Modify Auto Procedure Calibrator Selection Algorithm

Modification Request #4 (C04 2008)

1. Introduction

The auto procedures AutoPeak, AutoFocus, AutoPeakFocus and AutoOOF find a calibration source that satisfy all user requirements (via parameters to the AutoX astrid procedures) within a specified great-circle angular distance of the search center. The auto procedures currently use an algorithm for selecting calibration sources only checks the distance along a great circle and does not also check what the azimuth and elevation distances to avoid large slews. This may result in unreasonably long slews from the source to the pointing source if the target source is near the zenith. For sources at high zenith, a source may be selected which requires a higher antenna velocity than the maximum allowed. This MR specifies the algorithm that solves both these problems.

2. Background

The present algorithm finds all suitable (satisfying all user requirements on frequency, flux-density limit, etc.) calibrators within a specified great-circle angular distance of the search center. At very high elevations (near the zenith), this can yield calibrators requiring excessively long azimuth slews and/or high tracking rates for scans such as "AutoPeak" Also, if no calibrator met all of the input criteria, no calibrator was returned and AutoPeak halted.

3. Requirements

3.1 General Requirements

3.2 Algorithmic Requirements

always return at least one calibrator. If no calibrator satisfies all of the user requirements, the algorithm will iteratively relax the requirements on flux density and search radius and tries again. A warning will be printed if the resulting calibrator(s) does not meet all of the user requirements. This is new and how Jim Condon wrote the algorithm. Do we really want to do this? We need astronomer input!
Add a new argument to the autoP procedures called "hlst"
- hlst is the local sidereal time in decimal hours which is needed to estimate azimuth slew times and elevations
- If hlst < 0., the new CALFIND still returns all suitable calibrators within the specified great-circle angular distance of the search center, without regard to their current azimuths and elevations. Specifying the current hlst between 0 and 24 hours enables two new features:
  - estimates the azimuth and elevation slew times needed to move from the search center to the calibrator. If the estimated slew time exceeds the larger of (a) one minute and (b) 0.1 minutes multiplied by the search radius in degrees, then the calibrator is rejected.
  - If the current elevation of a calibrator is greater than 88 degrees, the algorithm will reject that calibrator on the grounds that the AutoPeak azimuth tracking rate may be too high for the GBT to measure accurate local pointing corrections. The elevation cutoff of 88 degrees corresponds to a zenith angle of 2 deg ~ 1/30 of a radian and a maximum azimuth rate of ~ 7 deg/min for sidereal tracking.
  - If the current elevation of a calibrator is less than 7 degrees, reject the calibrator.

The default value of hlst ("hlst=None") will mean that the software will use the current LST.
In the case where no calibrator satisfies the criteria, I would suggest that a dialog box appears giving the user 4 choices 1)increase search radius, 2)decrease flux limit, 3)both, 4)bail out.

There should be a version which lists all calibrators that satisfy the criteria; this will be useful for testing versus Condon's program.
- But normally, AutoPeakFocus will use only the calibrator that is closest to the specified location.

4. Design

Translate Jim Condons calfind3.f fortran code into python code in astrid.

The following link gives details of Condon's algorithms in non-Fortran language.
Notes on Condon's algorithms

4.1 OffsetScan

5. Deployment Checklist

A user-level description of the modified calibration selection algorithm is to be included with the Release Notes.
The Auto Procedures documention should be updated where appropriate to indicate the change to the cal source selection algorithm.

6. Test Plan

Testing should have as its primary goal confirmation that the new algorithm successfully selects calibrations sources which do not require velocities greater than allowed and it does not select sources which require slew times greater than than mazimum of (one minute or 0.1 minutes multiplied by the search radius in degrees).

6.1 Internal Testing

Existing unit tests which are directly affected by these changes will be reviewed, and modified if necessary. New Unit tests will be added if they provide substantial benefit. The simulator will be used to test the basic operations of the the Auto procedures. Verification of the actual improvements to the algorithm will be performed by the sponsor.

Written	- MelindaMello - 2008 June 9
Checked
Approved by Sponsor
Approved by CCC
Accepted/Delivered by Sponsor	symbol - name - date

Symbols:

Use %X% if MR is not complete (will display )
Use %Y% if MR is complete (will display )

CCC Discussion Area

-- MelindaMello - 09 Jun 2008

Topic ModificationRequest4C408 . { Edit | Attach | Ref-By | Printable | Diffs | r1.3 | > | r1.2 | > | r1.1 | More }

Revision r1.3 - 02 Jul 2008 - 17:37 GMT - FrankGhigo	Content copyright © 1999-2007 by the contributing authors. All material on this collaboration platform is the property of the contributing authors.