<<So we have developed technology for cars that can drive themselves. Our automated cars, manned by trained operators, just drove from our Mountain View campus to our Santa Monica office and on to Hollywood Boulevard. They’ve driven down Lombard Street, crossed the Golden Gate bridge, navigated the Pacific Coast Highway, and even made it all the way around Lake Tahoe. All in all, our self-driving cars have logged over 140,000 miles. We think this is a first in robotics research.
Our automated cars use video cameras, radar sensors and a laser range finder to “see” other traffic, as well as detailed maps (which we collect using manually driven vehicles) to navigate the road ahead. This is all made possible by Google’s data centers, which can process the enormous amounts of information gathered by our cars when mapping their terrain.>>
http://googleblog.blogspot.com/2010/10/what-were-driving-at.html?utm_source=feedburner&utm_medium=feed&utm_campaign=Feed:+blogspot/MKuf+(Official+Google+Blog) |
Wow!
I've been aware of the DARPA contests for self-driven vehicles, of course, but had no idea anyone was working on the problem from this direction.
What Google is doing differently is to build a computerised model of the road, instead of making the car work out the whole thing from its realtime sensors. It's a natural way for them to do it, I suppose, given their experience with rangefinding sensors on their street view cars.
Now that the technology has reached the point of a successful "concept demonstration", it's inevitable that we will be using self-driven vehicles for real in the near future. Maybe in 15 years, 20 years or 25 years, but not possibly any longer. |
Funny, I've read something like that yesterday. Self driving car tested in traffic in Germany, the University of Braunschweig.
http://www.nu.nl/auto/2352057/auto-stuurt-zelf-straten.html
http://translate.google.nl/translate?js=n&prev=_t&hl=en&ie=UTF-8&layout=2&eotf=1&sl=nl&tl=en&u=http%3A%2F%2Fwww.nu.nl%2Fauto%2F2352057%2Fauto-stuurt-zelf-straten.html |
The Braunschweig car needs the driver to press a button to tell it whether or not to stop at red lights. That doesn't sound like "self-driving" to me. |
Just noticed on Techcrunch that Google was testing with the cars almost a year ago. Well this is the future!
http://techcrunch.com/2010/10/09/google-car-video/ |
the google car are driving a previous mapped routes. the Braunschweig car not. |
![[put at-character here]](image/at.gif) hebbet:
> the google car are driving a previous mapped routes
That's true, but I don't think it's a negative point. Google has shown that it's quite practical to map the road network in considerable detail.
As I wrote above:
> What Google is doing differently is to build a computerised model
> of the road, instead of making the car work out the whole thing
> from its realtime sensors
In my opinion, that's the difference that's going to make Google's approach become a practical reality in the near future.
Any other approach is too ambitious.
When a human drives, it is also by using a model of what to expect. For example, when we take a flyover at a freeway exit, we have a strong expectation that the flyover will provide a continuous trafficable surface until we once again reach ground level. We don't re-learn how roads work every time when we make a journey.
Of course this sometimes fails, for example if an earthquake breaks the flyover, but I think an automated guidance system can be at least as good as a human in adopting a "failsafe" strategy in this situation.
|
> Now that the technology has reached the point of
> a successful "concept demonstration", it's inevitable
> that we will be using self-driven vehicles for real in
> the near future. Maybe in 15 years, 20 years or 25
> years, but not possibly any longer.
If it will end up being cars instead of other forms of transportation, let's hope it will be smaller cars causing less pollution. Right now typical cars seem to be typically only 50% or less occupied, speaking in terms of available seats (number a complete guess, but you know what I mean). |
When automatic navigation reaches the point where it is much safer than human driving, cars can be made much lighter because they will no longer need to be a collision-resistant enclosure. That will certainly reduce pollution.
Pollution will also be reduced when automatically-driven cars can follow very close behind the car in front, benefiting from a reduction in air resistance due to slipstreaming.
Driving speed could be adjusted with full knowledge of the cycle of each traffic light along the route, resulting in a smoother ride and again reducing pollution.
Automated driving also makes car-sharing much easier. Instead of needing to pick up the car from an agreed depot, the shared car will simply drive itself to your house when you need it.
From an environmental point of view, automatic driving is a big win. |
> When automatic navigation reaches the point where it
> is much safer than human driving, cars can be made
> much lighter because they will no longer need to be a
> collision-resistant enclosure.
I wonder what the theoretical limits to this are. For instance, if the car drives at 100 km/h and a deer suddenly jumps on the road with no way to brake and little pathway to steer clear. Though perhaps the car could have smart heat sensors seeing the deer even before it appears to the eye. (K.I.T.T. was always really good with animals if I remember correctly.)
Perhaps one day the bigger problem will be the psychological one. Right now, there's also no "need" to drive a Hummer, yet perhaps some people feel more powerful driving really big cars.
http://patdollard.com/wp-content/uploads/2009/06/hummer_h6_players_edition_dw.jpg |
