On the Fundamental Properties of Linear Parameter-Varying Dynamic Systems Under Parametrical Multi-Perturbations. Applications to Time-Delay Systems. Preliminary Results

TL;DR

This paper presents a unified approach to determine admissible parametric perturbations in linear time-varying systems that preserve fundamental properties like controllability and observability, with applications to delay systems.

Contribution

It introduces a method to compute perturbation bounds ensuring system properties are maintained, extending to systems with delays and complex parameter variations.

Findings

Perturbation radii are calculated simply.

Applicable to systems with multiple delays and parametric uncertainties.

Maintains properties like controllability and observability under perturbations.

Abstract

This paper deals with a unifying approach to the problems of computing the admissible sets of parametrical multi perturbations in appropriate bounded sets such that some fundamental properties of parameter-varying linear dynamic systems are maintained provided that the so-called (i.e. perturbation-free) nominal system possesses such properties. The sets of parametrical multi perturbations include any combinations of parametrical multi perturbations in the matrix of dynamics as well as in the control, output and input-output interconnection matrices which belong to some prescribed bounded domain in the complex space. The various properties which are investigated are controllability, observability, output controllability and existence of minimal state-space realizations together with the associate existence or not of associate decoupling, transmission and invariant zeros. All the matrices of parameters including the nominal and the disturbed ones which parameterize the dynamic system may be real or complex. The radii of the multi- parametrical perturbations are calculated in a simple way. The obtained results are then applied to systems subject to a finite number of point internal delays and parametrical multi perturbations by comparing the state- space descriptions of such systems with the general descriptions previously investigated. In particular, the contributions of the delays to the spectral descriptions are assimilated to the contributions of a set of varying parameters in a domain for the general description.