Energy based modeling, control and reconstruction of soft continuum arm

Chang, Heng-Sheng

Permalink

https://hdl.handle.net/2142/129231 Copy

Description

Title

Energy based modeling, control and reconstruction of soft continuum arm

Author(s)

Chang, Heng-Sheng

Issue Date

2025-04-28

Director of Research (if dissertation) or Advisor (if thesis)

Mehta, Prashant G

Doctoral Committee Chair(s)

Mehta, Prashant G

Committee Member(s)

• Gazzola, Mattia
• Krishnan, Girish
• Belabbas, Mohamed Ali
• Admal, Nikhil Chandra

Department of Study

Mechanical Sci & Engineering

Discipline

Mechanical Engineering

Degree Granting Institution

University of Illinois Urbana-Champaign

Degree Name

Ph.D.

Degree Level

Dissertation

Keyword(s)

• Control
• Soft Robotics
• Soft Continuum Arm

Abstract

This thesis presents a comprehensive framework for modeling, controlling, and reconstructing the shape of soft continuum arms, with a particular focus on biomechanical systems such as octopus muscular arms and engineered systems such as pneumatic soft manipulators. The work addresses key challenges in soft robotics, including the complex dynamics of highly deformable structures, the control of underactuated systems, and accurate, real-time shape estimation from noisy measurements. The foundation of this research is an energy-based modeling approach using Cosserat rod theory. The proposed framework provides a unified representation for passive elasticity and active actuation in terms of stored energy functions. The model is applied to two testbed systems: a biomechanical octopus arm incorporating longitudinal, transverse, and oblique muscles, and the BR2 pneumatic soft manipulator, demonstrating the versatility of the approach across biological and engineered systems. Building on this modeling framework, an energy shaping control method is developed to achieve complex three-dimensional motions. The proposed approach implicitly solves the matching conditions associated with underactuated systems and provides a stabilizing control law for desired configurations. The effectiveness of this method is demonstrated through simulations of an octopus arm in performing reaching and grasping tasks. Lastly, the problem of shape estimation in soft robotics is addressed. Two complementary posture reconstruction methods are presented for this problem. The first is an iterative algorithm that solves the smooth reconstruction problem, based on optimal control theory, to estimate continuous strains. The second is a fast reconstruction method utilizing a neural network framework, enabling real-time performance on the shape estimation task. Both methods are validated on simulated and physical soft robotic systems. This thesis contributes to the advancement of soft robotics by providing a cohesive framework for modeling, control, and shape estimation of soft continuum arms. The proposed methods offer new capabilities for the design and operation of highly dexterous and adaptable soft robotic systems across various applications.

Graduation Semester

2025-05

Type of Resource

Thesis

Handle URL

https://hdl.handle.net/2142/129231

Copyright and License Information

Copyright 2024 Heng-Sheng Chang

Owning Collections