In this project, you will study the prediction of end-to-end throughput in a WiFi/PLC network. You will apply machine learning techniques on a dataset of real measurements, and study which features (link capacity, node distance, interference ratio, etc.) are useful and how accurate a prediction can be. Based on the results, you will implement a real machine-learned routing protocol.
This project will enable you to apply machine learning techniques to a real-world problem. Background in networking is not required, but if you are interested by networking, this project will give you an opportunity to acquire knowledge on the subject.
The student must have followed a machine learning class.
The student must know Python.
Knowledge of Linux tools is a plus.
Knowledge and/or interest in networking is a plus.