no. 16 · Tuesday 17 November 2026
Contraction Analysis and Control of Polynomial Dynamical Systems
Abstract
Contraction analysis has found broad applications in synchronization, tracking, observer design, and nonlinear feedback control. However, establishing contraction conditions for polynomial dynamical systems remains challenging due to their state-dependent Jacobians. In this article, we develop a tensor-based framework for contraction analysis of polynomial dynamical systems by exploiting the structure of orthogonally decomposable (odeco) coefficient tensors. By transforming the dynamics into modal coordinates, we derive necessary and sufficient semicontraction conditions and characterize state-dependent semimetrics. We establish fundamental limitations on exponential contraction and prove polynomial decay of differential displacements on forward-invariant regions. To accommodate more general polynomial dynamical systems, we extend the framework to include linear dynamics and non-odeco tensor perturbations, deriving computationally tractable contraction certificates with explicit robustness guarantees. We further develop convex feedback synthesis conditions based on linear matrix inequalities for general input matrices. We demonstrate the effectiveness and computational scalability of the proposed framework through numerical experiments on both synthetic systems and nonlinear elastic relaxation models.
- comments
- 16 pages, 4 figures
- fields
- Optimization and Control