Speaker
Description
The Electron-Ion Collider (EIC), which is scheduled to begin operations at Brookhaven National Laboratory in the 2030s, will explore the quark and gluon structure of nucleons and nuclei, including the emergence of gluon saturation at high densities. Through high-energy collisions of polarized electrons with protons and nuclei, the EIC will provide new insights into the structure and dynamics of strongly interacting matter. The ePIC experiment is being developed as the first detector at the EIC to realize this broad physics program. This study aims to develop algorithms and a software framework for online physics event identification for the ePIC experiment.
The ePIC experiment adopts a streaming data acquisition system based on continuous detector readout to maximize its physics reach. Detector data are aggregated into fixed-duration timeframes (0.6 ms) that contain detector responses from multiple physics interactions together with beam-induced background and detector noise. In this approach, physics events are identified in software rather than by hardware triggers. Together with AI-based methods, this approach also enables real-time experimental monitoring and control, as well as calibration and full event reconstruction within a few weeks of data taking.
In this study, we are extending the EICrecon software framework to support streaming reconstruction. At the core of EICrecon, JANA2 provides a hierarchical data processing model that allows data to be processed at different levels, including the timeframe level and the physics-event level. EICrecon builds on this capability to process data from continuous detector readout in timeframes and to identify temporally localized physics-event candidates within each timeframe. The selected candidates are then passed to the standard event-level reconstruction chain. We evaluated the performance of the developed event-identification algorithm using simulations that incorporate beam-induced backgrounds at realistic rates. In this presentation, we will describe the streaming event-identification approach implemented in EICrecon and present its event-identification efficiency. We will also briefly discuss how this approach has been applied to recent ePIC test-beam measurements.
