@article{M039ED8CD,
title = "Research on Automatic Modulation Recognition Using Vision Transformer",
journal = "The Journal of Korean Institute of Communications and Information Sciences",
year = "2024",
issn = "1226-4717",
doi = "10.7840/kics.2024.49.8.1074",
author = "Minju Lee, Myoungho Chae, Wansu Lim",
keywords = "automatic modulation recognition, Vision transformer(ViT), Transformer encoder, wireless communication systems, Constellation",
abstract = "Automatic Modulation Recognition (AMR) is a technology that plays a key role in wireless communication systems, contributing to improving the efficiency of data communication and enhancing the reliability and security of wireless communication systems. Recently, due to the development of deep learning technology, research using deep learning has been actively conducted in the field of AMR. In this paper, we propose an AMR technique based on the ViT (Vision Transformer) model, which has excellent time series data processing capabilities. The ViT model divides the input image into patches, which are small image units, and assigns an order to each patch, which is used as an input to the transformer encoder. By doing so, the ViT-based AMR model learns the characteristics of each modulation scheme and automatically recognizes the modulation scheme. By using the ViT-based AMR model, we were able to achieve an average classification accuracy improvement of about 2% even at low SNR."
}