The fiber-reinforced pneumatic twisted-and-coiled actuator (FR-PTCA) is a novel soft actuator that overcomes key limitations of existing designs such as the Cavatappi muscle. By incorporating a fiber reinforcement to enhance tube anisotropy, the FR-PTCA achieves significantly improved performance. It delivers higher energy efficiency (over 19%), lower required input pressures (130 psi), and maintains a large contraction strain (up to 70%). In contrast, the Cavatappi muscle, while offering desirable attributes such as compliance, low cost, and large linear contraction with minimal radial expansion, suffers from low energy efficiency (~9%) and demands high-pressure inputs (over 240 psi).

Twisted and Coiled Actuators (TCA) are lightweight, low-cost artificial muscles made by twisting and coiling polymer fibers. When electrically heated, they contract along their length, mimicking natural muscle motion. Their high power-to-weight ratio and flexibility make them ideal for soft robotics. My research focuses on designing free-stroke TCAs that achieve large, untethered contractions and incorporate self-sensing for closed-loop control without external sensors. I apply these actuators in diverse robotic systems, including soft robotic arms, bistable jumping robots, and morphing structures. Additionally, I develop models to capture the mechanical and thermal behavior of TCAs and TCA-driven robots, enabling precise control and adaptive functionality in soft robotic applications.

Back support devices (BSDs) have the potential to mitigate overexertion in industrial tasks and also to provide assistance to people with weak back muscle strength in daily activity. While state-of-the-art active BSDs can offer a high assistive force, they are bulky and heavy, making them uncomfortable for daily use. On the contrary, passive BSDs are compact but need manual adjustment to be versatile. This work presents a hybrid soft BSD that can provide task-oriented assistance by tuning the stiffness (0.58 N/mm, 0.92 N/mm, and 1.7 N/mm) and slack length (0 mm to 67 mm) in a compact design. The tunable stiffness allows for selecting a task-specific force profile, and the slack tuning will ensure that the device enables unhindered movement when assistance is not required.

We work on Koopman-based modeling and control introduces a physics-informed, data-driven framework for managing the complex, nonlinear dynamics of soft robots. By employing the Koopman operator theory, he transforms the nonlinear behavior of soft continuum manipulators into a higher-dimensional but approximately linear representation. This enables the use of efficient linear control techniques, such as convex optimization, for real-time shape control. His approach enhances computational efficiency while preserving the expressiveness needed to model soft robotic deformations, bridging the gap between accurate modeling and practical, fast control strategies.

Physics-based modeling in soft robotics centers on developing analytical frameworks that accurately capture the behavior of soft robots actuated by Twisted-and-Coiled Actuators (TCAs). Leveraging Cosserat rod theory, his models incorporate the nonlinear, coupled interactions between soft structures and multiple TCAs routed in complex geometries. This enables the prediction and control of soft robot motions—such as bending, twisting, and grasping—with high fidelity. His modeling approach balances mechanical realism with computational efficiency, providing a foundational toolset for design, simulation, and control of multifunctional soft robotic systems.