Are Kalman filters (or other state estimation algorithms) used for global localization?
In the ICRA Duckietown paper, global localization is discussed in Section IV-C. It discusses using the AprilTags to determine the absolute pose of the robot. It does not mention anything about using a state estimation algorithm. However, since the AprilTag detection is subject to uncertainty in the robot’s camera, isn’t something like a Kalman filter needed to account for the uncertainty? If an estimation algorithm is not used, then why?
Thank you for any replies.

Asked question