This paper presents an approach to vision-based self-localization using the combination of particle filter and preprocessed image database. The robot uses particle filter with odometery data and landmark pose for position tracking. Furthermore, it uses the image database to globally localize itself when it has been kidnapped or when no landmark is visible.
Published in:
Computational Intelligence in Robotics and Automation (CIRA), 2009 IEEE International Symposium on
Date of Conference: 15-18 Dec. 2009