Kalman-filter-based track fitting in non-uniform magnetic field with segment-wise helical track model

TL;DR

This paper introduces a segment-wise helical track model for Kalman-filter-based track fitting in non-uniform magnetic fields, improving momentum resolution for future collider detectors like ILD.

Contribution

It extends the traditional helical track model to account for magnetic field non-uniformity by using segment-wise approximations and transformations within a Kalman filter framework.

Findings

Successfully implemented in KalTest software.

Enhanced track fitting accuracy in non-uniform magnetic fields.

Applicable to high-precision collider experiments like ILC.

Abstract

In the future International Linear Collider (ILC) experiment, high performance tracking is essential to its physics program including precision Higgs studies. One of major challenges for a detector such as the proposed International Large Detector (ILD) is to provide excellent momentum resolution in a magnetic filed with small (but non-negligible) non-uniformity. The non-uniform magnetic field implies deviation from a helical track and hence requires the extension of a helical track model used for track fitting in a uniform magnetic field. In this paper, a segment-wise helical track model is introduced as such an extension. The segment-wise helical track model approximates the magnetic field between two nearby measurement sites to be uniform and steps between the two sites along a helix. The helix frame is then transformed according to the new magnetic field direction for the next step, so as to take into account the non-uniformity of the magnetic field. Details of the algorithm and mathematical aspects of the segment-wise helical track model in a Kalman-filter-based track fitting in the non-uniform magnetic field are elaborated. The new track model is implemented and successfully tested in the framework of the Kalman filter tracking software package, KalTest, which was originally developed for tracking in a uniform magnetic field.