On insects, birds, and drones

Mini-workshop on insects, birds, and drones
Thursday 22 Sept, 11:00-12:00, room MED 1418

Visually guided morphing-wing flight in birds
Prof. Graham Taylor, Oxford University

Birds have been aptly described as “une aile guidée par un oeil”, but this simple summary belies their underlying complexity. Their wings are dynamic, morphing structures, supplemented by a morphing tail; their guidance is optimized through learning; and their eyes are just one part of a multi-functional visual system used for tasks including perching, prey pursuit, collision avoidance, and gap negotiation. In this talk, I will summarise what we have learned of these different aspects of bird flight by combining motion capture technologies, inertial measurement units, and precision GPS/GNSS loggers with physics-based simulation, video rendering, and machine learning. I will conclude by discussing prospects for future bio-inspired research on bird-like morphing-wing air vehicles.

On flying robots and insects: The tortuous story of optical flow control
Prof. Guido de Croon, Technical University Delft

Flying insects are capable of feats that are currently beyond the reach of flying robots. Tiny fruit flies are able to fly, avoid obstacles, navigate, find food or shelter, learn, and interact socially with one another – all with only ~100,000 neurons. It may come as no surprise that robotics researchers draw inspiration from nature to achieve autonomous flight of small drones. In particular, monocular vision-based control has received ample attention, as a single camera can be light-weight and power-efficient while providing rich information on the environment. It is well-known that flying insects make ample use of the monocular visual cue of optical flow, which captures the relative motion between the eye and objects in the environment. During my presentation, I will tell the tortuous story of optical flow control, starting with the biological finding that honeybees keep optical flow constant while descending – leading to soft landings without relying on height measurements. I will talk about the intertwined successes and failures of such a landing strategy on flying robots, and how robotics and biology can form a mutual inspiration for getting closer to understanding insect intelligence and creating similarly intelligent flying robots.