Biomimicry is the study of nature to help solve human challenges. It seeks to emulate patterns and behaviours existing in nature and applying them for the design and production of man-made materials and structures.
Examples include the modelling of wind turbines after the fin of humpback whales and designing the nose cap of the Japanese bullet train to resemble the beak of the Kingfisher bird.
More recently, researchers from the University of California have applied the gliding techniques used by birds to help a drone fly and gain height. To better understand how birds glide through the air using thermal currents, the researchers applied machine learning techniques. The drone processed a large amount of input data to determine what actions to take to allow it to glide in the air for as long as possible. Input data included the change in upward wind strength and the impact to the drone both in terms of its vertical and horizontal lengths. Sensors took these measures ten times per second.
According to the results of the study, the drone was able to understand how to optimize its position and catch warm updraughts after only 15 hours of flight. This was done without any information being fed to the drone to direct its actions. These results not only indicate that machines can learn strategies quickly and in real-time, but that soaring birds are able to either perform complex planning calculations or rely on additional navigational cues.
This research provides clues as to what might be possible with the next generation of “smart” navigation which could be used for drones as well as for autopilot systems used for conventional aircrafts. It could also lead to futuristic autonomous gliders, powered by solar energy as recently developed by Airbus, to provide an alternative to surveillance or communication satellites.