State Estimation and Control of Continuum Robots with Applications in Force Sensing and Tissue Palpation
Date Issued
May 2026
Author(s)
Gaston, Joshua Brady
The University of Tennessee
Advisor(s)
Daniel Caleb Rucker
Additional Advisor(s)
Eric Barth
Tayfun Ertop
Dustin Crouch
Abstract
Soft and continuum robots offer unique advantages for safe interaction, adaptability, and operation in confined environments due to their highly compliant structures. However, their underactuated nature and effectively infinite degrees of freedom pose significant challenges for modeling, control, and state estimation, particularly for large flexible robots. This dissertation addresses these challenges by developing mechanics-based modeling, control, and estimation frameworks that enable robust task-space control, full-state estimation from sparse measurements, and intrinsic force sensing without dedicated distal sensors. First, a task-space control framework is developed for planar continuum robots using a highly underactuated discrete rod mechanics model formulated in maximal coordinates. This approach enables straightforward and computationally efficient control design. Input–output feedback linearization combined with sliding mode control achieves robust task-space tracking despite significant modeling uncertainty, inaccurate initial conditions, and output-only sensing. A model-based observer is introduced to estimate the full system state from sparse measurements, enabling closed-loop control with minimal sensing requirements. This modeling and estimation framework is then extended to three-dimensional continuum robots using a symmetric constrained-Lagrangian formulation with quaternion-based orientation representation and constraint stabilization. A physically intuitive observer based on virtual forces and moments is developed to estimate the robot’s dynamic state. Experimental validation on a large tendon-driven continuum robot demonstrates accurate state estimation during dynamic actuation using optical tracking. Finally, the dissertation leverages the inherent compliance of continuum robots to enable intrinsic force sensing for surgical robotic palpation. Deflection-based force-sensing probes are designed and modeled using quasi-static Cosserat and Kirchhoff rod formulations, enabling estimation of both bending and axial contact forces without dedicated tip force sensors. Experimental validation on a surgical continuum robot demonstrates accurate force estimation and supports applications such as tissue stiffness mapping and tumor localization. Collectively, this work advances the modeling, control, estimation, and sensing capabilities of continuum robots, providing a unified, sensor-efficient foundation for robust autonomy in medical and unstructured environments.
Degree
Doctor of Philosophy
Major
Mechanical Engineering
File(s)![Thumbnail Image]()
Name
GASTON-PRIMARY-2026.pdf
Size
6.31 MB
Format
Adobe PDF
Checksum (MD5)
18a9ebe3ad2085954fa6ce883fddd7bd