BER Degradation Prediction Based on SOP Monitoring Enabling Local and Remote Transceiver Reconfiguration

Machine-learning algorithms are proposed for excessive bit-error-rate prediction and the estimation of its evolution over time, based on state-of-polarization (SOP) real-time monitoring. The accuracy of the developed algorithms is demonstrated, which enable its use for local and remote transceiver reconfiguration.