Muhtadin, Mauridhi Hery Purnomo : DETEKSI FITUR DAN PENENTUAN LOKASI ROBOT PEMAIN SEPAK BOLA BERBASIS PENANDA YANG TIDAK UNIK

Muhtadin ST., MT.
Prof.Dr.Ir. Mauridhi Hery Purnomo M.Eng.

Year

2014

Published in

The 2nd Indonesian Symposium on Robot Soccer Competition

External link

-

Type

Seminar Nasional

Keywords

feature localization triangulation


Abstract

This paper describe about feature extraction in RoboCup 2013 environment. The features are the ball, the goals and field line. By knowing the features, robot can estimate his position (x, y, #952) in the field. Coordinates (x, y) obtained from the triangulation method, and #952 obtained from the orientation sensors which planted in the robot. If it was equipped with a self-localization module, he could participate in global decision making for his team\s best interest, such as game strategy, team coordination and role assignment among the players. This research discussed about features extraction for the ball, goal and field boundary. Goal identification is done by applying hough transform to detect lines that represent goalposts and crossbar. Polynomial curve fitting method used to get the distance between robot and the goal. To improve our robot performance, an orientation sensor was added. This sensor can be used to distinguish between oponent#8223s side and our side. Then using non unique landmark and triangulation method, we do localization for soccer robot. As result, based on experiment robot can do localization with 0.1154m error.