Computer Vision - Feature-Based Nav - Fig N

Feature-Based Navigation

A capability of autonomous flight in the national airspace assumes unmanned aerial system (UAS) knowing where it is all the time. With a GPS (or even a DGPS) sensor onboard this task seems very easy to be achieved. However, in GPS-degraded or GPS-denied environment accurate navigation becomes a real problem. The research efforts in this area include processing the onboard electro-optical (EO) / infrared (IR) data and geographical maps (the data bases of geographical features) to provide a continuous estimate of UAS relative position with respect to the map (the map features). To this end, Fig.N shows an example of the preliminary results allowing mapping geographical data about two airports UAS operates around with video data. In this case, once airport is identified a feasible landing maneuver is computed and executed to bring UAS onto the final approach to the chosen runway (picked by the analysis of winds direction).

Figure N. Identifying airport and bringing UAS to a final approach leg.