Slashdot Mirror
Grand Theft Auto V Is Being Used To Help Teach Self-Driving Cars (bloomberg.com)
An anonymous reader quotes Bloomberg:
In the race to the autonomous revolution, developers have realized there aren't enough hours in a day to clock the real-world miles needed to teach cars how to drive themselves. Which is why Grand Theft Auto V is in the mix... Last year, scientists from Darmstadt University of Technology in Germany and Intel Labs developed a way to pull visual information from Grand Theft Auto V. Now some researchers are deriving algorithms from GTAV software that's been tweaked for use in the burgeoning self-driving sector. The latest in the franchise from publisher Rockstar Games Inc. is just about as good as reality, with 262 types of vehicles, more than 1,000 different unpredictable pedestrians and animals, 14 weather conditions and countless bridges, traffic signals, tunnels and intersections...
The idea isn't that the highways and byways of the fictional city of Los Santos would ever be a substitute for bona fide asphalt. But the game "is the richest virtual environment that we could extract data from," said Alain Kornhauser, a Princeton University professor of operations research and financial engineering who advises the Princeton Autonomous Vehicle Engineering team.
The idea isn't that the highways and byways of the fictional city of Los Santos would ever be a substitute for bona fide asphalt. But the game "is the richest virtual environment that we could extract data from," said Alain Kornhauser, a Princeton University professor of operations research and financial engineering who advises the Princeton Autonomous Vehicle Engineering team.
Hmmm. Maybe not that one. Maybe don't teach your AI with simulations that include the option of going on a shooting spree?
Your eyes are far better at matching light frequencies between both eyes to get the depth mapping correct. Your standard camera can only distinguish 24 bits of light frequency. At that level you get somewhat of a depth map but not a very good one.
Lasers try to get around that limitation by using a frequency the camera can easily pick up and compare between the two images. If you could use the whole image and any frequency, you'd be a lot better off.
That's ultimately the challenge: getting cameras that are not only incredibly sensitive to light frequency, but also very high resolution. Or they'll need to get the cameras looking around just like your eyeballs.
In a 3D mapped world, all the depth information is 100% accurate.
They'd need to render 48-64 bit color to emulate what might be possible in the real world to get accurate depth information.
Work Safe Porn
So will Waymo vehicles elect to stop for prostitutes or just run them over?
Laws are rules for the court, but merely a bottom bar to hit for life. Think beyond laws in your actions always.
This is how Sky Net, et al become cynical
It's all fun and games until your physical car decides to stop for hot coffee.
"First they came for the slanderers and i said nothing."
remember cops are bad in that game
... i would have said that it was far more likely that it was *Uber* that was taking "Grand Theft Auto" to heart...
So they're teaching their cars how to find prostitutes and avoid the cops?
this video.
It turns out that Rockstar is not so happy about them using their software with out asking first:
http://mashable.com/2017/04/21...
“We welcome discussions about the use of our technology to help further academic research, but it’s obviously not appropriate for corporations to take our work and use it for their own financial interests or for researchers to distribute unlicensed copies of our code as part of their work without first seeking our permission,” the company said.
Nowhere in the article is it claimed that Waymo uses GTA V in any way. Waymo uses simulators, and some academics used GTA V.
I worked with Alain Kornhauser about thirty years ago, first taking his robotics course as an undergraduate, later managing his robotics lab as an employee, and then again even later (briefly) as a grad student tangentially as part of a group doing self-driving car research focused mainly on a neural networks approach. I had also been hanging around Red Whittaker's group making the first ALVAN (Autonomous Land Vehicle) around 1986 before going back to Princeton to work as an employee.
While I did not contribute much of significance to that self-driving car group (I had other interests), I had suggested we train cars to just drive one specific route based on videos from driving that route a variety of times. I guessed that most daily commutes are just along the same route and so that could be a big win. But he dismissed that idea for some reason I'm still not sure I understand. Still think it made a lot of sense though for the resources we had at the time.
About ten years ago I suggested he get his PAVE students to write software to drive Gran Turismo as a challenge. Not much response from him on that then though. Glad to see his is finally doing that -- although with much better game/simulation software now.
I also suggested he could make PAVE the free and open source software hub for self-driving vehicle software to address some concerns I outlined back in 2001 in the essay to the Markle Foundation:
http://pdfernhout.net/on-fundi...
From the email I sent Alain in 2007-02-02:
"Glad to read of your group's successes with the Grand Challenge. I've long thought a fun project for your students would be to write software that takes visual input from a a PlayStation 2 driving game like "Gran Turismo" :-) , but it is cheap, easy, and safe to do in an undergraduate lab with limited supervision. And the racing game simulators just keep getting more and more realistic. And if that challenge becomes too easy, you can then add noise to the video signal to make it harder... Or introduce lags or noise in the USB steering. And then start working on controlling ATV Off Road Fury or the the Snowmobile racing games, and so on. Or have kids write software to control one game and then give them only one day to make it work for another... Probably lots of good science and engineering and education to do there on a (relatively) small budget."
http://en.wikipedia.org/wiki/G...
http://www.gran-turismo.com/
(direct via video out to video capture, or even through a camera focused on a TV) and processes that image to drive the simulation via a USB hookup into the PlayStation. Not quite the real thing (and Red Whittaker might rightfully scoff at that approach as ignoring much of the challenge of making real hardware survive in a tough environment
I mentioned that idea again to him in 2011-06-18 when I was looking for jobs:
"Or maybe you need someone to do more work on cars that drive themselves, which sounds like more fun? :-) Except that PAVE stuff is all student run, and good for that approach, so I can see you probably won't need someone for that. I still feel getting students interested in writing open source software to process images from the latest driving simulator games is a good (safe) project that might advance the state-of-the-art in automotive intelligence in a very positive way. :-) I'm sure it would at lead to lots of funny press though ("Students at Princeton are seriously playing with video games", and so on). Whether that is good or bad depends on your point of view, perhaps."
Anyway, glad to see that idea finally getting some traction. :-)
While he did not take some of my ideas that seriously, I did not take his
A 21st century issue: the irony of technologies of abundance in the hands of those still thinking in terms of scarcity.
https://cias.rit.edu/media/upl...
Luminescence can be used to create a depth map using two stereoscopic images.
3D movies and VR headsets don't work for everyone and create headaches for a variety of reasons, one probably being the lack of bit depth.
Work Safe Porn
https://cias.rit.edu/media/upl...
Luminescence can be used to create a depth map using two offset images.
3D movies and VR headsets don't work for everyone and create headaches for a variety of reasons, one probably being the lack of bit depth. It's much harder to judge depth when the colors are so close together.
Work Safe Porn
Top 10 Most Expensive Cars In The World 2017-
https://www.youtube.com/watch?v=DrtuZX6GnVc
Doesn't anyone ever learn ANYTHING from horror novels and movies?
"Self-driving car" + "Behavior from Grand Theft Auto V" sounds a LOT like a recipe to make more instances of Christine. All will be fun and games until several high school students go missing, and the "paint" on their cars seems to be a slightly different shade of red than usual.
But of course adults won't listen to warnings in time to prevent any of that. In the movies, they never do.
and how many people want their car to drive like GTA?
The original software based education tool: Carmageddon
NPC cars in GTA V usually drive quite "neatly" and predictably. Not sure it will be an adequate representation of real life where you have morons doing stupid things every once in a while.
Unless they are using the online version of the game, in which chase there will probably be too much chaos on the streets.