Author:
Philippe Feyel
Affiliation:
Safran Electronics & Defense, Massy and France
Keyword(s):
LPV/LFT Synthesis, H∞ Synthesis, μ Synthesis, Fixed-Structure, Multiple Plants, Small Gain Theorem, Evolutionary Computation, Robust Control, Adaptive Control.
Related
Ontology
Subjects/Areas/Topics:
Adaptive Signal Processing and Control
;
Evolutionary Computation and Control
;
Informatics in Control, Automation and Robotics
;
Intelligent Control Systems and Optimization
;
Nonlinear Signals and Systems
;
Robotics and Automation
;
Signal Processing, Sensors, Systems Modeling and Control
Abstract:
This paper proposes to address the problem of fixed-structure gain-scheduled LPV/LFT controllers for plants with time-varying measurable and time invariant unmeasurable uncertainties. Due to the complexity of merging μ-technics with LPV/LFT approach, an alternative presented here consists in computing robust fixed-structure LPV/LFT controllers using the multiple plants framework instead of μ-technics. The complexity of this optimization problem is tackled with global evolutionary optimization. This paper shows that this approach is quite efficient and very simple to implement. The algorithm has been tested on the pendulum in the cart academic example.