Title: Control Schemes for Teleoperating Non-Anthropomorphic Robots in Virtual Reality

Date: Monday, March 13, 2023

Time: 12:00pm-2:00pm ET

Location: Klaus Advanced Computing Building 2100

Zoom link: https://gatech.zoom.us/j/91862681665?pwd=WThDRUVJQzFjU0RPNHE4eG1yZjBnZz09

Jennifer Molnar

Robotics Ph.D. Student

Woodruff School of Mechanical Engineering

Georgia Institute of Technology

Committee:

Dr. Sonia Chernova (Advisor), School of Interactive Computing, Georgia Institute of Technology

Dr. Harish Ravichandar, School of Interactive Computing, Georgia Institute of Technology

Dr. Seth Hutchinson, School of Interactive Computing, Georgia Institute of Technology

Dr. Charles Kemp, Coulter Dept of Biomedical Engineering, Georgia Institute of Technology

Dr. Yiğit Mengüç, College of Engineering, University of Oregon

Abstract:

Virtual reality (VR) has emerged as a popular teleoperation interface for robotic systems due to the immersive nature of VR interactions. In VR, users use their hands, or hand-held controllers that capture their hand movement, to interact with the environment. However, just as living animals have evolved specialized embodiments that differ from those of a human (e.g. tentacles, wings, prehensile tails, etc.), specialized robotic systems are often designed with non-anthropomorphic embodiments (e.g. treads, wheels, arms with differing degrees of freedom, etc.). Such differences in embodiment between human hands and the robot introduce challenges for the design of intuitive user teleoperation interfaces. In this thesis, I will contribute a VR interface and computational framework for identifying a functional control scheme that enables the intuitive teleoperation of a diverse set of non-anthropomorphic robots. Specifically, my work will contribute:

• A comprehensive survey of control scheme designs for teleoperating both real and virtual non-anthropomorphic robots, to characterize common design principles and identify unexplored research areas.
• A VR “puppet-master” platform for identifying and validating user-generated mappings between a user’s hand and a simulated robot.
• A characterization of generalizable design principles for particular classes of non-anthropomorphic robots, as well as forms and/or motions that may require personalized control schemes or training protocols.
• A computational control scheme that enables teleoperation of a wide range of non-anthropomorphic robots.