No. 37 (00125)

Family name : Etzion
Given name : Erez
Affiliation : Tel-Aviv University, currently at CERN
Abbreviation : 
E-mail address : erez.etzion@cern.ch
Title : Using a neural network approach to improve the ATLAS muon 
reconstruction and triggering.
Authors : H. Abramowicz, Y. Benhammou, G. Dror, E. Etzion, D. Horn, L. 
Levinson, R. Livneh
Abstract : 
The extremely high rate of events that will be produced in the future
Large Hadron Collider (LHC)requires the triggering mechanism to take
precise decisions in a few nano-seconds. In the following we compare the
performance of the ATLAS dedicated electronic muon triggering system to
a neural network based solution. Both methods use the same limited
information from the muon chambers in the outer layers of the detector.
The neural network approach exhibits superior performance by
reconstructing the momentum of the muons passing through a highly
inhomogeneous magnetic field. We demonstrate the capability of the
neural network to perform this classification in a rapid and efficient
way. Furthermore, we show how a neural network can be used to tune the
hardware system built to perform this task.