https://pfspipe.ipmu.jp/jira This file is an XML representation of an issue en-us 8.3.4 803005 13-09-2019 https://pfspipe.ipmu.jp/jira/browse/INSTRM-1480 Instrument control development <p>In the November run, Robert reported the cobra spots are missed matched in some cobras.  See the following image.</p> <p> </p> <p><span class="image-wrap" style=""><img src="https://pfspipe.ipmu.jp/jira/secure/attachment/14526/14526_image.png" height="517" width="538" style="border: 0px solid black" /></span></p> INSTRM-1480

Fixing the cobra spot matching problem.

Task Normal Done Done karr chyan EngRun Wed, 8 Dec 2021 07:36:04 +0000 Tue, 10 May 2022 19:52:30 +0000 Fri, 22 Apr 2022 02:59:23 +0000 0 4 <p>This should be addressed after convergence  runs have been performed with the full fibre identification code, which can be done on the test stand. </p> <p>I've started testing the code against the motor map data, which provide interesting test cases. The theta runs provide a trivial case, as all the spots should be inside the region only one cobra can  access, while the phi runs gives examples where all the cobras are in the overlap regions. </p> <p>This case tests the closest to the last position algorithm for sorting out ambiguous positions.</p> <p>At the moment, I'm running in offline mode, to make sure the algorithms run as expected, before testing in MHS simulation mode to test integration into the system on the same data. </p> <p> </p> <p>A few minor changes have been pushed to the branch. Currently, we're sorting out some geometry/transformation issues which affect hte fibre identification. </p> <p>Geometry issues have been registered in ticket <a href="https://pfspipe.ipmu.jp/jira/browse/INSTRM-1551" title="Issue PFI-MCS coordinate transformations on bench" class="issue-link" data-issue-key="INSTRM-1551"><del>INSTRM-1551</del></a>; the transformation between MCS and PFI has some instabilities which make the calculation of the patrol region of individual cobras inaccurate. This causes problems when we are dealing with cobras close to the edge of the patrol region.  </p> <p>The identification algorithm has been updated to deal with uncertainties in the transform, and to account for the illuminated but non moving cobras. The updated code is in branch 1480, which needs to be be checked on the real system and merged to master. </p> <p>Can we merge this to master prior to checks on the real system? It would be good not to have too many open ticket branches.</p> <p>I can give it a check to make sure the interface has remained the same, then it merge fairly cleanly. </p> <p>I checked the interface, and made a few modification to McsCmd.py, in particular passing an additional parameter (the fiducial positions) to fibreId and adding a top level routine to switch methods in the fibre ID between target position and previous position modes. </p> <p> </p> <p>rebased and merged to master. Can close now, and open up smaller tickets for updates.</p> <p>merged? Does not look like that has been pushed....</p> <p>Ping on the merge/push? I see the tip at 8a37af5, but a merge whose text says it is a merge of <a href="https://pfspipe.ipmu.jp/jira/browse/INSTRM-1480" title="Fixing the cobra spot matching problem." class="issue-link" data-issue-key="INSTRM-1480"><del>INSTRM-1480</del></a> (2a05099) is not quite that. Very little actual code difference, but the git tree is quite confusing. After a <tt>git rebase master</tt> in the ticket branch, is the matching <tt>git push --force-with-lease</tt> always being done?</p> Relates INSTRM-1559 Development Rank 0|02qpq3:r3s00000004003i Sprint PreEngRun05 D PreEngRun05 E PreEngRun05 F Story Points 3.0