Συντάχθηκε 27-09-2013 18:20
από Balasi Panagiota
Email συντάκτη:
Ενημερώθηκε:
-
Ιδιότητα: -.
ΠΟΛΥΤΕΧΝΕΙΟ ΚΡΗΤΗΣ
Σχολή Ηλεκτρονικών Μηχανικών και Μηχανικών Υπολογιστών
Πρόγραμμα Προπτυχιακών Σπουδών
ΠΑΡΟΥΣΙΑΣΗ ΔΙΠΛΩΜΑΤΙΚΗΣ ΕΡΓΑΣΙΑΣ
ΝΙΚΟΛΑΟΥ ΚΑΡΓΑ
με θέμα
Αξιόπιστος Εντοπισμός Θέσης
για το Πρωτάθλημα Standard Platform του RoboCup
Robust Localization
for the RoboCup Standard Platform League
Δευτέρα 30 Σεπτεμβρίου 2013, 1:30μμ
Αίθουσα Εργαστηρίου Intelligence, Κτίριο Επιστημών, Πολυτεχνειούπολη
Εξεταστική Επιτροπή
Αναπληρωτής Καθηγητής Μιχαήλ Γ. Λαγουδάκης (επιβλέπων)
Καθηγητής Μίνως Γαροφαλάκης
Επίκουρος Καθηγητής Άγγελος Μπλέτσας
Abstract
RoboCup is an international competition that promotes research in the field of Robotics and Artificial Intelligence and focuses mainly on the game of soccer. In the Standard Platform League (SPL) all teams use identical robots, namely the Aldebaran NAO humanoid robots and focus on software development aiming to develop fully autonomous robots. A key aspect of robot autonomy is the ability of self-localization. Robot localization is the process of determining the pose (position and orientation) of a robot relative to a given map of the environment. Robots operating in the SPL league compete on a symmetric field of fixed size with same-colored goals and field lines. Localization can be quite challenging because of the high uncertainty in locomotion, as well as in object recognition, due to landmark ambiguities. This thesis addresses the problem of self-localization in SPL relying on extended Kalman filtering. Our localization algorithm uses observations corresponding to field landmarks and odometry information and provides an estimate for the robot's position in the field. To deal with the problem of data association a multiple-hypothesis tracking approach has been applied. An augmented state with odometry error parameters is used to cope with odometry errors. Our localization approach offers more accurate results compared to a previous approach based on particle filters. The output of localization is used in our RoboCup team “Kouretes” as input to the module estimating the global game state used for developing game strategies and assigning appropriate roles to each robot in the team.
Συντάχθηκε 30-09-2013 10:17
από Michail Lagoudakis
Email συντάκτη: lagoudakis<στο>tuc.gr
Ενημερώθηκε:
-
Ιδιότητα: ΔΕΠ ΗΜΜΥ.
Προσοχή! Αλλαγή αίθουσας!
Η παρουσίαση θα πραγματοποιηθεί στην Αίθουσα Συνεδριάσεων ΗΜΜΥ, Κτίριο Επιστημών, Πολυτεχνειούπολη.
