Research on multi-source fusion based seamless indoor/outdoor positioning technology

Abstract

GNSS has been applied widely. Yet because satellite signals are vulnerable and susceptible to blockage, the operability of GNSS in urban canyons are greatly hampered and GNSS even proves useless for indoor settings. This paper proposes system architecture for the integration of WLAN fingerprinting, visual positioning, baroceptor-derived altitude estimation and GNSS for seamless indoor/ outdoor positioning for vehicles and pedestrians. This architecture augments GNSS through the integration of terminal-side/network-side positioning and position/ measurement domain. After temporal and spatial synchronization, data from each sensor is filtered by sub filters and then processed by the main filter. The purpose of these operations is to provide accurate and continuous estimates of positions. Tests conducted in the new technology center of CAS show that the architecture proposed can achieve seamless indoor/outdoor positioning, with a better accuracy performance than any single-source method as the former still maintains accuracy and continuity when the later generates noticeable errors. Calculation shows that multisource fusion has an accuracy level of better than 1 m (outdoor)/3 m (indoor), hence capable of meeting users’ demand for seamless indoor/outdoor positioning.