← Back to list of papers of the 2019 EuroGNC conference

CEAS EuroGNC 2019

A Distributed Robust Optimal Control Framework Based on Polynomial Chaos
Patrick Piprek Technical University of Munich, Institute of Flight System Dynamics, Munich, Germany
Sébastien Gros Norwegian University of Science and Technology, Department of Engineering Cybernetics, Trondheim, Norway
Florian Holzapfel Technical University of Munich, Institute of Flight System Dynamics, Munich, Germany
Abstract:
This study is concerned with the development of a robust open-loop optimal control (ROC) framework that distributes different generalized polynomial chaos (gPC) sub-problems from the non-intrusive stochastic collocation (SC) method. This distributed open-loop optimal control (DOC) approach yields a number of smaller open-loop optimal control problems (OCPs) that can be solved independently of each other and are only connected by a small number of connection variables. These connection variables are introduced based on the specifics of the used cost and constraint functions and describe the coupling in the gPC expansion when e.g., calculating the variance. Overall, the definition as a DOC problem yields a faster and more reliable way to solve the ROC problem than by a full, connected problem. Here, the study shows the applicability of the proposed method in an air race example with the optimization of mean values and variances.
Keywords: Optimal control; Optimization; Robust control
 View PDF CEAS-GNC-2019-031


Patrick Piprek, Sébastien Gros, Florian Holzapfel: A Distributed Robust Optimal Control Framework Based on Polynomial Chaos. Proceedings of the 2019 CEAS EuroGNC conference. Milan, Italy. April 2019. CEAS-GNC-2019-031.
BibTeX entry:

@Incollection{CEAS-GNC-2019-031,
    authors = {Piprek, Patrick and Gros, Sébastien and Holzapfel, Florian},
    title = {A Distributed Robust Optimal Control Framework Based on Polynomial Chaos},
    booktitle = {Proceedings of the 2019 {CEAS EuroGNC} conference},
    address = {Milan, Italy},
    month = apr,
    year = {2019},
    note = {CEAS-GNC-2019-031}
}