The Colloquium Series of the Department of Computer Science, University of Wyoming presents Dr. Paul Oh Drexel University "Micro-Air-Vehicles for Flying in Near-Earth Environments" Thursday, November 18, 2004 ENG 1044 1:00 - 2:00 p.m. ABSTRACT Micro air vehicles are a new class of bird-sized aerial platforms. Called MAVs, the envisioned purpose is to have them act as robotic front observers that provide situational awareness in near-Earth environments like forests, buildings, caves and tunnels. Unlike contemporary miniature fixed-wing vehicles that fly in open air spaces, MAVs will be agile vehicles that can fly slowly and safely in cluttered environments. Making such airframes robotic demands sensor suites that can autonomously navigate the vehicle. MAVs have severe payload constraints, demanding the sensor suite be small, light and power-conscious. Additionally GPS, lighting and communications are often degraded in near-earth environment. As such, the sensor suite should operate in varied lighting and avoid distributed wireless computing. Flying insects, like honeybees, do not possess sophisticated visual or inertial measureents systems. Instead they avoid obstacles, regulate speed, compensate for wind gust, hover and follow terrain using optic flow. This sensing modality is simply the apparent visual motion experienced by an insect as it travels through the environment. Objects that are close will tend to appear to move faster than objects that are far away, and objects with which the insect are on a collision course will tend to appear as if they are rapidly increasing in size. Optic flow sensors, like those found in a computer mouse, are readily available. They often weigh less than 10-grams, are coin-sized, draw milliAmps and can interface with embedded microprocessors. This talk details proof-of-concept results applying optic flow sensors on MAV test beds. These 20-inch wingspan, 30-gram fixed-wing vehicles fly indoors at speeds as slow as 2 m/s in areas as small as 10x10 square meters. Demonstrations of autnomous avoid collisions, altitude regulation, take off and landing are presented. The results are promising and suggest broad impacts to all classes of unmanned aerial vehicles including lighter-than-air blimps, rotary- and flapping-wing vehicles. Bio: Paul Oh is an assistant professor at Drexel's Mechanical Engineering Department. He received his Bachelors in Mechanical Engineering (Honors) from McGill University (Montreal, Canada) minoring in Applied Mathematics in 1989 and thesis in control-configured aircraft. Awarded a government research fellowship, he received a Masters from Seoul National University (Korea) in 1992, with a thesis in adaptive control of electrohydraulic servomechanisms applied to gun-boat firing mechanisms. In 1999 he received a Ph.D from Columbia University for the visual servoing of a 5-DOF gantry hybrid robot for workcell monitoring. His current interests are in sensors and mechatronic design. In 2004, he received an NSF CAREER award for micro air vehicle control. He current chairs the IEEE Technical Committee on Aerial Robotics.