Why is modeling and localization seen as a combined task?
Why not use existing mapping approaches?
What are suitable localization features?
How to extract them?
How does it support high level tasks on the robot?
Back to first slide
View graphic version