Mark My Words

The third DARPA Challenge was this weekend. Three years ago, the Defence Department gave a $1 million prize, a challenge to anyone who could build a robotic car that could navigate 160 miles through the desert -- with no driver, no remote control, no "wires in the road" -- just on-board computing power and on-board sensors.

That year (2004), no vehicle made it more than five (5) miles.

The next year (2005), five vehicles made it through the complex course, at average speeds over 45 miles per hour.

Then DARPA offered about $2 million in prizes, but moved the contest to a simulated urban environment. Stop signs. Other vehicles. Simulated pedestrians.

This year (2007), three vehicles made it through under the time limit, and eight more made it through without incident. That's right -- almost no crashes or problems at all. The cars were so successful, that observers found it boring to watch. The cost of all of this research to the American taxpayer? About $3 million. And keep in mind that this was WITHOUT any supervisory computer control, which would make the system faster, more efficient, and safer.

I've been saying this for years, but let me get on the record again. No, we won't have fully robotic cars on our streets in 10 years. But in 10 years, most people will agree that robotic cars are coming. Ten years after that, automated vehicles will be on the roads somewhere in the world. Ten years after THAT, most vehicles in the developed world will be automated. If there are vehicles (outside of theme parks) that humans can drive, there will be so many computer-controlled safety features as to make the human's actions more like suggestions than actual commands.

And I can't wait. The idea of putting a 17-year-old in charge of 2000 lbs of metal, tearing down highways and back roads at 80 miles per hour, scares the crap out of me. But we will live to see a time when that's a thing of the past.

http://news.yahoo.com/s/nm/robot_cars_dc;_ylt=Aj8cZ5Ud7EYxW1V37sDjRqIjtBAF