Assistive Teleoperation: A New Domain for Interactive Learning - Robotics Institute Carnegie Mellon University

Assistive Teleoperation: A New Domain for Interactive Learning

Conference Paper, Proceedings of AAAI '12 Fall Symposium on Robots Learning Interactively from Human Teachers, November, 2012

Abstract

In assistive teleoperation, the robot attempts to predict the user’s intent and augments his or her input based on this prediction, in order to simplify the task. Our recent work on policy blending formalizes assistance as an arbitration of the user’s input and the robot’s prediction. In this extended abstract, after quickly overviewing our findings on the two main components of policy blending -- prediction and arbitration -- we discuss the important role that interactive learning can play in improving assistive teleoperation and human-robot collaboration in general.

BibTeX

@conference{Dragan-2012-7632,
author = {Anca Dragan and Siddhartha Srinivasa},
title = {Assistive Teleoperation: A New Domain for Interactive Learning},
booktitle = {Proceedings of AAAI '12 Fall Symposium on Robots Learning Interactively from Human Teachers},
year = {2012},
month = {November},
keywords = {teleoperation, human-robot interaction, shared autonomy, interactive learning},
}