Récepteur multiutilisateur basé sur l'apprentissage profond pour améliorer la liaison montante dans les réseaux LoRa
Résumé
This paper presents a deep learning-based multiuser receiver to improve the uplink in LoRa Network. The proposed
receiver can decode simultaneous transmissions from multiple users on the same frequency channel with the same spreading factor.
The objective of the proposed approach is to reduce the high computational complexity and error propagation issues of the classical
SIC method. Each step employs a specific convolutional neural network model to directly decode each user’s signal. Simulation
results show that the proposed receiver significantly reduces the error propagation and increases the number of connected devices in
the network compared to the classical SIC scheme.
Domaines
Sciences de l'ingénieur [physics]
Origine : Fichiers produits par l'(les) auteur(s)