ETD

Archivio digitale delle tesi discusse presso l'Università di Pisa

Tesi etd-06302016-110321


Tipo di tesi
Tesi di laurea magistrale
Autore
NASIMI, HIKMAT
URN
etd-06302016-110321
Titolo
The Fast Track Trigger Upgrade for the ATLAS Experiment High Rate Trigger Algorithms
Dipartimento
FISICA
Corso di studi
FISICA
Relatori
relatore Prof.ssa Roda, Chiara
Parole chiave
  • FTK
  • Associative Memory
  • Jet
  • LHC
  • ATLAS
  • Trigger
  • Pattern Matching
Data inizio appello
21/07/2016
Consultabilità
Completa
Riassunto
At LHC, in the context of the ATLAS experiment, for the so-called Phase-II, the instantaneous luminosity is expected to reach $5\cdot 10^{34} cm^{-2} s^{-1}$. In such conditions the overlapping collisions and the QCD multi-jet production makes triggering on heavy fermions (such as $b$ quarks and $\tau$ leptons) an extremely difficult challenge. The standard methods used at offline levels to select these objects such as $b$-tagging and primary vertex reconstruction require full granularity information from the tracker of the ATLAS detector and are therefore too time consuming to be implemented at trigger level. Within the ATLAS experiment, the FTK project has developed a custom processor that is capable of reconstructing all the tracks within an event after the Level-1 trigger. In fact FTK can provide real time tracking at 100 kHz using a highly parallelized and high performance system.


In my thesis I will describe the FTK system in detail, in the context of the the ATLAS experiment itself. I will also summarize the FTK performances showing that the early availability of reconstructed tracks immediately after the Level-1 increases the signal efficiency while maintaining the same background rate. I will also describe my work on the development of modules that allow to monitor and test the digital logic of the Associative Memory (AM) Board, the core of the FTK processing unit. The AM board in fact, is equipped with custom memory chips (AM chips) capable of ultra-parallized pattern matching. The memories receive the hit information from the tracker they give in output coarse reconstructed tracks that make the fast high quality track fitting, done by the FTK processing unit, possible.


Once the FTK system will be fully functional the High Level Trigger performance will be highly improved and it becomes important to optimize the Level-1 trigger to be able to profit at best of the improved trigger performance. For this reason a new approach on Level-1 selection of purely hadronic decays of the Higgs boson will be described in the second part of the thesis. The new strategies rely on global features of the event and not only in single thresholds.
The Randall-Sundrum Graviton with a mass of 500 GeV and decaying to a couple of Standard Model Higgs each one decaying to $b\bar{b}$ has been used as benchmark signal. The trigger algorithms developed follow two major strategies: the tagging of calorimetric clusters at Level-1 with muons from the semileptonic decay of the $b$ quark, done trough the exploitation of the newly installed topological trigger, and the selection based on clean multi-jet topological and kinematic configuration. Efficiencies on signal selection and background selection are calculated for both selection algorithms.
File