Berkeley Team Successfully Demonstrates Flying Unmanned Monitoring System

Assistant Professor Raja Sengupta describes Berkeley's UAV system to Congressman Kurt Weldon at the test site in Arizona.

On Wednesday August 20th, the faculty and students of the CEE Systems program and the ME control group demonstrated Berkeley's ability to autonomously control unmanned air vehicles (UAVs). There were two firsts. For the first time a UAV autonomously navigated a road based on machine vision. The demonstration took place in the desert, just outside Tucson, Arizona. The system worked in sunshine, cloud, and rain. For the first time a team of UAV's followed a moving bus, keeping watch over its surrounding area, and streaming video of the area into the bus. The demonstration was attended by Congressman Kurt Weldon, Vice-Chairman of the House Armed Services Committee, as well as scientists from MIT, UCLA, and a variety of other research laboratories.

Unprocessed view of the test road from the UAV camera. Inset shows the result after image processing to find the road and lane boundaries.

UAV's have attracted considerable interest as a viable way to collect environmental data from remote places, aerially monitor traffic conditions, protect military convoys, patrol borders, or track forest fires. Most of these applications have not been realized because the UAV's are unable to autonomously localize and navigate themselves with respect to roads, around obstacles or other terrain features. The demonstration showcased technologies that are a critical advance towards making these applications into realities. This research is undertaken by the ONR-UCB Center for Collaborative Control of Unmanned Vehicles.

Lab Notes Article on the UAV Demonstratrion

UC Berkeley CEE Systems Engineering Program

UCB Mechanical Engineering Control Group