SMuK 2023 – wissenschaftliches Programm
Bereiche | Tage | Auswahl | Suche | Aktualisierungen | Downloads | Hilfe
T: Fachverband Teilchenphysik
T 9: DAQ NN/ML – HW
T 9.3: Vortrag
Montag, 20. März 2023, 17:00–17:15, HSZ/0301
FPGA-based fast Machine Learning Triggers for Neutrino Telescopes — •Francesca Capel1,3, Christian Haack2,3, Lukas Heinrich2,3, and Christian Spannfellner2,3 — 1Max-Planck-Institut für Physik — 2Technische Universität München — 3ORIGINS Excellence Cluster
Neutrinos provide valuable insight into the origin and acceleration mechanisms of cosmic particles. They are able to traverse vast distances and dense environments on their way to Earth unimpeded, but are also challenging to detect due to their weakly interacting nature. Earth itself is used as detector, where large volumes are equipped with photosensors to detect the Cherenkov light induced by astrophysical neutrino interactions. Neutrino telescopes are located deep underwater or in the Antarctic ice to reduce the background rate, inducing often strict limits on power and bandwidth available for the detector. Trigger algorithms are inevitable to reject background signals and reduce the data stream to manageable rates. In this contribution we will present the potential of fast, intelligent machine learning triggers implemented on low power FPGAs for the usage as online trigger in neutrino telescopes. Our main objectives are an improved signal to background discrimination and improved sensitivity for low energy events.