Sensor-fusion and cooperation based vehicular navigation systems for GNSS degraded environments

Abstract

Vehicular application technologies, i.e., autonomous driving cars, have rapidly developed in recent years, as they provide convenience, location-related information, and safety to transportation users. Specially, vehicular navigation technology which provides precise positioning has become a vital component in most transportation systems and, with recent developments, now uses a Global Navigation Satellite System (GNSS). GNSS performance degradations can have various sources, but disconnections of the GNSS signal during signal outages (i.e., tunnels) and non-line-of-sight (NLOS) GNSS signals in urban environments (i.e., urban canyons) lead to significant errors. A number of studies have been conducted with regard to precise positioning methods, but they require a heavy computation cost, show low performance, or have high implementation costs to realize a viable vehicular navigation system. Therefore, in this thesis, two navigation systems which are robust against harsh signal environments and which outperform conventional navigation systems are introduced for vehicles. One of the navigation systems is a sensor-fusion-based navigation system of the type slated for use on Korean next-generation high-speed trains. It is modified to make it feasible for an actual train environment and its outperformance is demonstrated in a comparison to conventional navigation systems with empirical data. Moreover, we develop a non-line-of-sight (NLOS) GNSS signal-detection algorithm and propose cooperative navigation techniques that significantly enhance accuracy and multipath-robustness in urban environments. We discuss the theoretical derivation and realization of the proposed techniques and demonstrate their performance capabilities with numerous Monte Carlo simulations using field measurements with multipath delays. The proposed techniques significantly improve the positioning accuracy of all vehicles and outperform the conventional cooperative positioning technique in urban multipath environments.