Learning to Fly High: Reinforcement Learning for Soaring and Airborne Wind Energy

Explore the fascinating world of soaring flight and airborne wind energy through this insightful 33-minute conference talk by Antonio Celani. Delve into the complexities of aerodynamics in turbulent atmospheres and discover how reinforcement learning can be applied to develop near-optimal control strategies for both bird-like soaring and kite-based power extraction. Gain a comprehensive understanding of the subject through topics such as naturalistic observations, virtual bird simulations, kite cycles, and prototypes. Examine numerical simulations, dynamical system analysis, and future prospects in multi-agent navigation. Engage with schematic representations and participate in a thought-provoking discussion on this cutting-edge intersection of biology, physics, and artificial intelligence.

Introduction
Naturalistic Observations
How do they do it
Reinforcement learning
Pseudocode
Virtual Birds
Platformers
Kites
Kite Cycles
Prototypes
Numerical Simulation
Dynamical System Analysis
Future work
Multiagent navigation
Questions
Schematics
Preview Slide
Discussion