According to the scientific method, an experiment must begin with a problem. For participants at the Intel International Science and Engineering Fair, one of the best-known student competitions in the country, their experiments usually begin with a big problem.
For Ionut Budisteanu, the Romanian high school student who won the $75,000 Gordon E. Moore award at this year’s fair, his experiment started with two big problems: the vast majority of fatal and severe car accidents are the result of human error, and self-driving car technology, which many think can reduce the number of these accidents, is still very expensive. That is, it was very expensive before Budisteanu invented a self-driving car that is about $70,000 cheaper than Google’s version.
“Because I am a high school student and I was trying to create a homemade self-driving car, I couldn’t [use] the same technology [as Google].” Budisteanu said in an interview. “They use a very expensive 3D Lidar [a type of radar], so I decided to try to remove that part as much as possible.”
Although his initial attempts to eliminate the 3D Lidar entirely only yielded a 60-70 percent success rate, Budisteanu was able to significantly raise the success rate to near-perfect by using a much cheaper lidar system to recognize larger objects, and cameras and software to recognize curbs, signs and lane markings. Budisteanu’s innovation is the kind that often emerges at student competitions, where students who don’t have the deep pockets of companies like Google have to come up with a low-cost, and possibly easier-to-replicate solutions.
Budisteanu’s self-driving car is not only cheap to produce, it also gets more reliable when more people use the software, the 19-year-old says. It’s programmed to enter things the cameras recognize, like street signs, into a database which allows other cars to identify signs with more ease and confidence.
In addition to the prize money, Budisteanu’s experiment also gave him new connections in the world of computer science and Artificial Intelligence, which he plans to study during and after college.
While “computer science is a new theory,” Budisteanu said, “the first generation of experts in computer science has changed [society] and become a research tool for other sciences.”
Budisteanu hopes to continue producing innovations in artificial intelligence as a university professor and researcher. For now, though, the high school senior who has been to 130 national and 18 international science competitions is focusing on something else: graduating.
His current project is called the “graduation exam,” he quipped. “I need to study for it.”
A version of this news article first appeared in the Digital Education blog.