Stanford's 150 mph Racecar Drives Autonomously

Chris Gerdes says that for people to let an autonomous car drive them around the car should be at least as good as the best human drivers. Gerdes and his team at Stanford are developing autonomous racecars capable of 150 mile per hour speeds.

Where a professional driver uses all of his abilities and the abilities of the vehicle to drive as fast as possible, Chris wants to use all of these skills to avoid accidents.


http://www-cdr.stanford.edu/dynamic/bywirediag/Image/P1.jpg

The P1 is an electric vehicle fully built by students that can drift race. Using its rear-wheel drive and front wheel steer-by-wire the P1 can drift around corners and take tight curves while adapting to different surfaces.

Gerdes can take a map of a racetrack and a math model of a car and find the fastest way around the track using iterative solutions. Human drivers don't have the benefit of an algorithm testing the tradeoffs between going as fast as possible and keeping the car on the track.

To understand how the brains of racecar drivers operate, the team hooked electrodes to the drivers' heads. The test was based on the principle that alpha waves in the brain are generated when the brain is at rest and theta waves are generated when visual processing and decision making occurs. Testing found that much of track driving requires hard theta wave decisions but there are several instances when a driver goes on instinctive reaction alone.

This was the inspiration for Chris and his team to plug some intuition into the autonomous vehicles. The long term goal is now to give vehicles some components of reflexive involuntary actions when driving. Going philosophically deeper Gerdes asks us if we might want more from our car than just driving us around. This new possible balance of man and machine can hopefully be used to help develop some part of the human potential.


http://www.ted.com/talks/chris_gerdes_the_future_race_car_150mph_and_no_driver.html