3D-Pedestrian Indoor Navigation including Altimeter Measurements

Kurzfassung

ZUPT-aided inertial navigation system (INS) and step-and-heading system (SHS) formulations of pedestrian dead-reckoning system (PDR) focus on obtaining accurate step measurements and thereby disregard external displacements of the user in the odometry estimation. In the presence of elevators or inclined escalators rides as indoor floor transition activities, these systems fail to track the full trajectory of the pedestrian. This thesis presents a context-aware, ZUPT-aided INS that supports superposition of displacements from step measurements and movement from transportation platforms. The system involves activity recognition from acceleration as well as magnetic and barometric features to perceive and track external displacements on the user. Incomplete tracking of platform movements are compensated by proposals from a transition model assisted with vertical position feedback from a barometric altimeter. The barometer is the backbone of this approach, for that reason a discussion of its measurement errors is included. First results approve reliable detection and tracking of platform rides resulting in improved floor estimation throughout three test walks.