backtop


Print 15 comment(s) - last by CZroe.. on May 8 at 7:11 PM

It may be deployed in future Tesla vehicles

Tesla CEO Elon Musk is interested in self-driving technology for his future fleets -- and he's going to Google for advice.

Musk has been talking to Google about its work with self-driving cars and how to implement such a system for future Tesla vehicles. However, Musk doesn't want to call it "self-driving," but rather "autopilot" technology. 

“I like the word autopilot more than I like the word self- driving,” Musk said. “Self-driving sounds like it’s going to do something you don’t want it to do. Autopilot is a good thing to have in planes, and we should have it in cars.”

The idea behind such technology is to not only make driving more convenient, but also safer. Cars equipped with self-driving systems can react to certain situations and prevent a crash, for example. 

Google is the place to go for insight on the new technology, considering Google has launched self-driving projects over the last couple of years. Its test fleet consists of Toyota Prius', Audi TTs and Lexus RX450hs equipped with the self-driving technology.


Google's self-driving cars use LIDAR, which is a rotating sensor on the roof that scans more than 200 feet in all directions for a map of the car's surroundings; a position estimator sensor that helps locate its location on a map; four radar sensors to identify the position of distant objects, and a video camera to detect traffic lights as well as moving objects like pedestrians.

“The problem with Google’s current approach is that the sensor system is too expensive,” Musk said. “It’s better to have an optical system, basically cameras with software that is able to figure out what’s going on just by looking at things.” 

As of right now, Google's self-driving cars are allowed to drive on Nevada, California and Florida roads for testing. Last month, it was reported that Michigan may approve autonomous vehicle licensing soon. 

Musk sees autonomous driving -- or autopilot -- as the next step in the evolution of vehicles, but isn't working on development quite yet. 

“We’re not focused on autopilot right now; we will be in the future,” Musk said. “Autopilot is not as important as accelerating the transition to electric cars, or to sustainable transport.”

Source: Bloomberg



Comments     Threshold


This article is over a month old, voting and posting comments is disabled

RE: Easy for you to say.
By dnd728 on 5/8/2013 9:42:14 AM , Rating: 2
Mobileye has been doing this processing for years, on a single chip, with just one camera or several.
iOnRoad has a smartphone app that does roughly that.

I think it can be done, and with less than a supercomputer.


RE: Easy for you to say.
By lelias2k on 5/8/2013 9:55:38 AM , Rating: 2
Great points. I wasn't aware of iOnRoad and found it really cool.

Furthermore, the computational power that we have available nowadays is much higher than what we see in our daily use of a computer.

Think about high-end Xeon processors and graphic cards. We're talking about Teraflops available in one chip. If we can have systems that work with a cell phone chip, imagine with one of those!

And let's not forget that we're still years away from deployment of such technology, which means we'll have chips that are even more powerful.

I think it will be fine. :)


RE: Easy for you to say.
By invidious on 5/8/2013 11:15:24 AM , Rating: 2
Wrong. iOnRoad does not even come close to the level of image processing that would be required for a driving computer. The fact that it only uses one camera should make that abundantly obivious.

iOnRoad only monitors what is happening directly in front of the car. Mapping a 3D environment based on 360 degree snapshots is orders of magnitude more complex than measuring linear distance on one axis.

iOnRoad also does not make decisions, it only does analysis. Taking action requires a vast increase in computing complexity. Accelerating breaking and stearing are not binary choices. Each has a magnitude associate with it and many driving scenerios require manipulating two at once.


RE: Easy for you to say.
By dnd728 on 5/8/2013 1:04:16 PM , Rating: 2
:)
iOnRoad was only an example of how a 2.5 guys startup could pull the image processing on a generic CPU and a generic camera. Still…
Decision making seems rather irrelevant, since it would have to be done for LIDAR and Radars just as well.
Cameras are not one dimensional – they can cover way over 90 degrees (just horizontally), hence covering 360 degrees should not be "orders of magnitude more complex".
Mobileye, that I also mentioned, as I recall, has been able to process full 360 degrees even on their old systems. Their systems are integrated in many luxury cars.
And cameras would probably be installed anyway - for recognizing objects, reading road signs…


"The whole principle [of censorship] is wrong. It's like demanding that grown men live on skim milk because the baby can't have steak." -- Robert Heinlein

Did You Partake in "Black Friday/Thursday"?
Did You Partake in "Black Friday/Thursday"? 














botimage
Copyright 2014 DailyTech LLC. - RSS Feed | Advertise | About Us | Ethics | FAQ | Terms, Conditions & Privacy Information | Kristopher Kubicki