 |
|
Publications, papers published in April 1999/communications publiées en avril 1999
- 1
-
J.-F. Magni.
Multimodel eigenstructure assignment in flight-control design.
Aerospace Science and Technology (special issue "Flifht
control law design"), 3(3):141-151, April 1999.
This paper presents a control design approach based on eigenstructure
assignment by dynamic feedback. The proposed method permits the designer to
handle simultaneously robustness against real parameter variations (from a
multimodel point of view) and the use of structured gain including, as a
special case, scheduled gains. Furthermore, given a feedback gain, it is
possible to re-design it by eigenstructure assignment by re-assigning its
leading eigenstructure. This technique has two advantages. First, it can be
viewed as an efficient controller order reduction, second, the design
methodology can be initialized, for example, by H_infinity or mu-synthesis
which makes it possible to combine the advantages of techniques of different
nature.The proposed method alternates multimodel design and mu-analysis but
apart from these natural iterations, only algebraic computation is involved.
- 2
-
C. Döll, G. Ferreres, and J.-F. Magni.
-tools for flight control robustness assessment.
Aerospace Science and Technology, 3(3):177-189, April 1999.
The aim of this paper is to emphasize the usefulness of the
framework for the analysis of the robustness properties of flight control
systems. Model uncertainties may correspond, either to parametric
uncertainties (e.g. in the aerodynamic model or in the value of the trim
point) or to neglected high frequency dynamics (e.g.in the actuators or
sensors). We especially show that several classical problems (computation of
a phase or delay margin, non-linear analysis of a PIO effect, computation of
a frequency domain overshoot)can be extended to an uncertain aircraft model.
A flexible model of the aircraft can be moreover taken into account.
Le but
de cet article est de mettre en valeur la -analyse pour caractériser la
robustesse et les performances des systËmes de pilotage. Les incertitudes de
modèle correspondent, soit aux incertitudes paramètriques (induites par le
modèle aérodynamique ou le point d'équilibre), soit aux dynamiques négligées
aux hautes fréquences dues aux actionneurs ou capteurs). Plus précisément, on
montre que plusieurs problèmes classiques (tels que le calcul d'une marge de
phase ou de retard, l'analyse non-linéaire du phénomène de pompage piloté, le
calcul du dépassement dans le domaine fréquentiel) peuvent être étendus à un
modèle d'avion incertain. En outre, un modèle flexible d'avion peut aussi
pris en compte.
- 3
-
J.-M. Biannic and P. Apkarian.
Missile autopilot design via a modified LPV synthesis technique.
Aerospace science and Technology (special Issue : Flight Control
Law Design), 3(3):153-160, April 1999.
Robust control via Linear Time-Invariant LTI Hinfini or mu techniques
has been extensively studied in the past few years. In this area, the theory
is now quite mature. However, its limitation to linear time-invariant systems
often remains a weakness which may give rise to well-founded criticisms. This
has enforced the need for an extension of standard Hinfini synthesis
techniques to the class of Linear Parameter-Varying (LPV) plants. Many
references are now available in the literature on this topic and several
variants have been proposed to improve the original approach. The major
improvement consists in incorporating knowledge on the rate of variation of
the parameters. This constraint is of high practical interest and is
naturally expected to relax significantly the conservatism of the original
approach. Unfortunately, it often leads to computationnaly expensive
algorithms. In this paper, we propose a very simple approach which consists
in applying the original method to a modified LPV system. Thus, the
attractive numerical properties of the initial algorithm are preserved. In
return, the proposed method suffers from the lack of rigorous theoretical
justification. This remains yet a minor drawback if we consider the practical
interest of the method which is demonstrated on a realistic example, namely
an autopilot design.
|
