ryuwd/offline

Offline software for the Mu2e experiment

Hi! 👋

I'm more active over on CERN GitLab!

offline's Issues

Set up track parameter description for Millepede

So far: 5-parameters from FullDriftFit t0, A0, A1, B0, B1.

Set up free parameter description for all planes using AlignableObjects

Set up global parameters for all planes dx, dy, dz, and dalpha, dbeta, dgamma angles using AlignableObjects. This choice is made since the MisalignTracker utility currently expresses misalignments as a translation vector and rotations about x, y, z axis.

Each free parameter per plane must have a unique 'label' or unsigned integer id number, for Millepede to keep track of. My simple convention is this:

plane_id * 10 + parameter_id where 0 <= parameter_id < 6.

So, for plane ID 27 global parameter dy (id: 1), the label would be 271.

Millepede-II set-up

Set up mille binary file output for input to ./pede application, supplying the required info from an art module.
Millepede-II will be given the straight track trajectory kalman fit result for straight cosmics

• Set up unique labelling of plane degrees of freedom
• Error on the residual measurement - comes from Kalman fit result - radialErr()
• Move to new track parameterisation - Straight track Kalman fit
• Residual derivatives wrt kalman track parameters (https://arxiv.org/pdf/0810.2241.pdf)
• Appropriate choice of constraints to suppress weak degrees of freedom
• Use unbiased residual calculations from BTrk using 'inverse kalman filter step'

**Note Title**

Here is another custom note card.

Calculation of residual derivatives with respect to Plane global params and Track local params

Currently finding these for Millepede. Millepede needs partial derivatives of DOCA/DT wrt. each global and local parameter.

Optimised MLE unbiased residual calculation

Set up, once again, unbiased track residual calculation but only use FullDriftFit with the output of the fit on the full set of straws.

Running Minuit 2 is not suited to track fitting as mentioned here, and Kalman appears to be the way to go, but in the meantime I will attempt to see if it converges faster this way than before. It's also worth considering that algorithms optimised to perform MLE may converge a lot faster than a black box solver.

In anticipation of obvious merge conflicts, write this functionality in a separate file as a helper.

MilleWrapper class

For #9.