A CNN and LSTM-Based Approach to Classifying Transient Radio Frequency Interference

TL;DR

This paper introduces a novel deep learning approach combining CNNs and LSTMs to classify transient radio frequency interference sources, aiding in RFI mitigation for radio astronomy.

Contribution

It is the first to apply CNN and LSTM-based deep learning techniques to classify transient RFI sources in radio astronomy data.

Findings

Achieved promising classification accuracy on experimental RFI data.

Demonstrated potential for real-time RFI source identification.

Provides a foundation for developing automated RFI mitigation tools.

Abstract

Transient radio frequency interference (RFI) is detrimental to radio astronomy. It is particularly difficult to identify the sources of transient RFI, which is broadband and intermittent. Such RFI is often generated by devices like mechanical relays, fluorescent lighting or AC machines, which may be present in the surrounding infrastructure of a radio telescope array. One mitigating approach is to deploy independent RFI monitoring stations at radio telescope arrays. Once the sources of RFI signals are identified, they may be removed or replaced where possible. For the first time in the open literature, we demonstrate an approach to classifying the sources of transient RFI (in time domain data) that makes use of deep learning techniques including CNNs and LSTMs. Applied to a previously obtained dataset of experimentally recorded transient RFI signals, our proposed approach offers good results. It shows potential for development into a tool for identifying the sources of transient RFI signals recorded by independent RFI monitoring stations.