Successfully navigating roundabouts and obeying stop signs despite having no hands on the wheel, a team of Northeastern University robotics students won a self-driving car competition at the American Control Conference, a premier conference in robotics and automatic control. 

The team included first-year students Kian Behzad and Rojin Zandi, rising second-year student Elehah Motamedi, and Dinesh Murugan, who is in the final year of a master’s program. 

Milad Siami, assistant professor of electrical and computer engineering at Northeastern and faculty adviser to the team, says that he was “immensely proud” of “truly exceptional students.”

“This victory is a shining example of our team’s intellect, effective collaboration, and their relentless passion for pushing boundaries in autonomous systems,” Siami says.

Northeastern beat out Purdue University, Arizona State University and the University of South Florida for the top prize.

“We had no idea what other teams were doing … so we were completely blinded,” says Behzad, the team leader. “So, we had to do it with everything we had so we didn’t have any regrets at the end.”

The concept of control means automation, or to make a system do what you mean it to do—to behave as you would like—without manual manipulation, Behzad explained. 

The American Control Conference was a three-day event for the American Automatic Control Council, the U.S. national member organization of the International Federation for Automatic Control. 

The annual event, which was held this year in San Diego, California, is sponsored by numerous professional engineering and scientific groups including the American Institute of Aeronautics and Astronautics, the American Society of Civil Engineers, the International Society of Automation, and the Society for Industrial and Applied Mathematics

The students participated in a competition to successfully navigate a scaled-down, self-driving car through an unfamiliar, mini city complete with traffic infrastructure. The vehicle was provided to all participants by Quanser, a company that sponsored the competition, and it was the size of a typical remote-controlled car. But there was nobody with a remote. Instead, the vehicle had to map the city and surroundings, identify and appropriately respond to various traffic infrastructures. 
The preparations began months before, and the competition was held in three phases.

In the first phase, the students spent months developing coding algorithms that would enable the vehicle to navigate a virtual city on a computer simulator. To do this, the team built their own data set of traffic signs and landmarks and trained the model car to identify and respond to them. 

“Everything, we coded ourselves,” Zandi says. 

The second phase of the competition moved from the virtual to the real world—as the team had five or six hours to prepare for a test run on a physical model of the virtual city.

“We had to show them that we could do the same thing in the real world, and if we showed them, then we qualified for the third phase,” Behzad says.

In the final phase, the model city was completely rearranged. The team had a day to prepare.

“They completely altered the map,” Zandi says. “They changed everything, even the lighting. They made the city very dark … to make it harder to detect things.”

But again, the test run was successful. The team presented their results to the judges.

“They announced a winner, and at the end it was us,” Behzad says. 

The students said the competition was a valuable experience.

“I want to start writing a paper on this concept and some of us want to continue in this field,” Motamedi says. 

Cyrus Moulton is a Northeastern Global News reporter. Email him at c.moulton@northeastern.edu. Follow him on Twitter @MoultonCyrus.