The objective of this work is to find an adaptive Forward Error Correction (FEC) scheme that runs at the application layer, which will compensate for future channel inequalities. The Predictive FEC algorithm will take into account channel variability and then add parity bits directly at the application based on the estimated channel fluctuations.
The student will use MATLAB/Python to build:
- the channel predictor using Gaussian Process Regression models (or other ML) and,
- determine the amount of FEC to use given the predicted channel.
If proficient, the students will be encouraged to play with network simulators (NS-3, OMNET++ …) to build working solutions.
20% Theory, 60% Simulations, 20%Writing.
Alessandro Chiumento (firstname.lastname@example.org)