In search of global 21-cm signal using artificial neural network in light of ARCADE 2

Understanding the astrophysical nature of the<br/>first stars remains an unsolved problem in cosmology. The<br/>redshifted global 21-cm signal(T21) acts as a treasure trove to<br/>probe the cosmic dawn era – when the intergalactic medium<br/>was mostly neutral. Many experiments, like SARAS 3,<br/>EDGES, and DARE, have been proposed to probe the cosmic<br/>dawn era. However, extracting the faint cosmological signal<br/>buried inside a brighter foreground, O(104), remains challenging. Additionally, an accurate modelling of foreground<br/>and T21 signal remains the heart of any extraction technique.<br/>In this work, we constructed the foreground signal (TFG)<br/>from the global sky model and star formation history using<br/>Press–Schechter formalism to determine the T21 signal with<br/>excess radio background following ARCADE 2 detection.<br/>Further, we incorporated static ionospheric distortion into the<br/>total signal and calculated the signal measured by an ideal<br/>antenna. We then trained an artificial neural network (ANN)<br/>for the extraction of a T21 signal parameters signal measured<br/>by antenna with an R-square score (0.5523−0.9901). Lastly,<br/>we used a Bayesian technique to extract T21 signal and compared the finding with ANN’s extraction.