Sensor fusion for unmanned aircraft system navigation in an urban environment

When unmanned aircraft systems operate in urban corridors, navigation accuracy is a priority due to proximity of buildings, obstructions, and other infrastructure. In most environments a Global Positioning System (GPS)/inertial measurement unit combination along with an air data system can provide accurate navigation capability. However, this is not possible in urban corridors where GPS has well-documented degradation. Other sensors such as vision-based systems and Long-Term Evolution transceivers have shown to be useful in urban settings, but modeling them individually is difficult without an in-depth understanding of each sensor and the factors dictating its accuracy. This paper proposes a framework to model location-dependent accuracy of navigation and how this changes within the urban environment. Results show that persistent machine vision can provide accurate navigation capability, but LTE with its current measurement delay does not have a noticeable positive effect on navigation accuracy.