A Foot-Mounted Pedestrian Localization System with Map Motion Constraints

Abstract

The work proposes a navigation system for pedestrian indoor localization. The system employs a cascaded architecture, where the information from the foot-mounted inertial unit is combined with the map-induced motion constraints. As the first stage, an Extended Kalman filter with a zerovelocity update (ZUPT) is used to get incremental heading and position information from the inertial sensor measurements. This is followed by motion filtering based on map constraints implemented in a form of SIR particle filter. Compared to a pure inertial approach without the map filtering, the proposed method shows a significant performance improvement for typical indoor scenarios and is shown to be superior to a non-gyroscopic pedometer approach with map constraints.