Ariel Perez-Rosado

University of Maryland graduate student Ariel Perez-Rosado holds the robotic bird with wings that flap independently in Cole Field House. (Kenneth K. Lam, Baltimore Sun / June 7, 2013)

Robo Raven is making aviation history – again.

The robotic bird's maiden flights in December were a first in the history of flying machines, but lasted only seconds inside the Reckord Armory at the University of Maryland, College Park. Now the mylar and carbon fiber contraption is back, soaring higher, longer and more accurately — the first machine ever built that flies on wings that can move independently of each other, as real birds' do.

"Nobody has flown anything with independent wing control," before, said S.K. Gupta, a professor of mechanical engineering in the A. James Clark School of Engineering at College Park. He's been working on the project with his fellow mechanical engineering professor, Hugh Bruck, and their students.

They've been at it for nearly 10 years, starting with a one-motor bird with wings flapping simultaneously and moving on to a much greater technical challenge: a two-motor version with independent wings. Independent wings represent a big step ahead, making for a more nimble flier, potentially capable of speed as well as quick bird-like turns, dives and subtle in-flight wing adjustments to the wind.

In pursuit of this Holy Grail, the researchers have crashed many birds — perhaps 50 to 100 in all. Several were attacked and damaged in flight by hawks.

This August, Gupta and Bruck will present the first paper on the project to a conference in Portland, Ore., organized by the American Society of Mechanical Engineers.

Meanwhile, the work and the flying go on.

At Cole Field House at College Park on a recent warm weekday afternoon, mechanical engineering doctoral student Ariel Perez-Rosado launched the bird from the stands as if he were flying a paper airplane. His fellow doctoral student, Luke Roberts, stood on the arena floor working the toggle controls of a radio transmitter/receiver: one stick for the tail, one for the wings.

Robo Raven made slow circles through the air under the 100-foot ceiling, the flapping of the delicate mylar wings sounding like an umbrella quickly opening and closing. From a distance, it was a bird up there. You could be fooled, as apparently the hawks were when they attacked.

In several flights in the course of an hour, the bird reached heights of some 70 feet, staying aloft for a minute or so at a time, making quick turns, a dive, and graceful landings, but also hitting the artificial turf hard a few times. Once it flew into the face of a member of the engineering school's public relations staff.

Still, that was a big improvement since the quick Dec. 21 flights. The bird wasn't getting much lift then, and the team wanted to keep the flights short — 12 to 15 seconds — to avoid crashing within the limited indoor space. Even at that, the bird at one point slammed into a stack of chairs, said Roberts, who was there.

The repairs weren't too difficult, he said in an email, but "we wanted to figure out a better way to attach the wings, which took a few months to successfully implement."

It was part of years of laborious trial and error as team members shuttled between field flight tests and the laboratory. In hopes of speeding up wing design tests and trimming the bird casualty rate, Bruck developed a laboratory platform for the two-motor version that allowed the team to measure aerodynamic forces generated by the wings.

The challenges were compounded with the two-motor model, which meant much more weight, not just for the motors, but a bigger battery and onboard programmable controller to synchronize the wings. At first, the bird was too heavy to fly.

Using lightweight polymer materials, 3D printing and precision laser cutting, they trimmed the bird down to 291.6 grams, or roughly 10.3 ounces. That's much lighter than a real raven, which usually weighs more than two pounds. But as it turned out, the other robot specifications — wing flapping frequency, 44-inch wingspan, 24-inch length — were very close to the real thing. Hence, the machine was dubbed Robo Raven.

After the first baby flights in December, several more attempts failed. From winter through spring, the team kept at it, improving the wings and the weight balance, Gupta said. By April 19, it was ready for another go inside the armory.

"April's successful flight was actually quite unexpected, simply because we had tried so many designs that hadn't worked that well," said Roberts. And yet, this version of Robo Raven made several flights of 30 to 40 seconds.

The longest flight so far has been 3 minutes, 45 seconds, Gupta said, but they have to improve on that if Robo Raven is to be of any practical use. The robotic bird could potentially be used for surveillance, inspecting agricultural fields and storm damage, environmental study, and for studying the behavior of real birds. Recent research has been supported by Army and Air Force grants, but Gupta declined to speculate on potential military uses.

"We are not involved in any of those things," he said. "So we stay away from those discussions."

Doctoral student John Gerdes, who has been working on the project since 2008, is still thinking about the potential improvements ahead.

"If you look at a turkey vulture when it's soaring, making those little adjustments when it hits a thermal. If we could do that it would be awesome," he said. "That's kind of the grand vision of the future."

arthur.hirsch@baltsun.com