Performance Comparison of Deep Integration and Tight Coupling

Abstract

This paper presents a comparative analysis of tight coupling and deep integration. Three deeply integrated and two tightly coupled architectures are analyzed. The analysis of the different architectures is performed using the same grade inertial measurement units, satellite constellations, and GPS signal processing techniques (i.e., identical discriminators, coherent integration times, and C ∕ N0 ratio estimators). Three aspects of the architectures are studied. The first is the general performance improvement offered by deep integration over tight coupling. The second is the effect of using a federated filtering architecture instead of using a single, centralized filter. The third aspect analyzed is the performance of the algorithms when operated with and without vector tracking. The analysis shows that the architectures with vector tracking perform virtually identically. The algorithms using scalar tracking loops perform worse in comparison. However, all the architectures yield comparable performance when the C ∕ N0 ratio is below 29 dB-Hz. Copyright © 2013 Institute of Navigation.