Study of event topology for a new fast primary vertex finder for the ATLAS trigger

Abstract

This document presents a transform-based approach to primary signal vertex finding feasibility analysis. An algorithm to find primary vertices has been developed using only Inner Detector measurements. Based on the information available to the algorithm, a metric has been developed to identify the signal vertex. The feasibility analysis demonstrates the theoretical distinguishability of different signal events and pileup with this metric. The results show that the majority of event types are easily distinguished. The algorithm can be used in the high level trigger. At this stage of computation, event types can be identified with the trigger information, such that this algorithm can be chosen only for appropriate events. An implementation of the algorithm is used to run simulations with increasingly realistic scenarios. This allows the quantification of the algorithm performance and to attribute inefficiencies to specific effects. In the most realistic scenario, the distinction rate reduces only by a small margin, remaining higher than 90% for events with good results in the feasibility analysis, also for high pileup scenarios. Countermeasures are presented for the effects reducing efficiency.