Evaluation of the Human Body Mask Effects on GNSS Wearable Devices for Outdoor Pedestrian Navigation Using Fisheye Sky Views
Ni Zhu, AME-GEOLOC, University Gustave Eiffel; Athanase Bouronopoulos, IRSTV, École Centrale de Nantes; Thomas Leduc, Myriam Servières, Nantes Université, ENSA Nantes, École Centrale Nantes, CNRS, AAU-CRENAU, UMR; Valérie Renaudin, AME-GEOLOC, University Gustave Eiffel
Date/Time: Tuesday, Apr. 25, 4:23 p.m.
Abstract—An increasing number of Global Navigation Satellite Systems (GNSS)-based wearable devices emerged in our daily life in recent years. The utilities of these wearable devices vary from entertainment applications to safety and reliability-critical applications (such as blind people guidance, health monitoring, behavior-based soft mobility insurance, etc.). A key challenge for safety and reliability-critical applications using GNSS is the tracking of reliable and high quality signals in stringent environments. One particularity for wearable devices compared to other GNSS devices is that GNSS signals can be masked by not only the surrounding obstacles but also by the body of the users. Many research work addresses the impact of the surrounding obstacles on GNSS signal receptions whereas very few research studies the impact of human body mask.
The objective of this paper is to investigate the impact of the user body mask on wearable GNSS devices. By fixing a fisheye camera on GNSS wearable devices on two body parts of a pedestrian, i.e., hand and foot, the sky view representing the “point of view” of the device can be obtained. The images are first segmented to distinguish the sky from the obstacles. The GNSS satellites can be then classified as Line-of-Sight (LOS) and Non-Line-of-Sight (NLOS) by projecting them on the fisheye images. The experimentation results show that, in an open sky context, the mean percentage of the sky view for the receiver on the foot during the stance phase of a walking cycle is higher than that in the hand. This leads to, on average, a higher number of visible satellites for foot-mounted receivers and better satellite geometry. The trajectory of the foot-mounted device is generally more accurate with less uncertainty in all the evaluated environments (open sky, suburban, deep urban and urban canyon). This preliminary research leads to a counterintuitive outcome since the foot-mounted GNSS device is globally able to provide better positioning than the hand-held device. Accordingly, the foot-mounted GNSS device could be further considered as a more reliable fusion source for hybridization positioning systems used for safety-critical or reliability-critical applications.
Index Terms—GNSS, wearable devices, body mask, fisheye camera