Skip to Main Content
You are now in the main content area

Modeling and Embodied Intelligence for Bio-Inspired Shape-Morphing Soft Robotics

At the ELIXIR Lab, we are pushing the boundaries of field-deployable soft robotics through the development of rigorous modeling tools and embodied intelligence frameworks for bio-inspired, shape-morphing robotic systems. Our work centers on both slender-body and shell-type soft robots—flexible, compliant platforms designed to operate in unstructured, real-world environments.

Unlike conventional rigid robots, soft robots offer intrinsic advantages: they are safer, more adaptable, and capable of operating in constrained or unpredictable settings. These properties make them ideal for high-stakes applications, including search-and-rescue, infrastructure inspection, space exploration, and minimally invasive medicine. Yet, these same advantages introduce substantial challenges in modeling, control, and autonomy.

We believe that true autonomy in soft robotics can only be achieved through embodied intelligence, where cognition emerges from the physical interaction between a robot’s morphology, materials, and sensorimotor dynamics. Toward this goal, we develop sensorimotor feedback loops embedded within soft actuators, leverage distributed learning architectures that integrate local deformation sensing with decentralized control, and build sim-to-real pipelines that allow behaviors learned in high-fidelity simulations to be transferred directly to physical systems. By unifying mechanics, control, and learning, we aim to engineer soft robots that not only move like biological organisms—but adapt, learn, and thrive like them, even in unpredictable and unforgiving environments.

Our Key Contributions:

  • Developing physics-based geometric modeling frameworks for both Cosserat rod-based slender robots and shell-type soft robotic structures, capable of capturing nonlinear deformations, contact interactions, and environmental disturbances with high fidelity.
  • Introducing the first fully untethered electromagnetic soft walking robot, demonstrating the feasibility of autonomous, power-efficient soft mobility.
  • Currently developing a modular, fully soft worm-like robot inspired by peristaltic locomotion, designed for complex, uneven terrains where conventional wheeled or legged systems struggle. This platform enables studies in distributed control and embodied coordination.

Selected Publications

To come