RSS
Autonomous Vehicle Begins Drive From Italy To China
from the watch-out-on-the-roads... dept
It's been nearly five years since the famous DARPA "Grand Challenge" showed off some successful autonomous vehicles traveling 132 miles without a driver (this was a year after a similar attempt ended with no vehicles making it). The DARPA Urban Challenge followed, and not surprisingly, there's still plenty of research going on in the space. Slashdot points us to the news that two autonomous vehicles are now attempting to drive from Italy to China over the course of the next three months.
It's not quite the same as the Grand Challenge situation (which had no one in the car, and no lead vehicles or anything like that). In this case, each autonomous vehicle will be following a "lead van." Also, each of the autonomous vehicles will have a "driver" sitting in the driver's seat who is supposed to be able to quickly take over the vehicle should anything go wrong (which sounds like it might be a tougher job than you might imagine, as I would imagine boredom sets in pretty quickly). However, it will also involve real roads with real traffic. In other news, governments between Italy and China are recommending drivers stay off major roads for the next three months...
That said, the folks behind this project admit that "failure is part of the plan." They fully expect problems to arise, and part of the idea is to figure out where those failure points are, so they can work on correcting them. In fact, in a "test drive" before the caravan left, they already experienced problems when a car got between the lead van and the autonomous vehicle in a traffic circle.
It's not quite the same as the Grand Challenge situation (which had no one in the car, and no lead vehicles or anything like that). In this case, each autonomous vehicle will be following a "lead van." Also, each of the autonomous vehicles will have a "driver" sitting in the driver's seat who is supposed to be able to quickly take over the vehicle should anything go wrong (which sounds like it might be a tougher job than you might imagine, as I would imagine boredom sets in pretty quickly). However, it will also involve real roads with real traffic. In other news, governments between Italy and China are recommending drivers stay off major roads for the next three months...
That said, the folks behind this project admit that "failure is part of the plan." They fully expect problems to arise, and part of the idea is to figure out where those failure points are, so they can work on correcting them. In fact, in a "test drive" before the caravan left, they already experienced problems when a car got between the lead van and the autonomous vehicle in a traffic circle.
I have an idea...
Finding Co-Founders in a Recovering Economy
App Developer Conversations: Successful Fundraising Techniques
I have a business...
To Understand The Data Move It Closer
The Talent Wars: Today's Toughest Startup Challenge
I’m looking for research...
Geography Of The App Economy
The Growing App Market
Reader Comments (rss)
(Flattened / Threaded)
Cue XKCD ducks
[ reply to this | link to this | view in chronology ]
The "Grand Challenge" situation
[ reply to this | link to this | view in chronology ]
The U.S. is proposing...???
[ reply to this | link to this | view in chronology ]
Re:
[ reply to this | link to this | view in chronology ]
job
[ reply to this | link to this | view in chronology ]
solving the wrong problem
Driving over any terrain is a difficult task. Driving a vehicle that needs a road requires recognition, discrimination, judgment. Driving in traffic requires social instincts, the ability to make reasonably accurate guesses about what other drivers will do. Driving in a world where a wind-blown plastic bag is not very important but a running child must be avoided at all costs requires something close to animal intelligence.
If they really wanted to develop autonomous vehicles, they'd do it the smart way: start with the rovers they're designing for Mars, the dumb insectoid robots that can scuttle over anything. Then graduate to something that can exploit roads (and still doesn't care about running into things0. Then introduce random moving obstacles of all shapes, sizes and colors, including other robots, and teach the robots not to collide, maybe even to self-organize into efficient traffic patterns.... Operating on the same roads as human beings is a long way down the list.
[ reply to this | link to this | view in chronology ]
[ reply to this | link to this | view in chronology ]
[ reply to this | link to this | view in chronology ]
Add Your Comment