A case study on sensors and techniques for pedestrian inertial navigation

TLDR: In this article, a comparison of the performance in realistic conditions is carried out between foot-mounted and belt-mounted techniques given an inertial measurement unit (IMU) commercialized by XSens.

Abstract: The interest in location based services is growing in several applications. The literature exhibits a wide spectrum of technology to complement the well-knwon limitations of satellite based positioning systems in constrained environments such as indoors or urban canyons. This paper focuses on inertial sensors and systems to locate pedestrians indoor without infrastructure. The theoretical background of a recently developped belt-mounted inertial navigation system (INS) is carefully depicted here. The approach aims to facilitate the equipment and the mobility of the users while maintaining repeatable performance. Therefore, a comparison of the performance in realistic conditions is carried out between foot-mounted and belt-mounted techniques given an inertial measurement unit (IMU) commercialized by XSens. Then, this commercial IMU and an IMU based on ADXL345 and ITG3200 were compared, given the belt-mounted algorithm, in terms of positioning performance. The results, supported by dozens of experiments involving different participants, show that the belt-mounted technique is as efficient as the foot-mounted one since the average error in position is less than 2% of the travelled distance about 200m. Whereas their costs are very different, the commercial and the integrated IMU reach a similar accuracy. The belt-mounted device achieve repeatable and efficient pedestrian indoor positioning in real-time with low-cost inertial sensors.