A Machine Learning-Based Receiver for Mitigating Nonlinear Distortion in OFDM Systems for 5G

July 7, 2025·
Marcele O. K. Mendonça
,
Francisco J. Martín Vega
,
Geoffrey Eappen
,
Carla E. García
,
Abuzar B. M. Adam
,
Jorge Querol
,
Beatriz Soret
,
Symeon Chatzinotas
· 1 min read
Abstract
This work introduces a convolutional neural network (CNN)-based receiver designed for channel estimation, equalization, and symbol demapping in orthogonal frequency-division multiplexing (OFDM) systems. The proposed receiver is specifically developed to mitigate nonlinear distortions, particularly under low input back-off (IBO), thereby markedly enhancing signal quality and energy efficiency. By significantly reducing the bit error rate (BER) and outperforming traditional methods, the CNN-based receiver establishes itself as a highly robust solution for next-generation communication technologies. These advancements are pivotal for the development of reliable and efficient communication networks, particularly in the context of the fifth generation (5G) and beyond. Furthermore, this work contributes to the integration of terrestrial and non-terrestrial networks by addressing critical challenges in mitigating nonlinear distortions in satellite and hybrid communication systems using AI-driven techniques.
Type
Publication
IEEE International Mediterranean Conference on Communications and Networking (MeditCom)
publication

Abstract & Overview

This work introduces a convolutional neural network (CNN)-based receiver designed for channel estimation, equalization, and symbol demapping in orthogonal frequency-division multiplexing (OFDM) systems. The proposed receiver is specifically developed to mitigate nonlinear distortions, particularly under low input back-off (IBO), thereby markedly enhancing signal quality and energy efficiency. By significantly reducing the bit error rate (BER) and outperforming traditional methods, the CNN-based receiver establishes itself as a highly robust solution for next-generation communication technologies. These advancements are pivotal for the development of reliable and efficient communication networks, particularly in the context of the fifth generation (5G) and beyond. Furthermore, this work contributes to the integration of terrestrial and non-terrestrial networks by addressing critical challenges in mitigating nonlinear distortions in satellite and hybrid communication systems using AI-driven techniques.

Publication Metadata

  • Authors: Marcele O. K. Mendonça, Francisco J. Martín Vega, Geoffrey Eappen, Carla E. García, Abuzar B. M. Adam, Jorge Querol, Beatriz Soret, Symeon Chatzinotas
  • Venue: IEEE International Mediterranean Conference on Communications and Networking (MeditCom) (2025)
  • Research Areas: 6G-NTN, AI & Neuromorphic SatCom
  • DOI: 10.1109/meditcom64437.2025.11104481
  • Open Access Repository: ORBiLu – University of Luxembourg
6G-NTN AI & Neuromorphic SatCom