Institute of Communications and Connected Systems


Simplified Transceiver Architectures for High Capacity Optical Networks

This page has been archived and will no longer be updated

Holding image - Awaiting image submission

2 January 2016

Using simplified optical transceiver architectures and digital signal processing to enable an intelligent allocation of resources.

Funder RAEng
Amount £ 500 000

Project website https://www.ee.ucl.ac.uk/

Research theme logos - Intelligent High Capacity Networks
Research topics Optical Fibre Communications | Digital Signal Processing | Algorithms | Physical Layer


The phenomenal increase in the communications capability of electronic devices has been enabled by the ever-increasing speed of the global communications network, which is itself underpinned by optical fibre communications. The recent drive towards ubiquitous communications has been enabled by ever more efficient use of this installed fibre infrastructure, resulting in a reduced cost per bit transmitted.

However, in core networks, optical fibre communications is now approaching the limits of bandwidth efficiency and so a greater optical bandwidth will be used by employing more transceivers to increase overall capacity. Without transceiver optimisation, the cost per bit transmitted may stagnate and emerging applications, which rely on a significant increase in communication capacity, will simply not be possible due to cost constraints.

This project investigates low complexity optical transceivers (transmitters/receivers) as a solution to this issue. By using digital signal processing (DSP) to offset the shortcomings of simplified transceiver designs, significant reductions in the complexity of a transceiver can be achieved without a performance penalty. The core network capacity can then be increased whilst decreasing the cost per bit.

This work uses efficient simulation algorithms to investigate wide bandwidth communication channels, and develops cooperative DSP to exploit the large number of co-propagating signals in these scenarios; reducing computational complexity by minimising repeated operations.