Koray Celik, a doctoral student in computer engineering, has designed a computer processor which runs an autonomous helicopter. Photo: Chris Potratz/Iowa State Daily
By Heidi Ebert — Daily Staff Writer
Published: Tuesday, September 29, 2009 11:12 PM CDT
For one ISU student, a personal medical condition became the inspiration for his doctoral program.
Koray Celik, graduate student in electrical and computer engineering, suffered from vertigo, “the sudden sensation that you are unsteady or that your surroundings are moving,” according to Mayo Clinic Web site.
“What happens is the balance organs [in your ear] stop working and your head is not able to sense acceleration,” Celik said. “Now, most people will think that’s not a problem, but it is, because when you cannot sense that, you won’t know if you are falling. So when that happened for a couple months, I actually had to depend on my eyes for everything, including getting to the university and back and attending my classes.”
Vertigo was the reason Celik credits for inspiring him to think about how humans perceive the world around them, using only eyesight. The result of his thinking is a small helicopter that can navigate its own way around a building, using only one camera.
The project’s name is Vision Guided Autonomous Navigation for Unmanned Aerial Vehicles. Vision guided navigation is not the only feature of this high-tech helicopter — it is also capable of producing a floor map of the building it is in. Unlike human eyesight, however, it does not need to go into every room to produce this map. The navigation and the floor mapping together is called SLAM, or Simultaneous Localization And Mapping. It was designed for a military project for the Rockwell Collins Corporation.
The helicopter also contains facial-recognition technology. This is the same type of technology Apple has advertised for their new iPhoto software, only this facial recognition is conducted in a three-dimensional setting. The helicopter can hover in front of someone and construct an image of the person’s face, which is then run through a database in order to determine who the person is.
There are numerous applications to the project. The facial recognition technology could also be used to secure buildings and run background checks on all people inside. The technology could also be used in the military.
“Right now, we are working on a proposal with the U.S. Army, and they are planning to put the system on helmets, like a soldier’s helmet,” Celik said. “If you are walking into an unknown building, it will make sure you can get out. It will map the building for you.”
Celik chose to put the system on a helicopter first because he has always had a fascination with spinning mechanics. He recalls a story his dad told him of when he was younger.
“In 1982, when I was just a little baby and could barely walk, my dad took me to a grocery store. He asked me what I want ... I actually crawl on the ground and point at the refrigerator’s spinning fan. I said ‘that,’” Celik said. “I just have a fascination with spinning mechanics. I can’t explain why.”
Celik plans to complete the helicopter in two years, even though he graduates with his Ph.D, in the next year.
Fun facts about the helicopter
It is registered with IT because it has Wi-Fi and therefore had to have its own IP address, like a personal laptop.
It can check e-mail.
It has facial-recognition technology on it, which means it can recognize someone and complete a background check on that person just by “looking” at them.
It has a 1.2-horsepower motor.
It weighs only two pounds, at a maximum, because it has to be light enough to fly.
It has three radios. One is an emergency override, in case something goes wrong and it has to be controlled by remote. Another acts like a radio transmitter. The other is a Wi-Fi antenna.
It can fly up to 45 mph.
It has 11 control systems.
It can record video on a 4GB hard drive.
How it works:
The computer in the helicopter looks for features or “high energy points.” The points are basically areas of extreme contrast, such as an orange cord on a black carpet, or a red light on a fire alarm. The computer then registers hundreds of the points.
After the computer identifies the points, it cross-references them with its software, which contains an “architectural lines algorithm.” The algorithm is a program that allows the helicopter to “see” lines, such as the lines around a door or the line in between a floor and the wall. The helicopter then builds a map of the environment by overlapping the points and the lines. Distances are measured based on how these lines and points move when the helicopter moves.