Linear Feedback Controller for Homogeneous Polynomial Systems

A team of researchers from the University of Groningen, led by Shaoxuan Cui, has introduced a new control theory framework for a specific and complex class of nonlinear systems. Their paper, 'Linear Feedback Controller for Homogeneous Polynomial Systems,' tackles the challenge of stabilizing systems whose dynamics are described by homogeneous polynomials that admit an orthogonally decomposable (ODECO) tensor structure. While recent tensor-based methods provided good analysis for open-loop systems, designing effective closed-loop controllers and accurately estimating their stability regions (ROA) remained difficult, often relying on conservative or computationally heavy methods like linearization or sum-of-squares (SOS) programming.

The core innovation is a 'structure-preserving' linear feedback controller that aligns with the ODECO eigenbasis of the system's inherent tensor. This alignment is key, as it allows the researchers to derive closed-form mathematical expressions for system trajectories. This, in turn, enables explicit calculation of convergence and escape thresholds and delivers 'sharp' (non-conservative) characterizations of the closed-loop region-of-attraction. The paper also extends the theory to provide robustness guarantees against bounded disturbances. The proposed method represents a significant theoretical advance, offering a more elegant and computationally tractable alternative to existing techniques for this niche but important class of systems in control engineering.