An extension of Associative Memory approach to tracking with a drift-tube detector using timing information and its demonstration for HL-LHC ATLAS muon trigger: On behalf of the ATLAS Collaboration

Yunjian He,Yasuyuki Okumura,Masahiro Kuze,Yohei Yamaguchi,Takafumi Kodama,Masaya Ishino

doi:10.1109/nss/mic42677.2020.9508066

Abstract

High-speed online pattern recognition has been a fundamental challenge for triggering in High Energy Physics (HEP) experiments. The Associative Memory (AM) approach has been developed and used in HEP experiments for online track-finding with silicon detectors. We intend to extend the AM approach to tracking with a drift-tube detector, using the drift-time as a “new dimension” of observables in addition to spatial information. Our benchmark study demonstrates the feasibility of the extended AM concept, aiming at the online muon reconstruction with the ATLAS Monitored Drift-Tube (MDT) detector for the Phase-II Level-0muon trigger system for High-Luminosity Large Hadron Collider. The online muon reconstruction will consist of two parts: (1) a fast track-segment finding, and (2) the following track reconstruction to estimate the momentum of the muons. It is found that timing information can be integrated into the AM approach in a natural way, and the AM approach can fit the needs for the fast track segment finding. In terms of hardware specifications expected for the Phase-II Level-0 muon trigger system, an optimal pattern training scheme is developed to prepare an effective set of AM patterns that provide high efficiency in the track-segment finding while keeping a good resolution. Based on the system-level design of electronics, an optimal algorithm chain has been developed to minimise the latency for the track segment finding. The detailed design and performance study shows that the AM approach has the capability of a high-speed and high-performance track-finding with drift-tube detectors.

Highlights

M AINTAINING a high-performance trigger system under the high luminosity conditions is a unique challenge towards the High-Luminosity Large Hadron Collider (HL-LHC) era [1]
The Associative Memory architecture is based on Content Addressable Memory (CAM) cells to identify sets of hits which are correlated on trajectories efficiently
Our study shows that the use of timing information of drift-tube detectors naturally fits the Associative Memory architecture

Summary

INTRODUCTION

M AINTAINING a high-performance trigger system under the high luminosity conditions is a unique challenge towards the High-Luminosity Large Hadron Collider (HL-LHC) era [1]. The AM-based pattern recognition has been used to form combinations of the silicon detector hit information for track trigger applications in HEP experiments, such as CDF Silicon Vertex Tracker (SVT) [3] and ATLAS Fast TracKer (FTK) [4]. The precise online reconstruction will be seeded by the fast muon trigger outputs from the dedicated trigger chambers with fast-response: Resistive Plate Chambers (RPCs) in the barrel region; Thin Gap Chambers (TGCs) in the endcap region. It consists of a straight-track finding within each detector station out of three stations (“segment finding”) and a muon track reconstruction by combining the segments among the stations to evaluate the momentum of the muons

AM APPROACH WITH DRIFT-TUBE DETECTORS

System design aiming at the LHC-ATLAS Phase-II Level-0 muon trigger system

The optimal method of online muon reconstruction and the momentum estimation

PERFORMANCE STUDY

Findings

CONCLUSION