Speaker
Description
The next generation of neutrino experiments requires high-precision data acquisition systems capable of processing massive data volumes with minimal latency. This work presents the development of the CAIPORA (Compact Artificial Intelligence Placed On Reprogrammable Array), an intelligent trigger system implemented on Field-Programmable Gate Arrays (FPGAs) for the real-time identification of rare, time-dependent events within the photon detection system of a Liquid Argon Neutrino detector.
Utilizing simulated data from Scintillation detectors, we propose a hardware-accelerated approach to distinguish between astrophysical neutrino signals—specifically potential galactic supernova bursts—and radiological backgrounds. The study focuses on the implementation of low-latency algorithms, including simplified neural networks and pulse-shape discrimination, directly into the FPGA fabric.
By processing the scintillation light collected by the sensor arrays, CAIPORA serves as a non-human monitoring system with virtually instantaneous response times. This capability transforms the Liquid Argon Neutrino detector from a passive observer into a proactive detection agent, enabling the issuance of instantaneous alerts and integration into the SNEWS (Supernova Early Warning System). The results demonstrate that this intelligent trigger can make localized, high-speed decisions, significantly reducing data throughput while maintaining high efficiency for low-energy neutrino interactions.
| I read the instructions above | Yes |
|---|