Non-Iterative Geometric Approach to Nonlinear Parametric Least-Squares Adjustment With or Without Constraints.

摘要

A nonlinear parametric adjustment model with n observations and u parameters can be interpreted geometrically as the description of a u-dimensional model surface embedded in an n-dimensional flat observational space. Two geometric concepts have been identified for the least-squares resolution of this model. The first concept is linked to the Gauss form of the surface, which is the most direct representation of the parametric model; however, the solution is iterative in nature. The second concept is linked to the functional form of the surface, and it leads to a non-iterative solution.