Methods for hardware design, motion retargeting, system integration, and tactile sensing for teleoperated robots

Although robots’ decision making and manipulation capabilities have not improved to human levels, teleoperated robot systems offer a middle ground, combining humans’ intuition for manipulation and decision making with robots’ ability to work in relatively inhospitable environments. Teleoperated robots must efficiently translate their operator’s commands to the remote environment, and also carry sensor data from the environment back to the operator. In Chapter 2 of this thesis, I present contributions to AVATRINA, a humanoid teleoperated robot with an immersive virtual reality interface. I develop a set of task-oriented metrics to optimize the robot’s hardware design for better workspace coverage, and implement a hand retargeting system to allow intuitive control of a low-Degree-Of-Freedom (DOF) anthopomorphic robot hand. In addition, I describe a modular approach to robot software architecture, enabling rapid hardware and software iteration on the robot platform. Immersive teleoperated robots feed a wide array of information back to their operators, such as visual, audio, and haptic signals. Recently, high resolution vision-based tactile sensors have been developed that can provide touch information suitable for robotics applications. These sensors present an exciting new input modality for robots, giving them tactile information on a much finer level than joint torques or end effector wrenches. In Chapter 3 of this thesis, we present an exploration of one such sensor: The Punyo Soft-bubble sensor. It consists of an inflatable rubber membrane resistant to wear and tear, whose deformation is sensed by an embedded depth camera. A physics-based model of the sensor is developed to convert the observed depth image and pressure signals to dense contact force estimates. The model can be calibrated with only a few data samples, and generalizes to unseen contact geometries and trajectories.

Name thesis:degree_name

M.S.

Level thesis:degree_level

Thesis

Discipline thesis:degree_discipline

Computer Science

Grantor

University of Illinois at Urbana-Champaign

Year dc:date

2024

Author dc:creator

• Peng, Jing-Chen

Contributors dc:contributor

• Hauser, Kris K

• Robotics
• Teleoperation
• Tactile Sensing
• Finite Element Method

Statement dc:rights

• Copyright 2024 Jing-Chen Peng

Language dc:language

en, eng

Handle dc:identifier

https://hdl.handle.net/2142/124264

OAI record oai:identifier

https://www.ideals.illinois.edu/oai-pmh?verb=GetRecord&metadataPrefix=etdms&identifier=oai%3Awww.ideals.illinois.edu%3A2142%2F124264