key devices.  Several machine-learning algorithms, including Deep Neural Networks (DNN), Reinforcement Learning (RL), and Random Forest Tree have been explored for Quality of Transmission (QoT) prediction, network resource allocations, and other applications.

Programmable multi-dimensional networks

Programmable optical hardware is one of the key enabling technologies for dynamic networks.  This research will investigate programmable network architecture, reconfigurable node functions and programmable transmission equipment.  Our developed architecture-on-demand synthesized optical nodes and bandwidth programmable optical transmitters provides a solid foundation for hardware-programmable optical networks.  The research intends to bring programmability in multiple dimensions (i.e. TDM/WDM/SDM in optical networks).  The flexibility of network hardware can adapt the capability, according to network states and network requirements, to contribute to the vision of intelligent optical transport.

For further information, please contact:

Dr Shuangyi Yan

(Lecturer in Optical Communications)


Edit this page