A Self-Developed Indoor Three-Dimensional Pedestrian Localization Platform Based on MEMS Sensors

Shengbo Sang,Ruiyong Zhai,Wendong Zhang,Qirui Sun,Zhaoying Zhou
DOI: https://doi.org/10.1108/sr-07-2014-682
2015-01-01
Sensor Review
Abstract:Purpose – This study aims to design a new low-cost localization platform for estimating the location and orientation of a pedestrian in a building. The micro-electro-mechanical systems (MEMS) sensor error compensation and the algorithm were improved to realize the localization and altitude accuracy. Design/methodology/approach – The platform hardware was designed with common low-performance and inexpensive MEMS sensors, and with a barometric altimeter employed to augment altitude measurement. The inertial navigation system (INS) – extended Kalman filter (EKF) – zero-velocity updating (ZUPT) (INS-EKF-ZUPT [IEZ])-extended methods and pedestrian dead reckoning (PDR) (IEZ + PDR) algorithm were modified and improved with altitude determined by acceleration integration height and pressure altitude. The “AND” logic with acceleration and angular rate data were presented to update the stance phases. Findings – The new platform was tested in real three-dimensional (3D) in-building scenarios, achieved with position errors below 0.5 m for 50-m-long route in corridor and below 0.1 m on stairs. The algorithm is robust enough for both the walking motion and the fast dynamic motion. Originality/value – The paper presents a new self-developed, integrated platform. The IEZ-extended methods, the modified PDR (IEZ + PDR) algorithm and “AND” logic with acceleration and angular rate data can improve the high localization and altitude accuracy. It is a great support for the increasing 3D location demand in indoor cases for universal application with ordinary sensors.
What problem does this paper attempt to address?