Mainz 2012 – scientific programme
Parts | Days | Selection | Search | Updates | Downloads | Help
HK: Fachverband Physik der Hadronen und Kerne
HK 29: Schwerionenkollisionen und QCD Phasen
HK 29.5: Talk
Wednesday, March 21, 2012, 15:15–15:30, P 5
Application of a Kalman Filter in Track Reconstruction* — •Erik Krebs1, Jochen Markert1,2, and Joachim Stroth1,2 — 1Goethe-Universität, Frankfurt am Main — 2GSI Helmholtzzentum für Schwerionenforschung GmbH, Darmstadt
The Kalman filter is a mathematical method to estimate the evolution of linear, dynamical systems and has become an established method for track reconstruction. The Deterministic Annealing Filter is an extension of the Kalman filter that introduces competition between measurements and allows rejection of fake measurements.
An extended Kalman filter that takes multiple scattering and energy loss into account has been implemented for the HADES-experiment. The filter works with position information from reconstructed drift chamber segments and the results have been compared to a global Runge-Kutta track fit. To reduce systematic errors of the segment reconstruction for highly curvilinear tracks the Kalman filter has been revised to work directly with the wire information of the drift chambers. Additionally, track candidates may include measurements from multiple tracks. For these reasons, an extension of the Kalman filter, the Deterministic Annealing Filter, is used to discriminate measurements that do not belong to a track. Measurements are assigned weights and an annealing process is introduced to avoid local optima in measurement assignments.
*supported by BMBF grant 06 FY 9100 I, HIC for FAIR, EMMI and GSI