Linear Collider Forum



Home » Analysis and Reconstruction » Tracking & Vertexing » TrackFinder
Re: TrackFinder [message #570 is a reply to message #568] Wed, 02 August 2006 09:52 Go to previous message
jabernathy
Messages: 78
Registered: March 2006
Location: University of Victoria
wiene wrote on Wed, 02 August 2006 01:38

Hi Jason,

Why do you need the time of the maximum pulse contributing to a track for the likelihood fit?




When we have a row of signals on pads we determine which pad the track most likely went through by finding the highest peak in the row which contributed to this Track. Then, knowing the time of the highest peak we integrate the charge in each pad in the row over a small interval around that time.

It reduces the affect multiple signals and noise has on the collected charge by localizing the data sample.

Unfortunately it only works if you can either separate the signal into multiple pulses (one for each Track) or you know the estimated time the Track crossed the pad.

killenberg wrote on Wed, 02 August 2006 00:57



I think the idea was to separate a double pulse into two, and then two hits are reconstructed from the separated double pulses. But I don't know how to disentangle how much charge was in which pulse/hit. As far as I know, none of the existing algorithms do this separation.




That solution would make the TrackFitter easier for sure (I think Dean Karlen might do something like that in JTPC - I'll have to check).

If it's not possible to create two TrackerPulses because the signals are too hard to separate could it be possible to use two TrackerHits and then set the time for each hit to be the estimated peak time of each Track in the signal?

continued...



Would it be possible to have something like a track finder which produces a double track candidate, containing all hits from both tracks, and then do one likelihood fit for both tracks?




Theoretically I don't see anything wrong with that. Double tracks could just be passed as a single Track with 2 tracks returned in getTracks().

To sum it up, perhaps if the signals are separated enough in the TrackerData they can be split into two TrackerPulses / TrackerHits and the TrackFitter could just do a single Track fit to each. However if it is too difficult to determine how much charge to assign to each TrackerPulse the TrackFitter could use the double-fit technique.
 
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Previous Topic:INSTALL of MarlinTPC
Next Topic:TPCCondData: TPCConditions and ChannelCorrection
Goto Forum:
  


Current Time: Wed Nov 20 14:24:23 Pacific Standard Time 2019
.:: Contact :: Home ::.

Powered by: FUDforum 3.0.1.
Copyright ©2001-2010 FUDforum Bulletin Board Software