| PhD Seminar

Name of the Speaker: Mr. Shubham Paul (EE19D407)
Guide: Prof. David Koilpillpai R
Venue: ESB-244 (Seminar Hall)
Date/Time: 24th September 2024(Tuesday), 2:30 PM
Title: Enhanced Faster than Nyquist Signalling using RNNs

Abstract :

The decoding of Highly Compressed Faster-than-Nyquist (FTN) signals using Recurrent Neural Networks (RNNs) is the focus of this talk. The proposed method is compared with traditional Equalization methods. It is demonstrated that RNN equalizers achieve near-optimal performance and are able to decode FTN signals at compression factors where traditional DFEs fail. The use of these RNN equalizers allows us to reliably communicate

more than twice as many symbols as the Nyquist signalling system at the same power and with nearly the same error probabilities at low complexity. We also demonstrate the robustness of the proposed RNN equalizer to receiver non-idealities.