Andrei Sandu Undergraduate Dissertation 2015/16

NAO Robot Football - movement and coordination

Supervised by G.Brown

Abstract

Football or soccer is a sport currently played at international level, with an overwhelming  popularity in most countries of the world. The complexity of the game’s rules, the great diversity of tactics available and the high freedom for motion that this sport implies has made  it spectacular to watch, and very dynamic all the same. For this reason, the sport has been  used as a basis for a scientific endeavour related to biped robotics since 1997. The RoboCup   Robot Football League is an annual event dedicated to robotics, where the context of football as a sport is used to showcase and compare complex control systems and algorithms  that enable robots to perform complicated motions, make autonomous decisions and cooperate with other robots to simulate a football competition. The high dynamics of the game  requires complex motions, as well as agile response to sensors and coordination, hence the  event poses great scientific value each year.

The aim of this project is to produce the algorithms necessary for a functional NAO football  robot capable to compete in the RoboCup Standard Platform League for The University of  Sheffield. In achieving this, the project focuses on identifying and implementing the foundation for all robot’s motions, as well as discussing solutions for coordination, intelligent tactical  play, and overall integration of the system’s parts. The project achieved a number of custom  motions that are essential for football playing, including kicking motions and blocking shots,  as well as exploring advanced modelling solutions for gait generation. Nonetheless, the experimental results obtained provide a good measure of the fitness of the solutions implemented.