Important Innovations Collection

Physicists Train Autonomous Gliders to Soar Like Eagles Source:  Woods Hole Oceanographic Institution Photo Montage New Navigational Strategies for UAV's Physicists at the University of California San Diego are using reinforcement learning, a type of machine learning,  to train gliders to autonomously fly like birds.  Some birds like eagles and hawks soar rapidly to great heights with ease.  The gliders are being trained to do the same.  The scientists expect this approach of reinforcement learning - learning from performance and results - will lead to new artificial intelligence for UAV navigation. Riding the Wind Upward currents of warm air assist the birds' rapid, high ascents.  The scientists use reinforcement learning to train the gliders to autonomously ride "atmospheric thermals" or currents of warm air to soar as high as 2,300 feet.  The results provide a new navigational strategy for the development of UAV's, or unmanned aerial veh...

Greater Energy Efficiency Yale's Drone with Landing Gear Nature Inspired UAV Researchers at Yale University have developed a UAV that uses robotic talons inspired by birds.  The landing gear allows the drone to perch and rest in places that would normally be inaccessible.  The drone can attach to ledges and even hang upside down.  The landing gear allows the drone to rest and turn off its motor while still gathering information with its sensors and cameras. Purpose The purpose of the robotic talons is to enable the drone to conserve battery power and increase its energy efficiency.  A number of different talons can be attached to the bottom of the drone to land in different places and on different surfaces. Practical Applications The nature inspired talons give the drone far more access in the environment that it's working in and enable it to perform many more tasks.  The Yale team continues to develop the drone.  They are likely to empower it ...