cGAN Model-Based Radio Frequency Interference Mitigation for Radio Astronomy Data

Abstract

Radio astronomy is one of the essential branches of space sciences where astronomers explore the universe by collecting data using various tools. The radio telescope is one of the principal tools for receiving celestial objects' emissions. How-ever, radio frequency interference (RFI) detection, mitigation, and avoidance are some of the main challenges in astronomical radio data. Additionally, they are essential steps for selecting the best site to initiate the radio telescope. RFI mitigation is arduous because interference can take a wide range of forms and affects different scientific goals. The substantial challenges of handling large radio data volumes make it a good application of deep learning (DL). The research aims to mitigate the interference using a DL-based approach, specifically, conditional generative adversarial network (cGAN), because of its powerful ability to differentiate the interference and the clean data.