Computer Sciences alumnus David Hoffert (BS'2008), now a graduate student at Stanford University, is working on autonomous vehicles that are able to drive and avoid accidents better than most humans.
A key application is vehicle driver assistance. Hoffert's current work, as part of an autonomous rally racecar driving team, is concerned with mapping — finding the optimal path for a racecar to follow on a racetrack.
A key question is how to find the optimal path given only limited information about the racetrack; for example, only that part of the track and its surroundings that's visible from the car.
From first principles, David derives navigation rules that are couched in physical intuition of the dynamics of the system. More specifically, he has been working on mathematical functions to construct a good path, which formed the basis of his software for the autonomous vehicle featured in the video at http://www.youtube.com/watch?v=-CuVVZq9GfY
The team's goal for next summer is to drive their vehicle autonomously up Pikes Peak in Colorado. By the end of his time at Stanford, David hopes to enter thevehicle in the annual Race to the Clouds — and win!
More information on David and his work can be found at http://www.stanford.edu/~hoffert/