[feat] Allow LeastSquaresSolver to also calculate RMS error by coordinate subset #516
Description
Is your feature request related to a problem? Please describe.
Currently, the least squares solver calculates the combined RMS error of the entire estimation vector, which is simple, but doesn't really have a good physical meaning when there are multiple subsets in the estimation vector
Describe the solution you'd like
Allow the option for the user to have the RMS error (which is used as the convergence check criteria) to be calculated individually per coordinate subset that exists in the estimation vector, and have different convergence RMS thresholds for each subset.