Collaborative Research: MORPHOSIS: Modeling and cOntrol for Robotic and Programmable High-dimensiOnal Shape-morphIng Structures

Colorado State University, Fort Collins CO

Investigators

Abstract

This project supports fundamental research on shape-morphing structures that looks to dynamically transform their physical shapes into desired configurations in two or three dimensions. Such morphing structures can potentially revolutionize various fields by enabling materials, systems, or devices that actively adapt their form to suit different needs. For example, such structures could lead to new materials that adjust their stiffness in response to changing demands, robotic systems that reconfigure themselves to move through complex environments, or wearable devices that alter their shape for improved fit and comfort. By advancing the scientific understanding and engineering capabilities of such systems, this project directly promotes the progress of engineering science and supports national interests in health, security, and manufacturing. The research activities will also provide educational opportunities for undergraduate students, enhance engineering curricula, and inspire the next generation of scientists and engineers through outreach to K-12 students. The technical focus of the project is to develop a rigorous framework for modeling, planning, and controlling high-dimensional morphing structures composed of interconnected morphing rods. Each morphing rod combines a thermally driven artificial muscle with a variable-stiffness shape memory polymer to enable large, reversible deformations. The project will begin by designing modular rod geometries and mechanical connectors that enable flexible and reconfigurable assemblies. It will then formulate physics-based models that integrate reduced-order rod mechanics, artificial muscle dynamics, and connector constraints. To enable scalable and robust control, the project looks to develop data-driven models based on Koopman operator theory, enhanced by deep learning to automatically discover system observables. These models will be made robust by incorporating uncertainty in the learned representations and scalable through the use of graph neural networks that capture the connectivity of complex structures. The project seeks to then leverage these models to enable real-time control and optimal planning, determining which elements should be actuated or softened to achieve desired shapes. The integration of model-based design, data-driven modeling, and real-time control seeks to establish a new paradigm for shape-morphing systems, enabling them to operate with precision, versatility, and autonomy in a wide range of applications. This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.