ANALYSIS OF THE EFFECT OF BINNING ON EFFICIENCY IN ACCUMULATOR PLOTS USING THE HOUGH TRANSFORM

Abstract

The ATLAS detector at the LHC is involved in a search for physics beyond the Standard Model. This search involves high energy collisions of transient particles, and as such can be a challenge for the trigger system that detects the trajectories of the particles. Furthermore, a planned upgrade to the LHC will be increasing the collision rate and energies involved. Therefore, the trigger system needs to be reworked and improved. One plan for this upgrade involves custom algorithms implemented on accelerator cards with FPGAs. This report concerns an efficiency analysis of one such algorithm, the Hough Transform, with a straight-track approximation on a single-Muon dataset. We aim to find the binning parameters that maximize the efficiency of our algorithm while still being able to complete the computation in a reasonable time. To do this, we perform a comprehensive computational search of our parameter space. We find that the greatest efficiencies are found when fewer bins are allocated to each coordinate direction, but note that our dataset is very simple, and further work still needs to be done. In the future, we hope to benchmark our algorithm with these parameters on actual FPGA hardware.