forward tracking simulation
DESCRIPTION
Forward Tracking Simulation. Norman A. Graf ALCPG Cornell July 15, 2003. Overview. Develop track-finding strategies for the forward disk regions in SD. Detectors and support material defined. Ideal MC hits written out. Detector digitization post facto - PowerPoint PPT PresentationTRANSCRIPT
1
Forward Tracking Simulation
Norman A. Graf
ALCPG Cornell
July 15, 2003
2
OverviewDevelop track-finding strategies for the forward
disk regions in SD.Detectors and support material defined.Ideal MC hits written out.Detector digitization post facto• Study various readout schemes (pixel, stereo angle)• Hit merging and ghosting important.
Include beam backgrounds• Largest source of hits!
3
Tracking Strategy Find tracks in the 5-layer CCD pixel VXD,
extrapolate forward to pick up hits in the disks.
Conformal mapping of 3D hits simplifies pattern recognition.
Find tracks from outer radius in, fit, then extrapolate fitted track outwards.
Use Kalman filter to reconcile track hit prediction with measured strips.
4
Tracking in VXDPattern recognition for well-measured,
separated 3D points is not a problem.Five layers provide sufficient redundancy.Test pattern recognition in simplified events
events: 1, 10, 100, 200, 500 /event3, 5• 4 < < 176• 1GeV < E < 10GeV
5
Pattern RecognitionConformal-mapping technique applied to 3D
hits in VXD and forward disks.• Hits smeared by expected resolutions:• 5 in r and z for CCD
• 7 in r and r for FWD
• No hit merging!• No ghosts!
Treat as combined system:• Find VXD+FWD tracks in forward region.
6
1000 Events 500/Event
-3 -2 -1 0 10
100
200
300
400
500
600
700
800
Found-ShouldHaveFound
Missed ~250/500,000 (99.95%)
7100 200 300 400
1
10
100
1000 mean : 35.172 rms : 41.218
Tracking Time (miiliseconds)
Track-Finding Time/Event
500 /event700MHz
PIII
8
Hit-Merging in VXDCurrently record exact position of MC track’s
intersection with sensitive volume in simulations.
Smear with expected measurement resolution• Default is 5 microns.
Hits are currently distinct, even when they are within a pixel (20 microns!).
Real hits populate ~2×1 set of pixels.SLD experience being studied.
9
Adding BackgroundsBackgrounds arising from pairs hitting the
beampipe have been generated and passed through the full simulation packages.
One can overlay such events from an arbitrary number of beam crossings onto signal events.
10
Single Beam Crossing
11
Single Beam Crossing
12
Hit Densities vs. Radius
SD
T. Maruyama
13
0 50 100 150 2000
100
200
300
400
500
600
700
800
900
1,000
1,100
1,200
1,300
1,400
1,500
1,600
1,700
1,800
1,900
2,000
2,100 mean : 37.556 rms : 21.329
VXDHits - Layer1 number of hits
0 20 40 60 80 100 120 140 1600
100
200
300
400
500
600
700
800
900
1,000
1,100
1,200
1,300
1,400
1,500
1,600
1,700 mean : 62.165 rms : 20.584
VXDHits - Layer0 number of hits
0 20 40 60 80 100 120 1400
200
400
600
800
1,000
1,200
1,400
1,600
1,800
2,000
2,200
2,400
2,600
2,800
3,000
3,200
3,400
3,600
3,800
4,000 mean : 13.867 rms : 12.765
VXDHits - Layer2 number of hits
0 20 40 60 80 100 1200
500
1,000
1,500
2,000
2,500
3,000
3,500
4,000
4,500
5,000
5,500
6,000
6,500
7,000
7,500
8,000 mean : 6.1518 rms : 7.8351
VXDHits - Layer3 number of hits
14
VXD TrackingSystematic investigation of pattern
recognition in presence of full backgrounds not yet completed, but no problems experienced yet.
Effects of hit-merging to be studied.• not expected to be a major factor.
Efficiencies and resolutions as a function of and momentum being compiled.
15
Forward Tracking1. Extrapolate found tracks to hit strips.
2. Associate strip hits (either double-sided or back-to-back single-sided) in wedges of z-disks to form 3D space points.• Need systematic study of occupancies for various
designs.• Can we survive the ghosts? Grow as ~n2!
3. Use pixel hits if available. Detailed hit merging and ghosting needed!
16
Forward Disk DetectorsMany open issues:• Tiling of disks with wafers:• Phi segmentation?
• Radial segmentation?
• Mix of Si pixel and -strip detectors?• If pixel, APD or CCD?• If -strip, double-sided or back-to-back?• Strip orientations within wedges.• Shallow- or large-angle stereo?
17
Tiling Forward Disks
18
Strip Orientations
Shallow Angle Stereo Large Angle Stereo
19
Next StepsCurrently developing flexible tools to study
effects of disk tiling and strip orientations.Will start characterizing various detector
layouts in terms of efficiencies and resolutions.
Continue with impact on physics channels of interest for forward regions.
20
SummaryStrategies being implemented to handle
pattern recognition in the forward regions.Detector digitization infrastructure (hit
merging and ghosts) is almost in place.Recognize that detector design requires
reconstruction input.
Aim for flexible framework to allow iteration.