A least squares adjustment for long baseline interferometry
Abstract
Long baseline interferometry software and data, developed by the Canadian L.B.I. group at York University has been combined with a least squares adjustment package. The options have been implemented to accept an input of both weighted parameters and functional parameter constraints. The results are then analysed statistically, including a chi-square goodness-of-fit test in residuals, a rejection criteria for residual outliers, and chi-square test on the variance factor.
The package has been developed with closer regard to computer economy. Computer storage space has been reduced by 60% and processing time has been reduced by 96% compared with the previously used maximum likelihood adjustment routines. This increase in efficiency has resulted in an ability to input a larger number of observations and, accordingly, in an improvement in accuracy.