Wake stabilization with Machine Learning Control (xMLC) - Laboratoire Interdisciplinaire des Sciences du Numérique Access content directly
Conference Poster Year :

Wake stabilization with Machine Learning Control (xMLC)


Controlling a dynamical system is to give an input that will change the system towards a desired state, assuring different constraints such as noise rejection, uncertainty mitigation and robustness. To this extent, two approaches oppose each other : model-based control (linear control theory, ERA/OKID, etc), pioneered by Norbert Wiener (1894-1964) and machine learning control. Model-based control victories are mainly in opposition and phasor control, and examples including MIMO control and frequency crosstalk are dim. To override the intrinsic limitations of model-based control, a new approach is considered building on recent innovations in artificial intelligence : machine learning control. Machine Learning Control (MLC, [4]) is a model-free control method based on genetic programming and building on the pioneering work of Dracopoulous [5]. It's a biologicallyinspired method that mimics the Darwinian evolution to build fitter controllers, in the same way the bald eagle evolved through millions of years to be able to fly under gusty conditions. The cornerstone is the formulation of the control problem as a function optimization problem; such non-convex problem can then be solved thanks to MLC. MLC strength relies on its ability to build nonlinear control laws reproducing known control methods including model-based (ERA/OKID), open-loop strategies (multi-frequency forcing) and closed-loop strategies (phasor control, ARMAX) and also linear and nonlinear combinations of them.
Fichier principal
Vignette du fichier
2020VKI_CornejoMaceda_Poster.pdf.pdf (1.98 Mo) Télécharger le fichier
Origin : Files produced by the author(s)

Dates and versions

hal-03874030 , version 1 (27-11-2022)




  • HAL Id : hal-03874030 , version 1


Guy Y. Cornejo Maceda, Bernd R Noack, François Lusseyran, Marek Morzyński. Wake stabilization with Machine Learning Control (xMLC). von Karman Institute Lecture Series, Feb 2020, Rhode-Saint-Genèse, Belgium. ⟨hal-03874030⟩
11 View
5 Download


Gmail Facebook Twitter LinkedIn More