fix: GX2F: update track state parameters and converge #1702
Conversation
AJPfleger
added
Component - Core
Codecov Report
@@ Coverage Diff @@
## main #1702 +/- ##
==========================================
+ Coverage 49.03% 49.04% +0.01%
==========================================
Files 397 397
Lines 21564 21581 +17
Branches 9817 9824 +7
==========================================
+ Hits 10573 10585 +12
- Misses 4183 4186 +3
- Partials 6808 6810 +2
📣 We’re building smart automated test selection to slash your CI/CD build times. Learn more |
📊 Physics performance monitoring for 693359aFull report VertexingVertexing vs. muIVF seededIVF truth_smearedIVF truth_estimatedIVF orthogonalSeedingSeeding seededSeeding truth_estimatedSeeding orthogonalCKFCKF seededCKF truth_smearedCKF truth_estimatedCKF orthogonalAmbiguity resolutionseededTruth tracking (Kalman Filter)Truth trackingTruth tracking (GSF)Truth tracking |
There was a problem hiding this comment.
Very nice work! There are a few things that I think should be cleaned up, but in general I think we are good here.
I think you could think about either fixing the parameters type directly to BoundTrackParameters and removing the related code or going back to the generic, templated case. I would not stay somewhere in the middle.
) The previous implementation could not update the track state parameters. The updated Jacobian was overridden by `stepper.transportCovarianceToBound`. Therefore, every time the identity matrix was used in place of the Jacobian. A second problem was the handling of surfaces in the update steps. Now, during the first iteration, proxies are created for each surface. During further iterations those surfaces are visited again.
) The previous implementation could not update the track state parameters. The updated Jacobian was overridden by `stepper.transportCovarianceToBound`. Therefore, every time the identity matrix was used in place of the Jacobian. A second problem was the handling of surfaces in the update steps. Now, during the first iteration, proxies are created for each surface. During further iterations those surfaces are visited again.
The previous implementation could not update the track state parameters.
The updated Jacobian was overridden by
stepper.transportCovarianceToBound. Therefore, every time the identity matrix was used in place of the Jacobian.A second problem was the handling of surfaces in the update steps. Now, during the first iteration, proxies are created for each surface. During further iterations those surfaces are visited again.