Task leads: Ed Moyse (UMAS), Mark Owen (Glasgow)

The goal of this task is to fully exploit the extended coverage of the Level-0 muon trigger (T2.2) and the novel tracking infrastructure (ACTS) developed in T2.6 to improve the physics performance of the Event Filter muon track reconstruction. Migrating to ACTS should significantly reduce the computing resources needed for muon reconstruction and potentially provide an enhanced precision for the (combined) muon track fitting. Using ACTS will also facilitate porting parts of the muon track reconstruction onto accelerator hardware. Novel pattern recognition algorithms using Machine Learning may further improve the efficiency and technical performance of the muon reconstruction. We will also study the potential of porting such novel algorithms onto accelerators like GPUs or FPGAs. As a result ATLAS will be able to handle the increased rate of Level-0 muon candidates within the available processing resources to retain more interesting events with muons in the final state.