Granular Feedback Linearization for Multivariable Systems with Guaranteed Stability
Marcela O. Coelho, Lucas Silva de Oliveira, Valter J. S. Leite · 2025
Evolving Granular Feedback Linearization (EGFL) is a flexible and adaptive control strategy designed to manage nonlinear, multivariable, and uncertain systems while achieving desired performance. This article provides an overview of the key aspects of this technique and introduces a modification to its structure aimed at ensuring stability, even under challenging learning conditions. The proposed strategy for maintaining stability involves limiting the compensation action from the EGFL. Stability conditions for the new approach are developed using Lyapunov theory, resulting in convex conditions in terms of linear matrix inequalities. The effectiveness of the modified topology is evaluated through a case study involving helicopter dynamics. The results demonstrate that the proposed approach can maintain system stability, even in challenging learning scenarios.