University of Pittsburgh team designing drone to herd Roomba vacuums
The challenge seems impossible.
Use a drone to herd a flock of Roomba vacuums while avoiding other Roombas with large poles attached to them.
And do this autonomously, flying and finding the Roombas without the help of a human pilot or GPS.
“You just push a button and it takes off and does its thing,” said Aaron Miller, a senior computer science and physics major at the University of Pittsburgh and one of the students trying to crack this challenge.
“That’s the idea, at least,” added Levi Burner, a junior majoring in computer engineering.
A team of more than 30 Pitt students in the Robotics and Automation Society is competing in the International Aerial Robotics Competition this year. The decades-old challenge pits universities against each other to use flying robots and drones to solve increasingly difficult missions. The missions are designed to push the limits of aerial robotics, autonomy and artificial intelligence. It can take years for a team to solve a mission.
The first mission in 1991 — in which an autonomous flying robot had to move a metal disc from one side of an arena to the other — went on for five years until Stanford University completed it. Carnegie Mellon University completed Mission 2, using an autonomous drone to map a simulated toxic waste dump, identify the waste and bringing back samples, after two years. Mission 4, which started in 2001 and required autonomous drones to fly 3 kilometers to an abandoned village, find a marked building and send a sensor probe into the building, took eight years to complete.
This was the fourth year for Mission 7. In it, drones have to herd Roombas by landing on top of them to slightly charge their course or dropping in front of them to change their direction 180 degrees. The competition was at the end of July at Georgia Tech in Atlanta. Another competition took place in China.
It was the Pitt team’s first time attending. The team took off and landed autonomously but nothing else. It didn’t herd Roombas. No teams at Georgia Tech did.
“Just taking off and landing took us six months,” Burner said.
Levi Burner, a junior studying electrical engineering at University of Pittsburgh, monitors a drone as it flies autonomously. Members of Pitt’s Robotics and Automation Society practice manuevers on Saturday, September 23, 2017 at Benedum Hall in Oakland in preparation for the International Arieal Robotics Competition. (Photo by Jack Fordyce | Tribune-Review)
The Pitt team scored the most points out of all the teams at Georgia Tech. Judges awarded the team Best System Design.
The Pitt team has high hopes for next July and its second crack at Mission 7. They are custom building a drone with more motors and sensors. Team members are working on Roomba detecting technology that will help the drone find and interact with the robot vacuums. The team is also working on its kill switch in case something goes wrong.
Aaron Aupperlee is a Tribune-Review staff writer. Reach him at [email protected], 412-336-8448 or via Twitter @tinynotebook.