Alexandru Mardare Undergraduate Dissertation 2015/16

NAO Robot Football Perception

Supervised by G.Brown

Abstract

Vision can be considered the most important human sense because it provides a lot of information about the surrounding environment. As robots and machinery become more and more part of humanity's life, they must be equipped with sophisticated perception algorithms that will help them survive in the environment and perform their tasks without any impediments.

The aim of this project is to create an algorithm that will help a NAO robot detect its environment, such as the field lines and the goal posts on the field. He should be able to distinguish which goal belongs to his team, no matter they switch sides at half-time. Apart from line detection, a robot should be able to detect the ball in the field and the rest of the robots and to distinguish between teammates and opponents.

The main features that were implemented for this project are object detection, which include ball and robot recognition and also the perception of field lines. A basic post detection algorithm has also been implemented.