A computationally reduced version of the fast H∞ filter. (June 2018)

• Record Type: Journal Article
• Authors: Nishiyama, Kiyoshi
• Is Part Of: Automatica. Volume 92(2018)
• Page Start: 249
• Page End: 253
• Publication Date: 2018-06
• Subjects: H-infinity filter -- Fast algorithms -- Adaptive filters -- RLS algorithm -- Forgetting factor; Automatic control -- Periodicals; Automation -- Periodicals; 629.805
• DOI: 10.1016/j.automatica.2018.02.009
• Languages: English
• ISSNs: 0005-1098

Abstract

Abstract: The fast H ∞ filter is designed to overcome the performance limitations of the fast Kalman filter and the fast transversal filter used for system identification; however, its computational complexity exceeds O ( 7 N ), the complexity of the original fast transversal filter. In this work, a computationally reduced version of the fast H ∞ filter is derived as an a posteriori error formulation of the fast H ∞ filter using a conversion factor. The conversion factor-based fast H ∞ filter provides the same performance as the fast H ∞ filter at the expense of a computational cost comparable to that of the fast transversal filter.