TY - JOUR

T1 - Using deep learning to enhance event geometry reconstruction for the telescope array surface detector

AU - Ivanov, D.

AU - Kalashev, O. E.

AU - Kuznetsov, M. Yu

AU - Rubtsov, G.

AU - Sako, T.

AU - Tsunesada, Y.

AU - Zhezher, Y.

N1 - We would like to thank Anatoli Fedynitch, John Matthews, Maxim Pshirkov, Hiroyuki Sagawa, Gordon Thomson, Petr Tinyakov, Igor Tkachev and Sergey Troitsky for fruitful discussion and comments. We gratefully acknowledge the Telescope Array collaboration for support of this project on all its stages. The cluster of the Theoretical Division of INR RAS was used for the numerical part of the work. We appreciate the assistance of Yuri Kolesov in configuring the high-performance computing system based on graphic cards. The work is supported by the Russian Science Foundation grant 17-72-20291. This work was partially supported by the Collaborative research program of the Institute for Cosmic Ray Research (ICRR), the University of Tokyo.

PY - 2021/3

Y1 - 2021/3

N2 - The extremely low flux of ultra-high energy cosmic rays (UHECR) makes their direct observation by orbital experiments practically impossible. For this reason all current and planned UHECR experiments detect cosmic rays indirectly by observing the extensive air showers (EAS) initiated by cosmic ray particles in the atmosphere. The world largest statistics of the ultra-high energy EAS events is recorded by the networks of surface stations. In this paper we consider a novel approach for reconstruction of the arrival direction of the primary particle based on the deep convolutional neural network. The latter is using raw time-resolved signals of the set of the adjacent trigger stations as an input. The Telescope Array (TA) Surface Detector (SD) is an array of 507 stations, each containing two layers plastic scintillator with an area of 3 m(2). The training of the model is performed with the Monte-Carlo dataset. It is shown that within the Monte-Carlo simulations, the new approach yields better resolution than the traditional reconstruction method based on the fitting of the EAS front. The details of the network architecture and its optimization for this particular task are discussed.

AB - The extremely low flux of ultra-high energy cosmic rays (UHECR) makes their direct observation by orbital experiments practically impossible. For this reason all current and planned UHECR experiments detect cosmic rays indirectly by observing the extensive air showers (EAS) initiated by cosmic ray particles in the atmosphere. The world largest statistics of the ultra-high energy EAS events is recorded by the networks of surface stations. In this paper we consider a novel approach for reconstruction of the arrival direction of the primary particle based on the deep convolutional neural network. The latter is using raw time-resolved signals of the set of the adjacent trigger stations as an input. The Telescope Array (TA) Surface Detector (SD) is an array of 507 stations, each containing two layers plastic scintillator with an area of 3 m(2). The training of the model is performed with the Monte-Carlo dataset. It is shown that within the Monte-Carlo simulations, the new approach yields better resolution than the traditional reconstruction method based on the fitting of the EAS front. The details of the network architecture and its optimization for this particular task are discussed.

KW - ultra-high energy cosmic rays

KW - machine learning

KW - telescope array observatory

KW - ENERGY COSMIC-RAYS

KW - ARRIVAL DIRECTIONS

KW - FLUORESCENCE DETECTORS

KW - SCALE ANISOTROPY

KW - EEV

KW - DISTANCES

KW - SEARCH

KW - FLUX

U2 - 10.1088/2632-2153/abae74

DO - 10.1088/2632-2153/abae74

M3 - Article

VL - 2

JO - Machine learning-Science and technology

JF - Machine learning-Science and technology

SN - 2632-2153

IS - 1

M1 - 015006

ER -