Aachen 2019 – scientific programme
Parts | Days | Selection | Search | Updates | Downloads | Help
T: Fachverband Teilchenphysik
T 11: Neutrino-Astronomie I
T 11.9: Talk
Monday, March 25, 2019, 18:00–18:15, S10
Improving IceCube low energy event reconstruction — •Elisa Lohfink1, Maicon Hieronymus1, Sebastian Böser1, and Elmar Schömer2 for the IceCube collaboration — 1Institut für Physik, JGU Mainz, Deutschland — 2Institut für Informatik, JGU Mainz, Deutschland
Within the IceCube Neutrino Observatory and its low-energy extension (DeepCore), neutrinos with energies down to the GeV range can be reconstructed individually. The reconstruction is based on minimizing a six-dimensional likelihood space. Causality induced steep borders in the time-dimension make this likelihood space particularly difficult to evaluate, especially at reconstructed times later than the true event time. Currently, this challenge is overcome using MultiNest, a computationally expensive global nested sampling algorithm. In the new approach investigated here, directed likelihood sampling in the time-dimension is used to circumvent the issues induced by causality borders. This opens the possibility to use different, less costly minimization techniques and thus to significant speed-up of the reconstruction.