Indoor Pedestrian Positioning and Tracking using INS and RF Sensor Fusion

Abstract

In this digital era, electronic devices are developing into more sophisticated equipment to perceive the local and global information such as fitness and activity trackers as well as for navigational purposes. The spatial awareness is fast becoming the key element in improving human life and navigation solutions are growing in demands. While accurate outdoor navigation has been widely available through Global Positioning System (GPS), the GPS does not work in indoor environments due to weak received signal and multipath reflection. In this regard, a number of indoor positioning and tracking solutions have been proposed, e.g. inertial and radio frequency (RF) based navigation. Low-cost inertial sensors are economically attractive and miniature in size, thereby become a popular choice in indoor positioning solutions. However, such devices are prone to errors that result in a large position drift. For accurate RF based navigation, the techniques rely on expensive high-precision time hardware and complex deployment of the wireless transceivers. This thesis describes the development of indoor pedestrian positioning and tracking using Inertial Navigation System (INS) and RF sensor fusion. The proposed system fuses the positioning and tracking information from INS and RF to produce better positioning and tracking accuracies. This research tackles the existing issues on low-cost INS, position and heading drifts, by proposing two novel techniques. First, we propose a better stance phase detector that directly improve the position drift. Secondly, we develop an efficient turn detector that is threshold-less, thus robust to different group of pedestrians and operating conditions. The proposed turn detector provides information of turns made by pedestrians that greatly assists in correcting the heading drift. The introduction of both techniques in the Kalman-based navigation framework produce a better positioning and tracking performance compared to the state of the art. The unconvinced performance of INS based navigation in the long run or challenging indoor arrangements, resulted from accumulated drift over time and insufficient dynamic range of the inertial sensors, are discussed in literature. This research proposes the fusion between INS and RF based navigation technology known as Device Free Location (DFL). The DFL provides absolute positioning that benefits in resetting the drifted positioning of INS. The DFL network uses low-cost IEEE 802.15.4 wireless transceivers and deploys in square-shape arrangement.