Asymptotical Orthonormalization of Subspace Matrices Without Square Root
- Author(s): Hua, Yingbo
- et al.
Subspace computation is fundamental for many signal processing applications. A well-known tool for computing the principal subspace of a data matrix is the power method. During the iterations of the power method, a proper normalization is essential to avoid numerical overflow or underflow. Normalization is also needed to achieve desirable properties such as orthonormalized subspacematrices. A number of normalization techniques for the power method is reviewed, which include the conventional as well as nonconventional ones. In particular, a new method of normalization is introduced to achieve asymptotical orthonormalization of subspace matrices without the use of squareroot. This method is among a class of normalization methods that allow a simple adaptive implementation of the power method for subspace tracking.
Many UC-authored scholarly publications are freely available on this site because of the UC Academic Senate's Open Access Policy. Let us know how this access is important for you.