GNSS/INS Integration Methods

Printer-friendly versionSend by emailPDF version
TitleGNSS/INS Integration Methods
Publication TypeThesis
Year of Publication2011
AuthorsANGRISANO, A
Academic DepartmentDepartment of Applied Sciences
DegreePhD
UniversityParthenope
CityNaples
Abstract

In critical locations such as urban or mountainous areas satellite navigation is difficult, above all due to the signal blocking problem; for this reason satellite systems are often integrated with inertial sensors, owing to their complementary features. A common configuration includes a GPS receiver and a high-precision inertial sensor, able to provide navigation information during GPS gaps.Nowadays the low cost inertial sensors with small size and weight and poor accuracy are developing and their use as part of integrated navigation systems in difficult environments is under investigation. On the other hand the recent enhancement of GLONASS satellite system suggests the combined use with GPS in order to increase the satellite availability as well as position accuracy; this can be especially useful in places with lack of GPS signals.This study is to assess the effectiveness of the integration of GPS/GLONASS with low cost inertial sensors in pedestrian and vehicular urban navigation and to investigatemethods to improve its performance.The Extended Kalman filter is used to merge the satellite and inertial information and the loosely and tightly coupled integration strategies are adopted; their performances comparison in difficult areas is one of the main objectives of this work. Generally the tight coupling is more used in urban or natural canyons because it can provide an integrated navigation solution also with less than four satellites (minimum number of satellites necessary for a GPS only positioning); the inclusion of GLONASS satellites in this context may change significantly the role of loosely coupling in urban navigation. In this work pseudorange and Doppler measurements are processed in single point mode; hence no differential processing is performed and no base station is necessary.For bounding the MEMS INS errors during GPS/GLONASS outages, the use of motion constraints is introduced and evaluated. To carry out the analysis a complete navigation software has been developed in Matlab environment.The results show that the integration of GPS/INS provides satisfying performance only in good visibility conditions in both loose and tight architectures, while during GPS outages (longer than few seconds) very large errors are experienced; including GLONASS in GPS/INS integration produces significant solution improvements. As expected the tight integration shows better performance during GNSS outages, but the results obtained with GPS/INS integration in tight architecture and 3GPS/GLONASS/INS integration in loose approach are often similar, suggesting the use of the latter configuration owing to its relatively simple implementation. The motion constraints provide remarkable enhancements of the navigation solution, with RMS position error of a few meters in difficult urban canyons for both loose and tight architectures.

How to citate: 
ANGRISANO, A (2011). GNSS/INS Integration Methods. Department of Applied Sciences Parthenope