Çìê»áìê¹¹¹ìëë Ìöö Blockinòò ××× Óò Ððùððö Ùøóññøóò Òò Ããðññò Ðøøö

Cellular Automaton and Kalman Filter D. Emeliyanov1;2 I. Gorbounov2;3 I. Kisel4;5 1 Institute of Control S ien es, Profsoyuznaya ul., 65, Mos ow, 117806 Russia 2 Deuts hes Elektronen-Syn hrotron DESY, Notkestrasse 85, D-22607 Hamburg, Germany 3 University of Siegen, Department of Physi s, D-57068 Siegen, Germany 4 Joint Institute for Nu lear Resear h, Dubna, Mos ow region, 141980 Russia 5 Max-Plan k-Institut f ur Physik, Werner Heisenberg Institut, Fohringer Ring 6, D-80805 M un hen, Germany Abstra t This note is devoted to the new tra k re onstru tion pa kage OTR/ITR-CATS developed for the Pattern Tra ker of the HERA-B experiment. This pa kage employs a ombined approa h for tra k re onstru tion based on using the ellular automaton for tra k sear hing and the Kalman lter for tra k tting. A similar re onstru tion strategy has been already su essfully applied to the Vertex Dete tor System (VDS) [1, 2℄. However, hit eÆ ien ies and resolutions of the Pattern Tra ker lower than those of the VDS require mu h more deli ate implementation of the method. The developed pa kage has been tested on simulated data and ompared with two other ompetitive algorithms: default HERA-B re onstru tion pa kage RANGER [3℄ and re onstru tion algorithm TEMA [4℄ based on Hough transform. The results of the tests regarding re onstru tion eÆ ien y, a ura y of estimates and omputing time are presented. 1 Introdu tion The HERA-B dete tor is a xed target spe trometer designed for studies in the eld of heavy quark physi s. The pattern tra ker of the dete tor is a part of the main tra king system and in ludes outer and inner tra king dete tors whose di erent granularities mat h to di erent tra k densities.