The Track Extrapolation package in the new ATLAS Tracking scheme

Andreas Salzburger, CERN PH-ATC & University of Innsbruck


The transport of track parameters and their associated covariances is a fundamental and very frequent operation in most track reconstruction algorithms. This is due to the fact that most track fitting techniques are based on the first and second momentum of the underlying probabilty distribution of the track parameters in the estimation procedure. The correct treatment of energy loss and multiple scattering during this extrapolation process are together with the precise knowledge of the magnetic field crucial for the precision that can be achieved in track reconstruction.

For the high track density predicted for LHC experiments special emphasism has to be put on the stability and timing performance of extrapolation process. In addition, a very detailed knowlegde of material in the detector and the magnetic field such as consequently its correct treatment during the extrapolation process is needed to fullfill required physics performance.


Figure 1.: Schematic illustration of a propagation process of track parameters (point, arrow) and their associated convariances (ellipse, cone) between surfaces.

Concrete Description of the Applicaiton

In the context of the general restructuring of the ATLAS reconstruction software a new track extrapolation package has been developed to serve this functionality. The software was designed with high granularity in the various steps of the track extrapolation process. This enables systematic studies of single effects (detail of material description, magnetic field parameterizations, etc.) on the track reconstruction, and to learn how to regulate these effects in the simulation.

The track extrapolation can be in general divided into two different steps:

  • purely mathematical propagation
  • interaction with material during propagation

A correct propagation requires in addition the infrastructure to get the material passed by the track such as the magnetic field access at any stage. For this purpose a complete new software framework has been established, including the development of a fully navigable reconstruction geometry that is characterized through a connective volume geometry. Currently three propagation methods are in use, following a straight line or a helical track model, such as a Runge-Kutta based propagation. The material integration for this propagation methods has been implemented. In addition, a new propagation method that takes material effects in a continous way during the Runge-Kutta based propagation into account has been developed and integrated into the extrapolation scheme. This new propagation method has not been used so far in track reconstruction and will allow a more realistic treatment of the passage of muons through dense material. It will serve as a the main propagation method for combined Inner Detector - Muon System tracking (passage through Calorimeter, material in Muon System). The construction of this reconstruction geometry is closely bound to the common ATLAS detector description GeoModel to avoid duplication of detector description specifications.


Figure 2.: Illustration of a propagation with construction of a curvilinear frame for convariance propagation.

In prospect of the upcoming "as-build material description" of the ATLAS Data Challange 3, the material integration has been designed in a modular way, to be adopted and to allow studies of the impact of this more realistice material description on the ATLAS track reconstruction performance.

The intrinsic navigation of the track extrapolation enables a predictive extrapolation within this geometry. By changing the material interaction to be dependend on a random number generation the extrapolation serves as a track creation engine producing Monte Carlo tracks by using the reconstruction geometry. This is crucial for the validation of various tracking algorithms (track fitters, holes on track search, etc.). It also enhances dedicated studies of the effect of disagreement in the material description between simulation and reconstruction to a level that can not be achieved with full simulation. Results of ongoing studies using this newly developed Monte Carlo simulation in respect of algorithm validation and in comparison to fully simulated data will be given.

In the Combined Testbeam 2004 (CTB2004) the extrapolation package became the new standard for track finding and reconstruction of the Inner Detector. This allowed various studies on real taken data including different particle types and energies, as well as different setups of the magnetic field. This feedback was integrated into the the software validation.

The multiple use of the new track extrapolation in the ATLAS reconstruction and physics analysis software can be shown by the big number of client algorithms that have incorporated this package:

  • Different track fitters have been implemented on top of the extrapolation package and can be used for track fitting, re-fitting independently of sub-detector technology. These fittes will also be used in the commissioning phase of the ATLAS detector with cosmics runs.
  • The extrapolation from Inner Detector and Muon System tracks to the Calorimeter and the matching with Calorimeter clusters takes use of the extrapolation package and is widely used in the e/gamma reconstruction in ATLAS.
  • The vertex reconstruction for displaced vertices uses the correct transport of trackparameters and their associated covariances to the vertex canditate.
  • The holes-on-track search uses both, the navigation and the extrapolation from the new package.
  • parts of the ATLAS alignment software uses the new trak fitters for refittend and therefor directly depend on the new extrapolation
  • combined track fitting with tracks from the stand alone muon reconstruction has been done

This has a major impact on the overall physics performance of the ATLAS detector and the precision of various future measurements.

-- AndreasSalzburger - 29 Jun 2005

Edit | Attach | Watch | Print version | History: r2 < r1 | Backlinks | Raw View | WYSIWYG | More topic actions
Topic revision: r2 - 2005-06-29 - AndreasSalzburgerSecondary
    • Cern Search Icon Cern Search
    • TWiki Search Icon TWiki Search
    • Google Search Icon Google Search

    Main All webs login

This site is powered by the TWiki collaboration platform Powered by PerlCopyright & 2008-2020 by the contributing authors. All material on this collaboration platform is the property of the contributing authors.
or Ideas, requests, problems regarding TWiki? use Discourse or Send feedback