Electrical Engineering
      and Computer Sciences

Electrical Engineering and Computer Sciences

COLLEGE OF ENGINEERING

UC Berkeley

Learning to Move Autonomously in a Hostile World

Leslie Ikemoto, Okan Arikan and David A. Forsyth

EECS Department
University of California, Berkeley
Technical Report No. UCB/CSD-05-1395
June 2005

http://www.eecs.berkeley.edu/Pubs/TechRpts/2005/CSD-05-1395.pdf

This paper describes a framework for controlling autonomous agents. We estimate an optimal controller using a novel reinforcement learning method based on stochastic optimization. The agent's skeletal configurations are taken from a motion graph which contains seamless transitions, guaranteeing smooth, natural-looking motion. The controller learns a parametric value function for choosing transitions at the branch points in the motion graph. Since this query can be completed quickly, synthesis is performed online in real-time. We couple the local controller with a global path planner to create a system which produces realistic motion even in a rapidly changing environment.


BibTeX citation:

@techreport{Ikemoto:CSD-05-1395,
    Author = {Ikemoto, Leslie and Arikan, Okan and Forsyth, David A.},
    Title = {Learning to Move Autonomously in a Hostile World},
    Institution = {EECS Department, University of California, Berkeley},
    Year = {2005},
    Month = {Jun},
    URL = {http://www.eecs.berkeley.edu/Pubs/TechRpts/2005/5542.html},
    Number = {UCB/CSD-05-1395},
    Abstract = {This paper describes a framework for controlling autonomous agents. We estimate an optimal controller using a novel reinforcement learning method based on stochastic optimization. The agent's skeletal configurations are taken from a motion graph which contains seamless transitions, guaranteeing smooth, natural-looking motion. The controller learns a parametric value function for choosing transitions at the branch points in the motion graph. Since this query can be completed quickly, synthesis is performed online in real-time. We couple the local controller with a global path planner to create a system which produces realistic motion even in a rapidly changing environment.}
}

EndNote citation:

%0 Report
%A Ikemoto, Leslie
%A Arikan, Okan
%A Forsyth, David A.
%T Learning to Move Autonomously in a Hostile World
%I EECS Department, University of California, Berkeley
%D 2005
%@ UCB/CSD-05-1395
%U http://www.eecs.berkeley.edu/Pubs/TechRpts/2005/5542.html
%F Ikemoto:CSD-05-1395