Learning-based control of a cyberoctopus

Inspired by the remarkable capabilities of octopuses, my work focuses on designing control strategies for a simulated soft-bodied octopus-like robot, the CyberOctopus. The continuous deformation and intricate mechanics of soft arms pose challenges in developing dynamic control models, especially when engaging with moving objects or coordinating multiple arms. Additionally, existing studies on soft arm manipulation have primarily focused on end-effectors, leaving whole-arm manipulation largely unexplored. To address these challenges, my research employs a hierarchical framework with static and dynamic activation strategies, leveraging the inherent mechanical intelligence of the arms to facilitate whole-arm interactions with other arms, obstacles, and moving objects in the environment. The CyberOctopus demonstrates refined manipulation and locomotion tasks in real-time, with potential applications in underwater exploration, search and rescue operations, and manufacturing processes.

Name thesis:degree_name

Ph.D.

Level thesis:degree_level

Dissertation

Discipline thesis:degree_discipline

Mechanical Engineering

Grantor

University of Illinois at Urbana-Champaign

Year dc:date

2024

Author dc:creator

• Shih, Chia-Hsien

Contributors dc:contributor

• Gazzola, Mattia
• Mehta, Prashant
• Krishnan, Girish
• Chowdhary, Girish

• Learning-based Control
• Octopus-inspired Robotics
• Hierarchical Decompositioin
• Motion Primitives
• Soft Robot Manipulation
• Soft Robot Locomotion
• Multi-arm Coordination

Statement dc:rights

• Copyright 2024 Chia-Hsien Shih

Language dc:language

eng, en

Handle dc:identifier

https://hdl.handle.net/2142/127317

OAI identifier oai:identifier

oai:www.ideals.illinois.edu:2142/127317