Skip to main content

Dr Tianhua Xu

1. Digital signal processing and machine learning in optical communication networks

In optical communication, high performance computing and data center systems, as high-speed and high-order modulated signals are applied, system performance will be significantly degraded by transmission impairments, such as bandwidth limitation, dispersion, polarization dependent loss, channel fading, laser phase noise and nonlinear distortions from optical channels and components. In this project, reconfiguration of optical communication networks and data center systems will be investigated based on digital signal processing and machine learning techniques in conjunction with software-defined transceivers, to compensate for transmission impairments and to realize the optimum detection of optical signals.

2. Information rates and channel estimation of optical networks

In optical communication networks, signals of different users are often multiplexed at different wavelengths, and will interact with each other due to the linear and nonlinear effects in optical channels and devices. Thus it is of importance to develop accurate physical estimation for nonlinear channels and components to assess the achievable capacity and mutual information of optical communication networks. In this research, fundamental limits of optical transparent networks will be studied considering the linear and nonlinear physical impairments in the link, such as chromatic dispersion, polarization mode dispersion, laser phase noise, self-/cross-phase modulation, four-wave mixing, channel memory etc.

 

3. Optimization in elastic optical networks to maximize the throughputs

Based on software-defined transceivers and elastic optical networks, the transmission parameters, e.g. forward error correction (FEC) schemes, modulation formats, frequency separation (in flex-grid networks), optical launch powers and symbol rates etc. will be adapted and tailored to physical channels and components in the transparent wavelength routed networks. In addition, probabilistic shaping and geometric shaping will be applied to optimize the signals to have better tolerance against the degradations. All these degrees of freedom will be jointly optimized in conjunction with the routing of lightpaths through the optical network to maximize the overall capacity and resources.


Note: Should your application for admission be accepted you should be aware that this does not constitute an offer of financial support. Please refer to the scholarships & funding pages.