Euclidean distances and least squares problems for a given set of vectors Original Research Article

Given an m×nm×n matrix A, n Euclidean distances, those from each column to the space spanned by the remaining columns of A, are considered. An elegant relationship between A, these Euclidean distances, and the solutions of n simple linear least squares problems arising from A is derived. When A has full column rank, from this a useful expression for these Euclidean distances is immediately obtained. The theory is then used to greatly improve the efficiency of an algorithm used in communications.