Simulating Radio Frequency Interference
Abstract
In the last few decades, huge advancements have been made towards measuring our Universe. However, key areas that have yet to be touched upon is what is known as the cosmic "Dark Ages" and the reionization of Hydrogen that follows it. Taking place 377,000 years post-Big Bang, the Dark Ages were a time devoid of any light. Then, between 150 million to 1 billion years, neutral Hydrogen began to re-ionize as objects started to compress, producing a 21-cm line signal. The Hydrogen Epoch of Reionization Array (HERA) is a key instrument in studying this otherwise inaccessible time in our Universe by utilizing this signal. This may seem to be a simple task, however, the frequency in which this signal emits is extremely faint making it difficult to detect. A way to combat this is to attempt to filter out objects, both human-generated and celestial, that emit signals at higher frequencies. Working closely with the HERA collaboration, an attempt was made to filter out radio frequency interferences (RFI) from human-generated objects by creating a script using the Python language that would simulate various versions of RFI data. The RFI that was studied and later replicated were randomly distributed signals called Scatter RFI, bursts that would emit signals at a higher frequency called Bursty RFI, and Narrowband RFI that would occupy a single frequency channel, while remaining, for the most part, constant throughout the time period observed. In this poster, we describe the different classes of RFI that our code simulates and discuss the challenges in creating a realistic RFI model in simulation. We then discuss how the newly simulated data would be used to train a machine learning algorithm that will ultimately be used to filter out RFI from the data that is being collected by HERA.
- Publication:
-
American Astronomical Society Meeting Abstracts #233
- Pub Date:
- January 2019
- Bibcode:
- 2019AAS...23334924E