Experimental study on sensor fusion to improve real time indoor localization of a mobile robot

Abstract

Reliable indoor navigation of mobile robots has been a popular research topic in recent years. GPS systems used for outdoor mobile robot navigation cannot be used indoor (warehouse, hospital or other buildings) because GPS require an unobstructed view of the sky. Therefore a specially designed indoor localisation system for mobile robot is needed. This paper aims to develop a reliable position and heading angle estimator for real time indoor localization of a mobile robot. The proposed technique is achieved by fusing three different sensor modules based on infrared sensing, calibrated odometry and gyroscopes and applying the real time filtering technique Kalman filter. It can provide filtered and reliable information of a mobile robot's current location and orientation relative to its environment. Extensive experimental results are provided to demonstrate its improvement over conventional dead reckoning method.