An Autonomous Identification and Pre-distortion Scheme for Cognitive Transceivers using Bayesian Optimization

We propose a Bayesian optimization based DSP algorithm that autonomously tunes Volterra filter design and the parameters of digital pre-distortion schemes to enable transmitter optimal operation. We experimentally validate it for Volterra and memory polynomial filters in a 64 and 80 GBd DP-64QAM b2b setup.