AGILESim: Monte Carlo simulation of the AGILE gamma-ray telescope
V. Fioretti, A. Bulgarelli, M. Tavani, S. Sabatini, A. Aboudan, A. Argan, P. W. Cattaneo, A. W. Chen, I. Donnarumma, F. Longo, M. Galli, A. Giuliani, M. Marisal… 展开作者
V. Fioretti, A. Bulgarelli, M. Tavani, S. Sabatini, A. Aboudan, A. Argan, P. W. Cattaneo, A. W. Chen, I. Donnarumma, F. Longo, M. Galli, A. Giuliani, M. Marisaldi, N. Parmiggiani, A. Rappoldi
The accuracy of Monte Carlo simulations in reproducing the scientific performance of space telescopes (e.g. angular resolution) is mandatory for a correct design of the mission. A brand-new Monte Carlo simulator of the Astrorivelatore Gamma ad Immagini LEggero (AGILE)/Gamma-Ray Imaging Detector (GRID) space telescope, AGILESim, is built using the customizable Bologna Geant4 Multi-Mission Simulator (BoGEMMS) architecture and the latest Geant4 library to reproduce the instrument performance of the AGILE/GRID instrument. The Monte Carlo simulation output is digitized in the BoGEMMS postprocessing pipeline, according to the instrument electronic read-out logic, then converted into the onboard data handling format, and finally analyzed by the standard mission on-ground reconstruction pipeline, including the Kalman filter, as a real observation in space. In this paper we focus on the scientific validation of AGILESim, performed by reproducing (i) the conversion efficiency of the tracker planes, (ii) the tracker charge readout distribution measured by the on-ground assembly, integration, and verification activity, and (iii) the point-spread function of in-flight observations of the Vela pulsar in the 100 MeV - 1 GeV energy range. We measure an in-flight angular resolution (FWHM) for Vela-like point sources of $2.0^{+0.2}_{-0.3}$ and $0.8^{+0.1}_{-0.1}$ degrees in the 100 - 300 and 300 - 1000 MeV energy bands, respectively. The successful cross-comparison of the simulation results with the AGILE on-ground and in-space performance validates the BoGEMMS framework for its application to future gamma-ray trackers (e.g. e-ASTROGAM and AMEGO).
Preliminary Results of a New Deep Learning Method to Detect and Localize GRBs in the AGILE/GRID Sky Maps
AGILE是意大利航天局于2007年发射的一项空间飞行任务，目的是研究能量范围从$\sim20$ keV到$\sim10$ GeV的X射线和伽马射线现象。AGILE团队开发了一个实时分析管道，用于快速检测瞬态源，并跟踪通过通用坐标网络等网络收到的外部科学警报。我们开发了一种新的深度学习方法，用于检测和定位AGILE/GRID天空地图中的伽马射线暴（GRB）。我们使用带有GRB的天空地图训练模型，该地图在距离地图中心20度的半径内模拟，这大于GRBWeb目录中存在的99.5%的误差区域。我们还计划将这种方法应用于寻找引力波事件的对应物，这些事件通常具有更宽的定位误差区域。该方法包括两个深度学习模型，使用两个卷积神经网络实现。第一个模型检测和过滤包含伽玛射线暴的天空地图，而第二个模型定位其位置。我们使用模拟数据训练和测试模型。检测模型的准确率达到95.7%，定位模型的平均误差小于0.8度。我们配置了一个Docker容器，其中包含数据模拟所需的所有软件，并使用Amazon Web Service进行部署，以计算不同条件下的p值分布。通过p值分布，我们可以计算检测的统计显著性。
