LArSoft  v09_90_00
Liquid Argon Software toolkit - https://larsoft.org/
KHitMulti.h File Reference

Compound Kalman Filter measurement. More...

Go to the source code of this file.

Classes

class  trkf::KHitMulti
 

Namespaces

 trkf
 

Detailed Description

Compound Kalman Filter measurement.

Author
H. Greenlee

This class allows a collection of Kalman Filter measurements to function as a single measurement. This class inherits from KHitBase and fulfills the KHitBase interface. The following attributes are owned by base class KHitBase.

  1. Measurement surface.
  2. Prediction surface.

This class adds the following attributes of its own.

  1. A collection of underlying measurements.
  2. Measurement vector.
  3. Measurement error matrix.
  4. Prediction vector.
  5. Prediction error matrix.
  6. Residual vector.
  7. Residual error matrix.
  8. Inverse of residual error matrix.
  9. Kalman H-matrix.
  10. Incremental chisquare.

The only supported kind of underlying measurement is KHit<1> (more types could be added if needed). The dimension of the measurement space is dynamic (potentially large) and specified at run time.

The measurement vector, measurement error matrix, prediction vector, and H-matrix are simply the concatenation of the corressponding quantities from the underlying measurements. The measurement error matrix is block-diagonal with no correlation between the underlying measurements. The prediction error matrix is calculated from the error matrix of the track hypothesis and the full H-matrix (there will be correlations between measurements). Residuals and chisquare are calculated in the usual way.

Definition in file KHitMulti.h.