Skip to content

Draft: Reviving the parametrised Kalman filter

Lennart Uecker requested to merge leuecker-updateKF into master

This MR adapts the existing parametrised Kalman filter to the current event model. Using the parametrised Kalman filter on long tracks improves momentum resolution, mass resolution and ghost rejection using the chi2 of the tracks.

On local test this come with a (Standalone: 30%, Gaudi-Allen 5%) penalty in throughput.

Last presented in WP2 on 15.10.24.

Current set of parameter files

Chi2 Ghost vs real Tracks D0 mass peak momentum resolution momentum bias
Edited by Lennart Uecker

Merge request reports