Online muon reconstruction in the ATLAS Muon Spectrometer at the level-2 stage of the event selection

To cope with the 40 MHz event production rate of LHC, the ATLAS experiment uses a multi level trigger architecture that selects events in three sequential steps, increasing the complexity of reconstruction algorithms and accuracy of measurements with each step. The Level-1 is implemented with custom hardware and provides a first reduction of the event rate to 75KHz, identifying physics candidates within a small detector region (Region of Interest, RoI). The higher trigger levels, the Level-2 and the Event Filter, are running on dedicated PC farms where the event rate is further reduced to O(400) Hz by software algorithms. At Level-2, the selection of the muon events is initiated by the “MuFast” algorithm, which confirms the muon candidates by means of a precise measurement of the muon candidate momentum. Designed to use a negligible fraction of the Level-2 latency, this algorithm exploits fast tracking and Look Up Table (LUT) techniques to perform the muon reconstruction with the precision muon chamber data within the RoI. The quality of the track parameters is good and approaches that of the offline reconstruction in some detector regions, and enables Level-2 to achieve a significant reduction of the event rate. This paper presents the current state of the art of the L2 algorithm “MuFast” reviewing its performance on the event selection, and the algorithm optimization achieved by using data taken in 2010.