Realizations for fixed-point implementation of controllers and filters with peak-to-peak scaling

Author

Viassolo, Daniel E.

Author_Institution

Xerox Corp., El Segundo, CA, USA

Volume

1

fYear

2003

fDate

4-6 June 2003

Firstpage

83

Abstract

The problem of implementing a given linear time-invariant controller or filter in a finite wordlength processor is studied. We provide methods for finding state-space realizations that: (1) minimize the impact of quantization errors on performance, and (2) prevent the peak values of the controller states from exceeding processor ranges. These methods are illustrated through numerical examples.

Keywords

controllers; digital control; digital filters; fixed point arithmetic; quantisation (signal); state-space methods; controller states; digital controllers; digital signal processors; exceeding processor ranges; finite wordlength processor; fixed point implementation; linear time invariant controller; linear time invariant filter; numerical examples; optimal realization problem; peak to peak scaling; quantization errors; state space realizations; Arithmetic; Constraint optimization; Degradation; Digital control; Digital filters; Digital signal processing; Digital signal processors; Hydrogen; Nonlinear filters; Quantization;

fLanguage

English

Publisher

ieee

Conference_Titel

American Control Conference, 2003. Proceedings of the 2003

ISSN

0743-1619

Print_ISBN

0-7803-7896-2

Type

conf

DOI

10.1109/ACC.2003.1238918

Filename

1238918