Minutes of the E614 Software Meeting

                            6 March 2000


 Discussion of Kalman filtering :

       At first glance it appears that Kalman filtering can be applied 
       straightforwardly to our track finding/fitting routines.  There
       may be problems due to differing helix representations.

       We will need at least one transport matrix (5 x 5) to propagate
       the track from UV pair to UV pair.  An inhomogeneous B field can 
       be handled as long as it can represented as uniform between chamber 
       pairs.  Another matrix will be needed to describe multiple 
       scattering.

       The efficiency of the filtering process will depend critically on
       the setup of the code, so care is required.  Art, Bill, Farhana
       and Glen have volunteered to work on this problem.

       It is not clear whether ours is a linear or non-linear fitting 
       problem.  If non-linear then the matrices to be inverted may be
       large.

       How does Kalman filtering deal with our left-right hit ambiguity?
       For large angle tracks this will not be a problem.

 Other points:

       We need to start thinking about calibration code and analysis 
       procedures as the chambers will be running within the year.

       The latest CVS MOFIA release will not compile.  It is not just a 
       LINUX problem.

Created for the The Center for Subatomic Research E614 Project Projects Page.

Created by The CoCoBoard.