In what has to be the best college class ever, students at MIT learned how to program algorithms for self-driving cars during a four-week Independent Activities Period (IAP) course called Rapid Autonomous Complex-Environment Competing Ackermann-steering Robot (RACECAR).
The vehicles used in the RACECAR course featured a high-performance NVIDIA Jetson Tegra K1 embedded computer, a rich sensor suite to perceive the environment, the open-source Robot Operating System (ROS), and student-developed autonomy algorithms to navigate MIT’s tunnels.
The mini cars happened to be on a 1:10 scale, and the students were tasked with building the fastest robotic car possible. The winning vehicle had a near-perfect run of 49.64 seconds at an average speed of 7.1 mph. Three out of four teams successfully completed the 515-foot race course.
Check out the video of the winning vehicle below. And if you want to see a self-driving car crash repeatedly (who doesn’t), skip ahead to the 2:40 mark.