Invariance of subspace based estimators

Subspace-based estimates, i.e., estimates obtained by exploiting the orthog onality between a sample subspace and a parameter dependent subspace, have proved useful in many applications, including array processing and system i dentification. The purpose of this paper is to complement the already avail able theoretical results generally obtained in specific contexts. We discuss the generalization of the optimal weighted subspace fitting appr oach introduced by Viberg in the DOA estimation context; we exhibit some in variance properties of optimally weighted estimate, and me show the equival ence between subspace fitting and subspace matching.