Mobile robot map-based localization using approximate locations and the extended kalman filter (original) (raw)

Map-Based Localization approaches use a local map of the sensed environment that is matched against a previously stored map to correct the robot localization in the world. In many cases these methods are based on a probabilistic representation of the spatial uncertainty and use the Kalman Filter (KF) or the Extended Kalman Filter (EKF) to update the robot's location estimation. On the other hand, Fuzzy Logic has been widely used to generate robust and efficient navigational behaviors for mobile robots in spite of the presence of noise and non-linearities in the system. In this paper we introduce a map-based localization approach that combine a fuzzy robot's location, a possibilistic method to propagate the uncertainty in the robot's motion and the use of the EKF to decrease the spatial uncertainty when valid landmarks are found. Experiments in simulation and in the real world are shown to validate the proposal.