Muon Event Filter Software for the ATLAS Experiment at LHC

At LHC the 40 MHz bunch crossing rate dictates a high selectivity of the ATLAS Trigger system, which has to keep the full physics potential of the experiment in spite of a limited storage capability. The Level-1 trigger, implemented in a custom hardware, will reduce the initial rate S. Armstrong a , K. Assamagan a , A. dos Anjos a , J.T.M. Baines c , C.P. Bee d , M. Biglietti e , M. Bellomo w , J.A. Bogaerts f , V. Boisvert f , M. Bosman g , B. Caron h , P. Casado g , G. Cataldi i , G. Carlino dd D. Cavalli j , M. Cervetto k , G. Comune l , F. Conventi dd , P. Conde Muino f , A. De Santo m , M. Diaz Gomez n , M. Dosil g , N. Ellis f , D. Emeliyanov c , B. Epp o , S. Falciano p , A. Farilla q , S. George m , V. Ghete o , S. Gonzlez r , M. Grothe f , S. Kabana l , A. Khomich s , G. Kilvington m , N. Konstantinidis t , A. Kootz u , A. Lowe m , L. Luminari p , T. Maeno f , J. Masik v , A. Di Mattia p , C. Meessen d , A.G. Mello b , G. Merino g , R. Moore h , P. Morettini k , A. Negri w , N. Nikitin x , A. Nisati p , C. Padilla f , N. Panikashvili y , F. Parodi k , V. Perez Reale l , J.L. Pinfold h , P. Pinto f , M. Primavera i , Z. Qian d , S. Resconi j , S. Rosati f , C. Sanchez g , C. Santamarina f , D.A. Scannicchio w , C. Schiavi k , E. Segura g , J.M. de Seixas b , S. Sivoklokov x , R. Soluk h , E. Stefanidis t , S. Sushkov g , M. Sutton t , S. Tapprogge z , E. Thomas l , F. Touchard d , B. Venda Pinto aa , A. Ventura i , V. Vercesi w , P. Werner f , S. Wheeler h,bb , , F.J. Wickens c , W. Wiedenmann r , M. Wielers cc , G. Zobernig r . a Brookhaven National Laboratory (BNL), Upton, New York, USA, b Universidade Federal do Rio de Janeiro, COPPE-EE, Rio de Janeiro, Brazil, c Appleton Laboratory, Chilton, Didcot, UK, d Centre de Physique des Particules de Marseille, IN2P3-CNRS-Université d’AixMarseille 2, France, e University of Michigan, Ann Arbor, Michigan, USA, f CERN, Geneva, Switzerland, g Institut de Fsica d’Altes Energies (IFAE), Universidad Autnoma de Barcelona, Barcelona, Spain, h University of Alberta, Edmonton, Canada, i Dipartimento di Fisica dell’Università di Lecce e I.N.F.N., Lecce, Italy, j Dipartimento di Fisica dell’Università di Milano e I.N.F.N., Milan, Italy, k Dipartimento di Fisica dell’Università di Genova e I.N.F.N., Genoa, Italy, l Laboratory for High Energy Physics, University of Bern, Switzerland, m Department of Physics, Royal Holloway, University of London, Egham, UK, n Section de Physique, Université de Genve, Switzerland, o Institut fr Experimentalphysik der Leopald-Franzens Universitt, Innsbruck, Austria, p Dipartimento di Fisica dell’Università di Roma ’La Sapienza’ e I.N.F.N., Rome, Italy, q Dipartimento di Fisica dell’Università di Roma ’Roma Tre’ e I.N.F.N., Rome, Italy, r Department of Physics, University of Wisconsin, Madison, Wisconsin, USA, s Lehrstuhl fr Informatik V, Universitt Mannheim, Mannheim, Germany, t Department of Physics and Astronomy, University College London, London, UK, u Fachbereich Physik, Bergische Universitt Wuppertal, Germany, v Institute of Physics, Academy of Sciences of the Czech Republic, Prague, Czech Republic, w Dipartimento di Fisica Nucleare e Teorica dell’Università di Pavia e INFN, Pavia, Italy, x Institute of Nuclear Physics, Moscow State University, Moscow, Russia, y Department of Physics, Technion, Haifa, Israel, z Institut fr Physik, Universitt Mainz, Mainz, Germany, aa CFNUL Universidade de Lisboa, Faculdade de Cincias, Lisbon, Portugal, bb University of California at Irvine, Irvine, USA, dd Dipartimento di Fisica dell’Università degli studi di Napoli “Federico II” e I.N.F.N., Napoli, Italy to 75 kHz and is followed by the software based Level-2 and Event Filter, usually referred as High Level Triggers (HLT), which further reduce the rate to about 100 Hz. In this paper an overview of the implementation of the offline muon recostruction algortihms MOORE (Muon Object Oriented REconstruction) and MuId (Muon Identification) as Event Filter in the ATLAS online framework is given. The MOORE algorithm performs the reconstruction inside the standalone Muon Spectrometer, thus providing a precise measurement of the muon track parameters outside the calorimeters; MuId combines the measurements of all ATLAS sub-detectors in order to identify muons and provides the best estimate of their momentum at the production vertex. In the HLT implementation the muon reconstruction can be executed in ”full scan mode”, performing pattern recognition in the whole muon spectrometer, or in the ”seeded mode”, taking advantage of the results of the earlier trigger levels. An estimate of the execution time will be presented along with the performances in terms of efficiency, momentum resolution and rejection power of muons coming from hadron decays and of fake muon tracks, due to accidental hit correlations in the high background environment of the experiment. THE ATLAS HIGH LEVEL TRIGGER The LHC bunch crossing frequency will be 40 MHz and will have to be reduced to the order of 100 Hz by the ATLAS Trigger and Data Acquisition (TDAQ) systems in order to achieve the foreseen storage capability and meet the physics requirements of the experiments. The Level-1 trigger (LVL1) [1], implemented in a custom hardware, will make the first level of event selection, reducing the initial event rate to less than 75 kHz in less then 2.5 μs. Operation at up to about 100 kHz is possible with somewhat increased dead time. The result of LVL1 contains informations about the type of trigger and the position of possible particle candidates that causes the event to be accepted. The second (LVL2) and third level, called Event Filter (EF), are software based systems and are referred togheter as High Level Triggers (HLT). The HLT must reduce the event rate further down to O(100) Hz. Each selected event will have a total size of 1.5 Mbyte giving a required storage capability of a few hundred Mbyte/s. The LVL2 and the Event Filter differ in several important respects. The LVL2 is composed of a combination of high rejection power with fast, limited precision algorithms using modest computing power; the Event Filter instead has a modest rejection power with slower, high precision algorithms using more A T L -D A Q -C O N F20 05 -0 08