6G D2D Wireless Communication System Using ML
Electrical Engineering
Team 3
Matthew Heusmann, Jeffrey Davis, Robert Fortunato, John Cappolella, Fabrizzio Arguello
Summary
Our project investigates how machine learning can improve wireless channel estimation in a realistic urban Device-to-Device communication scenario. We compare classical Least Squares against a Random Forest model trained to reconstruct missing subcarriers using channel correlation. The results show that machine learning reduces pilot overhead and delivers more accurate channel estimates.