While there's a great deal of excitement surrounding the concept of autonomous delivery drones, the aircraft would likely all utilize GPS to navigate – and GPS satellites aren't always available. That's why Prof. Martinez Carranza has developed a new drone navigation system, that's based purely on visual observations.
Carranza is a researcher with Mexico's National Institute of Astrophysics, Optics and Electronics (INAOE). He developed the system as part of the RAFAGA (Robust Autonomous Flight of unmanned aerial vehicles in GPS-denied outdoor areas) project during his postdoc at The University of Bristol.
Users of the system start with a Google Maps-like satellite image of the area that they want the drone to traverse, and draw a flight path on that map – not unlike the autonomous flight software currently used by many personal drones.
Whereas those aircraft proceed to follow GPS waypoints, however, a RAFAGA-enabled drone uses its onboard camera to identify buildings or other features on the ground, matching them up with their aerial images on the satellite map.
Not only is the technology reportedly more reliable than GPS, but the required hardware (cameras, accelerometers and gyroscopes) is also simpler and less expensive.
Source: Investigacion y Desarrollo (Spanish)