(logo by Janette Forte)
This is the official site of the UT Austin Villa 3D Simulation team from the Department of Computer Sciences at the University of Texas at Austin.

This web page provides supplementary material to the following paper:

On Optimizing Interdependent Skills: A Case Study in Simulated 3D Humanoid Robot Soccer

Daniel Urieli, Patrick MacAlpine, Shivaram Kalyanakrishnan, Yinon Bentor, Peter Stone

Published in the Proc. of 10th Int. Conf. on Autonomous Agents and Multiagent Systems (AAMAS 2011) in Taipei, Taiwan.

The full paper can be found here


Abstract:

In several realistic domains an agent's behavior is composed of multiple interdependent skills. For example, consider a humanoid robot that must play soccer, as is the focus of this paper. In order to succeed, it is clear that the robot needs to walk quickly, turn sharply, and kick the ball far. However, these individual skills are ineffective if the robot falls down when switching from walking to turning, or if it cannot position itself behind the ball for a kick.
This paper presents a learning architecture for a humanoid robot soccer agent that has been fully deployed and tested within the RoboCup 3D simulation environment. First, we demonstrate that individual skills such as walking and turning can be parameterized and optimized to match the best performance statistics reported in the literature. These results are achieved through effective use of the CMA-ES optimization algorithm. Next, we describe a framework for optimizing skills in conjunction with one another, a little-understood problem with substantial practical significance. Over several phases of learning, a total of roughly 100--150 parameters are optimized. Detailed experiments show that an agent thus optimized performs comparably with the top teams from the RoboCup 2010 competitions, while taking relatively few man-hours for development.


Below are a few videos that demonstrate the work reported in this paper.



A demonstration of our an optimized back walk, optimized using CMA-ES, which runs in the speed of 1.03 m/s.
Download videos: avi mp4




A demonstration of our an optimized front walk, optimized using CMA-ES, which runs in the speed of 1.07 m/s.
Download videos: avi mp4




A demonstration of our an optimized Kick, optimized using CMA-ES. The agent kicks to 5.09m on average.
Download videos: avi mp4




The UTAustinVilla 2010 agent drives the ball to goal. This was used as a fitness evaluation procedure in our evolutionary optimization of agent walks.
Download videos: avi mp4

For any questions, please contact Patrick MacAlpine and Daniel Urieli.