HeadSLAM - simultaneous localization and mapping with head-mounted inertial and laser range sensors

Abstract

Self-localization of users and their wearable computers is essential to many applications, but so far, most implementation rely on a-priori information and pre-deployed infrastructures such as maps. We show how techniques from mobile robotics, namely simultaneous localization and mapping can be used to automatically generate both localization information and 2D environment maps using head-mounted inertial and laser range sensors. We present an initial implementation and the results of a number of experiments conducted in an office environment with focus on map degradation caused by shape ambiguities in the environment such as corridors.