We have all seen many times how birds land on the cornice: at the last moment they spread their wings wide and, like on a parachute, fall to the right place. But what is naturally and easily obtained from living beings is practically inaccessible for flying vehicles. Nevertheless, researchers at the University of Bristol believe that they "taught" the drone to perform this difficult bird maneuver.
In normal flight, a fixed-wing aircraft overcomes air resistance. Designers try to minimize the drag coefficient to make it easier for the aircraft to fly. However, when approaching the plane, it is necessary to slow down, for which it increases the resistance of the oncoming flow by deflecting it back. If the deflection angle is too large, the aircraft may lose lift and stall.
Scientists from the University of Bristol have developed adjustable tilt wings that provide the drone with the necessary lift during takeoff and braking during landing. To do this, they equipped their drone with the appropriate machine learning algorithms, which brought it closer to their natural prototype - birds.
Thus, the researchers proved that the combination of wing morphing with machine learning will help UAVs master the "bird" landing technique.