A combination of analysis techniques for efficient track reconstruction of high multiplicity events in silicon detectors

Research output: Article

1 Citation (Scopus)

Abstract

A novel combination of established data analysis techniques is proposed for the reconstruction of all tracks of primary charged particles created in high energy collisions. It uses all information available in a collision event while keeping competing choices open for as long as possible. Suitable track candidates are selected by transforming measured hits to a binned, three- or four-dimensional, track parameter space. The transformation is based on templates, taking advantage of the translational and rotational symmetries of a given detector. Subsequently, their number is further narrowed down by a Kalman filter-based technique. Track candidates and their corresponding hits form a highly connected network, a bipartite graph, where one allows for multiple assignments of hits to track candidates. The graph is cut into very many minigraphs by removing a few of its components. Finally, the hits are distributed among the track candidates by exploring a deterministic decision tree. A depth-limited search is performed, maximising the number of hits on tracks and minimising the sum of track-fit χ2. Simplified models of LHC silicon trackers, as well as the relevant physics processes, are employed to study the performance (efficiency, purity, timing) of the proposed method in the case of single or many simultaneous proton-proton collisions (so-called event pile-up), and for single heavy-ion collisions at the highest available energies.

Original languageEnglish
Article number113
JournalEuropean Physical Journal A
Volume54
Issue number6
DOIs
Publication statusPublished - jún. 1 2018

Fingerprint

detectors
silicon
collisions
protons
Kalman filters
piles
ionic collisions
charged particles
purity
templates
time measurement
physics
energy
symmetry

ASJC Scopus subject areas

  • Nuclear and High Energy Physics

Cite this

@article{130f379d5ee64fa192034bf7e7346065,
title = "A combination of analysis techniques for efficient track reconstruction of high multiplicity events in silicon detectors",
abstract = "A novel combination of established data analysis techniques is proposed for the reconstruction of all tracks of primary charged particles created in high energy collisions. It uses all information available in a collision event while keeping competing choices open for as long as possible. Suitable track candidates are selected by transforming measured hits to a binned, three- or four-dimensional, track parameter space. The transformation is based on templates, taking advantage of the translational and rotational symmetries of a given detector. Subsequently, their number is further narrowed down by a Kalman filter-based technique. Track candidates and their corresponding hits form a highly connected network, a bipartite graph, where one allows for multiple assignments of hits to track candidates. The graph is cut into very many minigraphs by removing a few of its components. Finally, the hits are distributed among the track candidates by exploring a deterministic decision tree. A depth-limited search is performed, maximising the number of hits on tracks and minimising the sum of track-fit χ2. Simplified models of LHC silicon trackers, as well as the relevant physics processes, are employed to study the performance (efficiency, purity, timing) of the proposed method in the case of single or many simultaneous proton-proton collisions (so-called event pile-up), and for single heavy-ion collisions at the highest available energies.",
author = "F. Sikl{\'e}r",
year = "2018",
month = "6",
day = "1",
doi = "10.1140/epja/i2018-12548-8",
language = "English",
volume = "54",
journal = "Zeitschrift fur Physik A Hadrons and Nuclei",
issn = "0939-7922",
publisher = "Springer New York",
number = "6",

}

TY - JOUR

T1 - A combination of analysis techniques for efficient track reconstruction of high multiplicity events in silicon detectors

AU - Siklér, F.

PY - 2018/6/1

Y1 - 2018/6/1

N2 - A novel combination of established data analysis techniques is proposed for the reconstruction of all tracks of primary charged particles created in high energy collisions. It uses all information available in a collision event while keeping competing choices open for as long as possible. Suitable track candidates are selected by transforming measured hits to a binned, three- or four-dimensional, track parameter space. The transformation is based on templates, taking advantage of the translational and rotational symmetries of a given detector. Subsequently, their number is further narrowed down by a Kalman filter-based technique. Track candidates and their corresponding hits form a highly connected network, a bipartite graph, where one allows for multiple assignments of hits to track candidates. The graph is cut into very many minigraphs by removing a few of its components. Finally, the hits are distributed among the track candidates by exploring a deterministic decision tree. A depth-limited search is performed, maximising the number of hits on tracks and minimising the sum of track-fit χ2. Simplified models of LHC silicon trackers, as well as the relevant physics processes, are employed to study the performance (efficiency, purity, timing) of the proposed method in the case of single or many simultaneous proton-proton collisions (so-called event pile-up), and for single heavy-ion collisions at the highest available energies.

AB - A novel combination of established data analysis techniques is proposed for the reconstruction of all tracks of primary charged particles created in high energy collisions. It uses all information available in a collision event while keeping competing choices open for as long as possible. Suitable track candidates are selected by transforming measured hits to a binned, three- or four-dimensional, track parameter space. The transformation is based on templates, taking advantage of the translational and rotational symmetries of a given detector. Subsequently, their number is further narrowed down by a Kalman filter-based technique. Track candidates and their corresponding hits form a highly connected network, a bipartite graph, where one allows for multiple assignments of hits to track candidates. The graph is cut into very many minigraphs by removing a few of its components. Finally, the hits are distributed among the track candidates by exploring a deterministic decision tree. A depth-limited search is performed, maximising the number of hits on tracks and minimising the sum of track-fit χ2. Simplified models of LHC silicon trackers, as well as the relevant physics processes, are employed to study the performance (efficiency, purity, timing) of the proposed method in the case of single or many simultaneous proton-proton collisions (so-called event pile-up), and for single heavy-ion collisions at the highest available energies.

UR - http://www.scopus.com/inward/record.url?scp=85049134416&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85049134416&partnerID=8YFLogxK

U2 - 10.1140/epja/i2018-12548-8

DO - 10.1140/epja/i2018-12548-8

M3 - Article

AN - SCOPUS:85049134416

VL - 54

JO - Zeitschrift fur Physik A Hadrons and Nuclei

JF - Zeitschrift fur Physik A Hadrons and Nuclei

SN - 0939-7922

IS - 6

M1 - 113

ER -