Seamless indoor-outdoor robust localization for robots

Abstract

In this paper we present a unified localization technique for indoor-outdoor environments that allows a seamless transition between a mapped zone using laser rangefinder on-board sensors and a GPS based localization zone. Different situations are detected during the indoor-outdoor transitions, in which the sensors used change and the localization estimator has to manage them properly for a continuous localization. The quality in the GPS measurements and the zone where the robot is localized are used to determine the best instant for switching the localization parameters for adapting to the situations.