Localization of a mobile robot with omnidirectional wheels using angular Kalman filtering and triangulation

Localization is one of the fundamental problems in mobile robot navigation. Several approaches to cope with the dynamic positioning problem have been made. Most of them use an extended Kalman filter (EKF) to estimate the robot pose -position and orientation- fusing both the robot odometry and external measurements. In this paper, an EKF is used to estimate the angles, relative to the robot frame, of the straight lines from a rotating laser scanner to a set of landmarks. By using this method angles are predicted, between actual laser measurements, by means of the time integration of its time derivative, which depends upon the robot kinematics. Once these angles are estimated, triangulation can be consistently applied at any time to determine the robot pose. In this work, a mobile robot with three amnidirectional wheels -that consist of two spherical rollers- is considered. Computer simulations showing the accuracy of this method are presented.