A Radio Modulation Data Augmentation Method Based on Improved CGAN

In recent years, the rapid development of 6G, edge computing, Internet of Things and other technologies has made wireless communication technology an important direction of modern scientific and technological progress. The research of wireless communication technology based on machine learning is facing the common pain point problem of lacking large-scale and high-quality data sets. The application of data-driven learning algorithms is restricted by its inherent characteristics, and wireless communication data has problems such as high collection difficulty, high cleaning cost, and large label data. Therefore, in order to develop wireless communication technology based on machine learning, we first need to meet the needs of high-quality wireless communication data sets. This paper proposes a radio modulation data augmentation method based on improved conditional generative adversarial networks (RMDA-CGAN), and verifies the effectiveness of the augmented data set through classification tasks.