A Modified Convolutional Neural Network Model for Automatic Modulation Classification
Yogesh Beeharry, Didier Gael Daryl Emilien · 2025
In the evolving landscape of 5G/6G communication systems, modulation type detection is crucial for adaptive signal processing, ensuring efficient spectrum utilization and robust data transmission. Machine Learning (ML) and Deep Learning (DL) offer powerful solutions for automatic modulation classification (AMC), leveraging large datasets and complex models to improve accuracy and adaptability. These techniques enable systems to autonomously recognize modulation schemes such as quadrature amplitude modulation (QAM), phase shift keying (PSK), and orthogonal frequency division multiplexing (OFDM), even in challenging conditions with noise, interference, and varying channel characteristics. In this paper, a Convolutional Neural Network (CNN) model with enhanced hyper-parameters is presented. An accuracy of 90.8% is obtained from test data using the proposed model.