Title: Autonomous Environment Manipulation to Facilitate Task Completion

Martin Levihn
Robotics Ph.D. Student
School of Interactive Computing
College of Computing
Georgia Institute of Technology

Date: August 21, 2014 (Thursday)
Time: 10am-12pm EST
Location: CCB 345

Committee:
Dr. Henrik Christensen (Advisor), School of Interactive Computing, Georgia Tech
Dr. Frank Dellaert, School of Interactive Computing, Georgia Tech
Dr. Charles Isbell, School of Interactive Computing, Georgia Tech
Dr. Magnus Egerstedt, School of Electrical and Computer Engineering, Georgia Tech
Dr. Tomás Lozano-Pérez, Department of Electrical Engineering and Computer Science, MIT
Dr. Leslie Kaelbling, Department of Electrical Engineering and Computer Science, MIT

Abstract:
A robot should be able to autonomously modify and utilize its environment to assist its task completion. Mobile manipulators and humanoid robots have two distinct capabilities: locomotion and manipulation. While these capabilities are usually utilized in isolation, we propose to tightly couple them. A robot should be able to use its manipulation capabilities to move obstacles out of its way or even utilize environment objects as tools to make way. However, there are two primary challenges in developing algorithms that can achieve such behavior on real robot systems: the inevitable inaccuracies in perception as well as actuation that occur on physical systems and the exponential size of the search space.

To address these challenges, we first extend the previously introduced domain of Navigation Among Movable Obstacles (NAMO), which allows a robot to move obstacles out of its way. We extend it to handle the underlying issue of uncertainty. In the second part of this proposal we then broaden the scope of the environment interactions. We introduce the EnvironmenT Aware Planning (ETAP) domain which enables the robot to utilize environment objects as tools to accomplish its task rather than just moving them out of the way.