Probabilistic graphical fusion of LiDAR, GPS, and 3D building maps for urban UAV navigation

Because of the recent interest in unmanned air vehicle (UAV) commercialization, there is a great need for navigation algorithms that provide accurate and robust positioning in urban environments that are often Global Positioning Systems (GPS) challenged or denied. In this paper, we present a probabilistic graph-based navigation algorithm resilient to GPS errors. Fusing GPS pseudorange and Light Detection and Ranging (LiDAR) odometry measurements with 3D building maps, we apply a batch estimation approach to generate a robust trajectory estimate and maps of the surrounding environment. We then leverage the maps to locate potential sources of GPS multipath and mitigate the effects of degraded pseudorange measurements on the trajectory estimate. We experimentally validate our results with flight tests conducted in GPS-challenged and GPS-denied environments.