Reinforcement Learning Helps Gliders Find Thermals

Sep-23-18 Robotic gliders able to autonomously locate and ride on thermals could allow the crafts to soar over long distances like birds.

To offer the gliders the same navigation advantages as birds, the team from UC San Diego turned to reinforcement learning, a form of machine learning in which the ‘agent’ learns from experience. Armed with that framework, the team programmed the gliders’ flight to include to a specific pitch and banking angle, and outfitted the gliders with an onboard computer that recorded the gliders’ experiences in the air. That data was then used to compile a navigational strategy that could be applied to larger gliders.

According to Professor Massimo Vergassola, “Our results highlight the role of vertical wind accelerations and roll-wise torques as viable biological mechanosensory cues for soaring birds, and provide a navigational strategy that is directly applicable to the development of autonomous soaring vehicles.”

More Info about this Invention:

[NEWATLAS.COM]
[UCSDNEWS.EDU]

---