Enhanced Acceleration Phase Tracking for Moving Platform Detection in 3D Indoor Navigation

Abstract

State of the art inertial navigation systems (INS) use pedestrian dead reckoning (PDR) to estimate the trajectory of a user in indoor environments. These algorithms enforce a human gait model on the inertial measurements, in order to restrain the accumulation of sensor errors. However, in the presence of external transportation platforms like evators or escalators, these assumptions cause the PDR to ignore the external displacement and it loses track of the user’s trajectory. In this paper, we follow a heuristic based approach for the detection of external Transportation platforms in indoor pedestrian navigation. We propose a method for reliably detecting platforms that is independent from bias errors. Different acceleration bias estimation techniques are compared based on accuracy and robustness. With the new detection technique 95% of the moving platforms are detected properly, where the remaining 5% are not detected because of missing or short acceleration phases during stance phases.