Details

    • Sprint:
      Active Development

      Description

      The value of crEpsilon was always 0.0 inside SolarRadiationPressure::GetDerivatives(). crEpsilon is used in calculating the partials for the A matrix. I believe that the value of crEpsilon should match the value of the epsilon in the BatchEstimator.

        Gliffy Diagrams

          Attachments

            Activity

            Hide
            dcooley Steve Cooley added a comment - - edited

            JM: Just confirming that this ticket should still be open?

            • Discussion of GMT-6716, Error in Cr calculation
            o John to review Tuan’s code changes.
            o Does it affect TSF and Cd solve-for code also? Maybe, John will check.
            o John to look at sample Cr estimation cases to see how this change affects estimation.
             Use simulate case
             Use SOHO case in “Do Not Distribute” folder
             Do Not merge to Master yet
             Work with Tuan to write up Math specs for this. See also my math specs in Repo.

            Show
            dcooley Steve Cooley added a comment - - edited JM: Just confirming that this ticket should still be open? • Discussion of GMT-6716 , Error in Cr calculation o John to review Tuan’s code changes. o Does it affect TSF and Cd solve-for code also? Maybe, John will check. o John to look at sample Cr estimation cases to see how this change affects estimation.  Use simulate case  Use SOHO case in “Do Not Distribute” folder  Do Not merge to Master yet  Work with Tuan to write up Math specs for this. See also my math specs in Repo.
            Hide
            jmcgreevy John McGreevy added a comment -

            The description above is from my first review of Tuan's Cr calculation changes. I'll review the revisions Tuan makes to stop crEpsilon always having a value of 0.0 once those are ready.
            I know it also affects Cd, but I didn't check the TSF solve for.

            Show
            jmcgreevy John McGreevy added a comment - The description above is from my first review of Tuan's Cr calculation changes. I'll review the revisions Tuan makes to stop crEpsilon always having a value of 0.0 once those are ready. I know it also affects Cd, but I didn't check the TSF solve for.
            Hide
            dcooley Steve Cooley added a comment -

            Tuan to write up math description of what need to be changed.

            Show
            dcooley Steve Cooley added a comment - Tuan to write up math description of what need to be changed.
            Hide
            tdnguyen Tuan Nguyen added a comment -

            Code to fix this bug was uploaded to CrCdSolvefor branch. I ran tests for the following cases:

            RunDef.Folders =

            { 'Navigation' 'FRR-22_BatchEstimator' 'FRR-23_KalmanFilter' 'FRR-24_MeasurementSimulator' 'FRR-46_TrackingFileSet' 'FRR-56_ThrustHistoryFile' 'FRR-59_SpacecraftNavigation' 'FRR-57_ThrustSegment' 'FRR-58_Receiver' 'FRC-20_RunEstimator' 'FRC-21_RunSimulator' }

            ;
            against all measurement. It passed all those test cases.

            Show
            tdnguyen Tuan Nguyen added a comment - Code to fix this bug was uploaded to CrCdSolvefor branch. I ran tests for the following cases: RunDef.Folders = { 'Navigation' 'FRR-22_BatchEstimator' 'FRR-23_KalmanFilter' 'FRR-24_MeasurementSimulator' 'FRR-46_TrackingFileSet' 'FRR-56_ThrustHistoryFile' 'FRR-59_SpacecraftNavigation' 'FRR-57_ThrustSegment' 'FRR-58_Receiver' 'FRC-20_RunEstimator' 'FRC-21_RunSimulator' } ; against all measurement. It passed all those test cases.
            Hide
            tdnguyen Tuan Nguyen added a comment -

            It needs to regression test and verify test result before it merges to master branch.

            Show
            tdnguyen Tuan Nguyen added a comment - It needs to regression test and verify test result before it merges to master branch.
            Hide
            dcooley Steve Cooley added a comment -

            John M to do thorough code review (Cd, Cr, TSF) as it relates to this "epsilon" issue.

            Show
            dcooley Steve Cooley added a comment - John M to do thorough code review (Cd, Cr, TSF) as it relates to this "epsilon" issue.

              People

              • Votes:
                0 Vote for this issue
                Watchers:
                4 Start watching this issue

                Dates

                • Created:
                  Updated: