Reduction principles and the stabilization of closed sets for passive systems

Authors: Mohamed I. El-Hawwary, Manfredi Maggiore · Year: 2009 · Venue: arXiv:0907.0686 (extended version, 25 pp.); published in condensed form as IEEE Trans. Autom. Control 55:982–987, 2010, DOI 10.1109/TAC.2010.2041684 Raw: md · pdf

Summary

Solves the set stabilization problem for passive systems: given a closed (not necessarily compact) set that is open-loop invariant and contained in the zero level set of the storage function, when does a passivity-based feedback make stable, attractive, or asymptotically stable for the closed loop? The engine is a pair of novel reduction principles — for semi-attractivity (Theorem 3.1) and semi-asymptotic stability (Theorem 3.2) — extending Seibert–Florio’s compact-set reduction principles (seibert1995reduction) to closed, possibly unbounded sets, with a cascade stability criterion as a corollary (Corollary 3.3). The passivity results generalize Byrnes–Isidori–Willems’ equilibrium theory and Shiriaev–Fradkov’s compact-set results, via a -detectability notion (sufficient conditions in Proposition 4.12; asymptotic stabilization in Theorem 4.14). The corpus copy is the 25-page arXiv version whose Appendices A–D carry the full proofs; the TAC version is the 6-page condensed note.

Key Claims

  • Reduction principle for semi-attractivity (Theorem 3.1). For nested closed invariant sets , attractivity of relative to plus the stated conditions lifts to semi-attractivity in the full state space; conditions (ii)/(ii’) are also necessary. Proved in Appendix C.
  • Reduction principle for semi-asymptotic stability (Theorem 3.2). The asymptotic-stability analogue, extending Seibert–Florio’s Theorem 4.13 / Corollary 4.11 to closed non-compact sets.
  • Lemma 2.5 (uniform semi-attractivity vs semi-asymptotic stability). A uniform semi-attractor [relative to ] is semi-asymptotically stable [relative to ]; if the flow is locally uniformly bounded near , the two are equivalent. Proved in Appendix A; used in Section 4.2 and Appendix D.
  • Passivity-based set stabilization (Theorem 4.14). Conditions for a passivity-based feedback to asymptotically stabilize , with -detectability replacing positive-definiteness of the storage function — novel even in the equilibrium case.

Method

Control-affine passive systems , on ; stability of sets phrased via the flow , Birkhoff limit sets and Ura prolongational limit sets; semi-definitions (semi-attractor, semi-asymptotic stability) handle non-compact .

Regime note. A general nonlinear-systems / set-stability paper — no spacecraft or manipulator model — so the free-flying vs free-floating distinction does not arise. Relevance is as a tool for certifying stability of cascade / hierarchical designs.

Relevance to thesis

Sibling and precursor to elhawwary2013reduction: the 2013 Automatica paper reuses this machinery for reduction-based backstepping, while this paper carries the passivity-side results and the full proofs. Filed to support the Lean formalization work (ctrllib), which cites Lemma 2.5 and its Appendix A proof as the ground truth for the semi-attractivity/semi-asymptotic-stability equivalence. The reduction principles are candidate tools for certifying the hierarchical closed loop of the free-flying space manipulator as stability of a set, complementing the cascade results of panteley2001growth.

Connections

Topics: Cascaded Systems · Lyapunov Stability Sources: elhawwary2013reduction · seibert1995reduction · khalil2002nonlinear

Key Equations / Quotes

From the abstract:

“In this paper we explore the stabilization of closed invariant sets for passive systems, and present conditions under which a passivity-based feedback asymptotically stabilizes the goal set. Our results rely on novel reduction principles allowing one to extrapolate the properties of stability, attractivity, and asymptotic stability of a dynamical system from analogous properties of the system on an invariant subset of the state space.”

Lemma 2.5 (verbatim, arXiv v1):

“Let be a closed set which is positively invariant for in (4), and let be a closed set. If is a uniform semi-attractor [relative to ], then it is semi-asymptotically stable [relative to ]. Furthermore, if is locally uniformly bounded near , then is semi-asymptotically stable [relative to ] if, and only if, it is a uniform semi-attractor [relative to ].”

Open Questions

  • Which of the semi-definitions (semi-attractor, semi-asymptotic stability) does the Lean formalization adopt verbatim, and where do they diverge from Mathlib’s existing stability vocabulary?