International Journal of applied mathematics and computer science

online read us now

Paper details

Number 1 - March 1997
Volume 7 - 1997

Adaptive discrete-time identification of continuous-time systems using adjusted integration

Zdzisław Kowalczuk

Abstract
A new approach to the design of on-line discrete-time systems of identification of continuous-time systems is presented. The problem concerns digital signal processing suitable for generation of the regression vector in discrete-time recursive estimation of continuous-time process parameters based on a mixed, discrete-continuous regression model. The approach is related to discrete approximation of continuous-time systems, where a kind of matching to the input and output signals is taken into account. In particular, the normal integrating operator technique is utilised. Two methods of "tempering" the characteristics of discrete-time integrators are proposed. One method consists in the transformation of an original system of matched IIR filters into an equivalent system of FIR filters. The purpose of the other method is simplification of calculations by applying stabilised closed IIR forms of the integrating operators. Such matched-and-tempered normal integrating operators are referred to as the adjusted integrators. Experimental results, obtained in an analogue and digital simulation environment, of the application of the two methods to discrete-time parameter estimation of nonstationary continuous-time systems corrupted by different noise processes illustrate the usefulness of the approach.

Keywords
-