| Name: | Anshuman Tripathi |
| Affiliation: | Indian Institute of Technology Indore |
| Conference ID : | ASI2022_738 |
| Title : | Extraction of HI 21cm signal from Low-Frequency Radio Observations using Artificial Neural Networks. |
| Authors : | Anshuman Tripathi, Abhirup Datta, Madhurima Choudhury, Suman Majumdar |
| Abstract Type: | Oral |
| Abstract Category : | General Relativity and Cosmology |
| Abstract : | The evolution of our Universe from the Dark ages to the Epoch of reionization (EoR) remains unexplored. Detection of redshifted HI 21-cm signals is one of the primary scientific goals of radio telescopes like the MWA, SKA, and a promising probe for studying these epochs. The HI signal can be measured by averaging over the entire sky using a single radio telescope, or it can be measured using an interferometer. However, detection of 21-cm is an observational challenge due to the much brighter foreground of the galactic and extragalactic sources. At low frequencies, the Earth's ionosphere distorts the signal significantly. We use Neural Networks to extract the global 21-cm signal parameters from the composite all-sky averaged signal, containing foreground and ionospheric effects. Ionospheric effects are added considering the effects of refraction, absorption, and thermal emission. This trained model can help us to achieve better results for ground-based observations. Our initial results show significant accuracy in recovering the signal parameters from Cosmic Dawn (CD) and Epoch of Reionization (EoR) using the synthetic observations of global signal experiments. |