On the formulation and solution of robust performance problems

Alexander Lanzon, Michael Cantoni

    Research output: Contribution to journalArticlepeer-review

    Abstract

    A new approach to robust performance problems is proposed in this paper. The approach involves the optimisation of the so-called performance weights subject to a constraint, formulated in terms of the structured singular value, which ensures the existence of a stabilising feedback compensator that achieves robust performance with respect to the optimised performance weights and an uncertain plant set. Optimising over the performance weights in this way gives rise to an algorithm for systematically trading-off desired performance against specified plant uncertainty and performance limitations due to plant dynamics. The algorithm also yields a robust controller. The designer is only required to specify the uncertain plant set and an optimisation directionality, which is used to reflect desired closed-loop performance over frequency in terms of a corresponding cost function. Design of this directionality appears to be simpler than designing the performance weights directly. © 2003 Elsevier Ltd. All rights reserved.
    Original languageEnglish
    Pages (from-to)1707-1720
    Number of pages13
    JournalAutomatica
    Volume39
    Issue number10
    DOIs
    Publication statusPublished - Oct 2003

    Keywords

    • μ-Synthesis
    • ℋ∞-Control
    • Optimising performance
    • Performance weight synthesis
    • Skewed-μ

    Fingerprint

    Dive into the research topics of 'On the formulation and solution of robust performance problems'. Together they form a unique fingerprint.

    Cite this