Key
<<<

Tuesday 12pm, 3 February 2015

0de25c4

Mechanism Design and Simulation for ULTRA Spine, a Tensegrity Robot | A Haptic Slip Feedback Device Using Interwoven Belts

Drew Sabelhaus | Colin Ho & Jonathan Kim

PhD Student - Berkeley Institute of Design, NASA

Abstract

Talk1: The Underactuated Lightweight Tensegrity Robotic Assistive Spine (ULTRA Spine) project is an ongoing effort to create a compliant, cable-driven, 3-degree-of-freedom, underactuated tensegrity core for quadruped robots. This work presents simulations and preliminary mechanism designs of that robot. Design goals and the iterative design process for an ULTRA Spine prototype are discussed. Inverse kinematics simulations are used to develop engineering characteristics for the robot, and forward kinematics simulations are used to verify these parameters. Then, multiple novel mechanism designs are presented that address challenges for this structure, in the context of design for prototyping and assembly. These include the spine robot's multiple-gear-ratio actuators, spine link structure, spine link assembly locks, and the multiple-spring cable compliance system.


Talk2: We introduce a novel haptic display designed to reproduce the sensation of both lateral slip and torque slip on a user’s fingertip. The device simulates three-degrees-of-freedom of slip by actuating four interwoven tactile belts on which the user’s finger rests. We present the specifications for the device, the mechanical design considerations, and initial evaluation experiments of the device. We generated the design of the device based on various psychophysical works in the literature, and conducted experiments on user discrimination of angled lateral movements, and direction of rotational torques. The initial feedback and results from our preliminary experiments suggest that our device has the potential to accurately simulate both tangential lateral slip and torque slip. We discuss plans for further improvement of the device, as well as future user studies and possible applications.

Bio

Talk 1: Drew is a Graduate Research Fellow at the University of California, Berkeley, pursing a PhD in mechanical engineering. He researches the design and control of a certain class of compliant robots, called tensile-integrity (“tensegrity”) systems, in collaboration with the Intelligent Robotics Group at NASA Ames Research Center. Drew previously worked on centimeter-scale mobile robots with the Micro-Robotics Lab at the University of Maryland, College Park, where he received a B.S. in mechanical engineering with a minor in computer science. His other research interests include control of highly nonlinear systems, hybrid systems identification and control, mechanism design for cable-driven robots, physics simulations for robotics, and best practices for diversity and inclusion in engineering disciplines.


Talk 2: Colin Ho is a Graduate Research Fellow at the University of California, Berkeley, pursing a PhD in mechanical engineering and a designated emphasis in new media. His research includes design of mechatronics systems for robotics and haptic interfaces, human-centered design for IoT applications, and new media art platforms for artistic performance and intervention. Colin previously led the development of MSLED, an underwater robot that explored Antarctic subglacial lakes, and researched autonomous underwater vehicle sampling techniques at Arizona State University, where he received a B.S. in mechanical engineering. He is also passionate about open-source robotics projects, and serves as a core developer for OpenROV.



Jonathan Kim is a 4th year undergrad in the EECS department at Berkeley. He is very interested in the intersection of EECS and medicine, particularly in how people working in the field of EECS can help engineer better tools for physicians to use. His research includes the development of haptic devices for telesurgical applications, user interfaces for surgical robots, as well as point-of-care diagnostic devices for HIV/AIDS treatment.