Title :
Digitizing method with application to the design of fractional-time prediction models
Author :
Braileanu, Grigore I.
Author_Institution :
Electr. & Comput. Eng. Dept., Gonzaga Univ., Spokane, WA, USA
Abstract :
An interpolation based digitizing method has been previously developed especially for the design of narrowband frequency selective filters. This method performs a trigonometric polynomial interpolation along an extended time window that ends at the current sampling time. The present paper reconsiders the basic extended window design (EWD) idea in the light of digital control systems requirements and proposes a generalized EWD (G-EWD) method. The main results of this paper are: (i) the closed-form G-EWD expressions derived for an interpolation space spanned by any set of linearly independent functions of time, (ii) a computationally efficient linear algebra generation of the G-EWD digital equivalents, and (iii) the derivation of optional prediction models whose output values are computed at fractional time instants beyond the current sampling time. Finally, the paper compares the proposed G-EWD method to an optimal frequency-domain digitizing method.
Keywords :
digital control; discrete time systems; interpolation; linear algebra; polynomial approximation; digital control systems requirements; digitizing method; extended time window; extended window design; fractional time instants; fractional-time prediction model design; fractional-time prediction models; interpolation based digitizing method; interpolation space; linear algebra generation; linearly independent functions; narrowband frequency selective filters; optimal frequency-domain digitizing method; optional prediction models; trigonometric polynomial interpolation; Control systems; Delay; Digital control; Digital filters; Interpolation; Narrowband; Polynomials; Predictive models; Sampling methods; Transfer functions;
Conference_Titel :
Decision and Control, 2004. CDC. 43rd IEEE Conference on
Print_ISBN :
0-7803-8682-5
DOI :
10.1109/CDC.2004.1428744