REAL TIME TRACKING OF AN OMNIDIRECTIONAL ROBOT - An Extended Kalman Filter Approach

Abstract

This paper describes a robust localization system, similar to the used by the teams participating in the Robocup Small size league (SSL). The system, developed in Object Pascal, allows real time localization and control of an autonomous omnidirectional mobile robot. The localization algorithm is done resorting to odometry and global vision data fusion, applying an extended Kalman filter, being this method a standard approach for reducing the error in a least squares sense, using measurements from different sources.