Robotic Audi TTS Hits 120 mph On Track

By Keith Buglewicz | August 21, 2012
There's a lot of news out there these days about autonomous cars. Google has them, Cadillac is experimenting with them, and supplier Continental has even let an autonomous Volkswagen Passat drive itself across the country. The future is almost here, and it's apparently behind the wheel. But there's still a prejudice against self-driving cars that says they'll be slow-moving, slow-reacting, and definitely needing a hand just above the wheel when everything goes kablooey. While we'll admit we still want a human nearby to take command in case the computer system driving the vehicle suddenly goes BSOD, we're coming around to the idea that an autonomous car doesn't have to be slow. In fact, it can be downright fast. Stanford University has been working on a self-driving Audi TTS sport coupe for some time now; we first reported on it back in November, 2009. The team has made great strides since then, and recently demonstrated its creation pounding out fast laps at Thunderhill Raceway Park outside San Francisco. Known as "Shelly," the autonomous Audi hit speeds of 120 mph on the track, and posted lap times within a few seconds of professional drivers. OK, so a bunch of eggheads made a fast car go fast by itself. So what? The practical applications of what Stanford--and corporate partner Volkswagen--are up to go well beyond a publicity stunt. "The basic idea here is also applicable to safety systems," said Stanford associate professor Chris Gerdes, leader of the Dynamic Design Lab that built the car. "If we can have cars that will drive up to the limits, and recover if they go past, this is something that could actually help ordinary drivers, for instance on a slippery road." So should professional race car drivers hand in their helmets? Not quite. While quick, the autonomous Audi TT doesn't have the smoothness of a human driver. People can use intuition and prior knowledge of a track or how a car behaves to smooth out transitions between braking, turning, and acceleration, thus giving our squishy brains a distinct advantage. But that doesn't mean the Stanford team is giving up on making its car quicker than people. Last weekend, it wired up two drivers at the Rolex Monterey Motorsports Reunion to see how their bodies react on the track. The goal is to use the data to smooth out the autonomous car's reactions, and make it faster. That's an ambitious goal, and the team has already come pretty far. Check it out for yourself below.

Source: Stanford, YouTube