Deep Learning for AGILE Anticoincidence System's Background Prediction from Orbital and Attitude Parameters

AGILE is an Italian Space Agency (ASI) space mission launched in 2007 to study X-ray and gamma-ray phenomena in the energy range from ∼20 keV to ∼10 GeV. The AGILE AntiCoincidence System (ACS) detects hard-X photons in the 50 - 200 keV energy range and continuously stores each panel's count rates in the telemetry. We developed a new Deep Learning (DL) model to predict the background of the AGILE ACS top panel using the satellite's orbital and attitude parameters. This model aims to learn how the orbital and spinning modulations of the satellite impact the background level of the ACS top panel. The DL model executes a regression problem, and is trained with a supervised learning technique on a dataset larger than twenty million orbital parameters' configurations. Using a test dataset, we evaluated the trained model by comparison of the predicted count rates with the real ones. The results show that the model can reconstruct the background count rates of the ACS top panel with an accuracy of 96.7%, considering the orbital modulation and spinning of the satellite. Starting from these promising results, we are developing an anomaly detection method to detect Gamma-ray Bursts when the differences between predicted and real count rates exceed a predefined threshold.