search

miranov25 / fastMCKalman

Fast simulation and performance parameterization - primary for ALICE3 and DUNE
MIT License
0 stars 1 forks source link

Logical fallacy in the method used to trigger Mirror in the reconstruction: found and implemented new method #13

Closed Trenchcoat95 closed 2 years ago

Trenchcoat95 commented 2 years ago

Logical fallacy in the method used to trigger Mirror in the reconstruction: since the particle has been mirrored in the simulation, checking if it will reack the same radius doesn't make sense (after the mirroring it will be the same). We should check if the step between the previous radius and the next radius is less than a certain amount instead (now set arbitrarily at 0.0001).

miranov25 commented 2 years ago

The fastMCKalman is not "searching" it uses points as they are in the MC. In case in MC the PropagateToMirror used also in the reco we will use it. In the future maybe better to put there flag, instead of using radius comparison ..

In real reconstruction the PropagateToMirror will be called in case angle too big ...

miranov25 commented 2 years ago

Real live use case (no ideal MC)

In case of loopers for tracks with the big inclination angle the clusters will be splitted to "several cluster" with no well defined position.

The measurements at such rows will be not relyable/precise. I real live - significant arc step will be used - at position where tan(phi)padlength >kpadwidth

The red point, blue point and "gray" points will be used in different way ..

image

miranov25 commented 1 year ago

Relates to #9