Performance Degradation and Failure Detection Methods for Optical Networks Based on Machine Learning

Large network operators have to deal with dynamic network service changes, including scheduled or unscheduled outages, while keeping efficient service levels with different requirements. Currently available restoration techniques remain limited to after the fact detection of hard or catastrophic failures when the service has already been impacted. Machine learning technologies have been explored as a potential solution to enable proactive fault management and performance prediction. This project will develop innovative machine learning methods in the context of optical networks, which will allow accurate prediction ahead of time for proactive fault management and maintenance, before actual degradation and failure occur.

Faculty Supervisor:

Christine Tremblay;Christian Desrosiers

Student:

Partner:

Ciena Canada (Saint-Laurent, QC);TELUS Communications

Discipline:

Engineering

Sector:

Information and cultural industries; Manufacturing

University:

École de technologie supérieure

Program: