Learning to Improve Capture Steps for Disturbance Rejection in Humanoid Soccer

TL;DR

This paper presents an online learning approach to enhance push-recovery in humanoid soccer robots, addressing balance recovery during dynamic gameplay and reporting on experimental results and competition performance.

Contribution

It introduces a novel online learning method for improving push-recovery in humanoid robots and provides insights from systematic experiments and RoboCup 2013 results.

Findings

Improved push-recovery capabilities demonstrated in experiments.

Successful localization despite goal color ambiguity.

Competitive performance in RoboCup 2013.

Abstract

Over the past few years, soccer-playing humanoid robots have advanced significantly. Elementary skills, such as bipedal walking, visual perception, and collision avoidance have matured enough to allow for dynamic and exciting games. When two robots are fighting for the ball, they frequently push each other and balance recovery becomes crucial. In this paper, we report on insights we gained from systematic push experiments performed on a bipedal model and outline an online learning method we used to improve its push-recovery capabilities. In addition, we describe how the localization ambiguity introduced by the uniform goal color was resolved and report on the results of the RoboCup 2013 competition.