When I was 19, I was in Rhode Island attending college while trying to figure out ways to make the radio in my '87 Honda Prelude louder. I wound up being forced to sell that car when I got bad grades my first semester and nearly lost an academic scholarship. Romanian student Ionut Budisteanu is doing quite a bit more than I, and probably a bit more than most of you as well. Budisteanu recently competed in the Intel International Science and Engineering Fair, where he won the top prize of a $75,000 scholarship. His focus? Making autonomous car technology more affordable.
As it stands, autonomous car tech is still rather expensive. The fancy high-resolution 3D radar used by Google, for example, runs in the range of $75,000. Budisteanu says his system can be created for around $4,000.
How does he do it? By replacing the high-resolution system with one that utilizes a much lower resolution. He has created an artificial intelligence program that can pick out curbs, lane markers, and all the other obstacles that drivers face on the road. That data is processed through a webcam while a low-resolution radar system handles the bigger objects such as cars and houses. All of the data collected is sent to a computer that is in charge of the system, and then decides on how the car should behave as it rolls driverless-ly through town.
In 50 simulations run through the system, the autonomous vehicle performed perfectly 47 times. Those three other times might scare you, but Budisteanu is already working to sort that out with a slightly higher-resolution radar system. He says the price will climb no higher than $4,000. The intelligent young man has attracted great attention through this competition. Budisteanu has received funding from a Romanian company that will allow him to begin working on a prototype.
Meanwhile, I'm still trying to figure out how to make the radio in my current car louder...