Skip to Main content Skip to Navigation
New interface

Identification des systèmes dynamiques multi-variables

Abstract : The aim of this thesis is the study and the development of identification methods for multi-variable dynamic systems. In the first chapter, we are interested in the cases of multi-input/multi-output, dynamic, discreet and invariant systems and those presented under the form of input-output relation. In this context, we have developed two methods : MCA (Moindres Carrés Alternés - Alternate Least Square). The first method is an iterative method comprising of several stages. The second and most general case is a recurrent mehtod which identifies the interconnected multi-variable systems. The advantages of both the methods proposed compared to those found in literature are discussed in order to test the robustness towards diffrent distirbances. In practice, the invariance hypothesis is not always verified. In order to overcome this problem in the case of linear systems with time-varing parameters, we have extended the MCMV method by using the multi-model technique. In the second chapter, we are interested in discreet linear systems represented by a state space model. Firstly, we develope a method which id based on the minimisation of the structure length ; secondly, we develope a new method which is based on the multi-model technique which allows us to follow the parameter variations ; and finally, in the third chapter, we present an extension of both new methods, MCA and MCMV, in the case of a Hammerstein model in which the non-linear part is appoximated by a polynomial function.
Document type :
Complete list of metadata
Contributor : Thèses UL Connect in order to contact the contributor
Submitted on : Wednesday, June 20, 2018 - 11:45:53 AM
Last modification on : Thursday, March 10, 2022 - 9:01:18 AM
Long-term archiving on: : Tuesday, September 25, 2018 - 7:48:27 AM


Files produced by the author(s)


  • HAL Id : tel-01748608, version 1



Mustapha Ouladsine. Identification des systèmes dynamiques multi-variables. Automatique / Robotique. Université Henri Poincaré - Nancy 1, 1993. Français. ⟨NNT : 1993NAN10209⟩. ⟨tel-01748608⟩



Record views


Files downloads