NavList:
A Community Devoted to the Preservation and Practice of Celestial Navigation and Other Methods of Traditional Wayfinding
From: Robin Stuart
Date: 2015 Mar 2, 17:51 -0800
Antoine (Kermit) has commented on the possible dangers of applying 3 parameter least squares fits to finding an estimated position in g30478 . As further input into the discussion I have carried out a virtual "double blind" experiment alluded to in g30382 .
Greg Rudzinski made 16 Sun sights over the course of a day ( g30368 ). Suppose that before doing those sights he had adjusted his sextant to introduce a certain index error (I.E.). Upon taking the sights the azimuths would be the unchanged but all his intercepts would be increased by I.E. If he used this set of observations to fix his position by means of 2 and 3 parameter fits ( g30219 ) how far away from his actual location would this be?
The attached plot provides the answer. The horizonal axis gives the I.E. that was introduced. Note that the 3 parameter fit is unaffected by the presence of a systematic offset and stays constant at 1.6nm. For this set of observations when the I.E. is below about 1' then the 2 parameter fit gives superior results but beyond that range the 3 parameter fit wins,
Robin Stuart
