Deep learning-based robust algorithm for automatic modulation recognition

• Main Author: Wang, Yucheng
• Other Authors: Teh Kah Chan
• Format: Final Year Project
• Language: English
• Published: Nanyang Technological University 2024
• Subjects: Engineering; Automatic modulation recognition
• Online Access: https://hdl.handle.net/10356/176262
• Institution: Nanyang Technological University

The project introduced a Deep Learning-based Automatic Modulation Recognition (DL- AMR) program employing various neural network architectures (CNN, ResNet, and LSTM) to enhance signal processing in 5G communication systems. It aimed to autonomously recognise modulation types of signals without prior knowledge, a task traditionally performed with manual feature engineering. The performance of these neural networks was compared by analysing accuracy in relation to signal-to-noise ratio (SNR). The discussion revolved around the classification accuracy of different modulation schemes at various SNRs, leading to the identification of the most effective DL-AMR model. Recommendations on future research in this area were highlighted finally.