Low cost inertial-based localization system for a service robot

Abstract

The knowledge of a robot’s location it’s fundamental for most part of service robots. The success of tasks such as mapping and planning depend on a good robot’s position knowledge. The main goal of this dissertation is to present a solution that provides a estimation of the robot’s location. This is, a tracking system that can run either inside buildings or outside them, not taking into account just structured environments. Therefore, the localization system takes into account only measurements relative. In the presented solution is used an AHRS device and digital encoders placed on wheels to make a estimation of robot’s position. It also relies on the use of Kalman Filter to integrate sensorial information and deal with estimate errors. The developed system was testes in real environments through its integration on real robot. The results revealed that is not possible to attain a good position estimation using only low-cost inertial sensors. Thus, is required the integration of more sensorial information, through absolute or relative measurements technologies, to provide a more accurate position estimation.