Data-Driven Spectral Submanifold Reduction for Nonlinear Optimal Control of Soft Robots

Modeling and control of continuum soft robots remains a challenging task due to their inherent nonlinearities and high degrees of freedom. These complexities hinder the construction of high-fidelity models appropriate for real-time control. While various model and learning-based approaches have been proposed to address these challenges, they lack generalizability and rarely preserve the structure of the dynamics. In this work, we propose a new, data-driven approach for extracting control-oriented models on invariant manifolds from data. We overcome the issues outlined above and demonstrate our superior performance of Spectral Submanifold Reduction (SSMR) vis-\'a-vis the state of the art.