Title :
Robust hyperbolic square root updating in RLS estimation
Author :
Bojanczyk, Adam W. ; Steinhardt, Allan O. ; So, Sing H.
Author_Institution :
Dept. of Electr. Eng., Cornell Univ., Ithaca, NY, USA
Abstract :
A stabilized version of the hyperbolic Householder transform method for the multiple vector downdating of Cholesky factors has been developed. This approach offers roughly a twofold savings over schemes based on the multiple application of single-vector downdating schemes, and it is just as stable. The proposed approach is found to exhibit numerical behavior superior to the standard scheme in simulations. It is concluded that there is no computational penalty incurred with stabilized hyperbolic Householder transforms; they enjoy an operation count identical to their conventional counterparts
Keywords :
least squares approximations; signal processing; transforms; Cholesky factors; RLS estimation; hyperbolic Householder transform method; hyperbolic square root updating; multiple vector downdating; signal processing; simulations; stabilised transforms; Covariance matrix; Equations; Filtering; Least squares approximation; Least squares methods; Resonance light scattering; Robustness; Signal processing algorithms; Transforms; Vectors;
Conference_Titel :
Spectrum Estimation and Modeling, 1988., Fourth Annual ASSP Workshop on
Conference_Location :
Minneapolis, MN
DOI :
10.1109/SPECT.1988.206210
