Title: Planning with Robot Capabilities and Affordances
Date: Wednesday, July 13, 2016
Time: 1:00pm
Location: MiRC 102B
Abstract:
This proposal outlines work to extend previous research on planning with affordances to incorporate uncertainty management techniques. First, we motivate the discussion by relating the author’s experience with various robot hardware and software configurations. Next, we report on agent-aware decompositions of the affordance representation, as well as shortcomings regarding operation in uncertain environments. Set-valued functions are introduced for dealing with this deficiency, and recent literature on related concepts and techniques is summarized. After diving into one method for determining the extent of such sets for dynamical systems under control and uncertainty, we propose how to compose these ideas into a novel and powerful system capable of planning manipulation techniques.
Committee:
Dr. Henrik Christensen (Advisor), School of Electrical and Computer Engineering, Georgia Institute of Technology
Dr. Stephen Balakirsky, Aerospace, Transportation and Advanced Systems Laboratory, Georgia Tech Research Institute
Dr. Patricio Vela, School of Electrical and Computer Engineering, Georgia Institute of Technology
Dr. Aaron Ames, School of Mechanical Engineering, Georgia Institute of Technology
Dr. Dmitry Berenson, Computer Science Department, Worcester Polytechnic Institute
Youtube Link: https://www.youtube.com/watch?v=CNay1sMDGYA