SoylentNews is people
SoylentNews
Automaker Daimler AG and parts supplier Bosch Group are teaming up to make driverless cars that they say could be on city streets at the start of the next decade.
The companies would combine expertise in car making, sensors and software so that people in a specific part of town could order a shared car through their smart phone. The driverless car would then come pick them up and take them where they want to go.
The companies said Tuesday the system would let people make better use of their time in cars and help those who do not have driver's licenses.
The auto industry and tech firms have been investing heavily in autonomous driving technology. Many basic elements of autonomous driving are already in use, such as driver assistance systems that can keep cars in freeway lanes or detect pedestrians ahead. But legal issues surrounding driver responsibility remain to be solved.
Other automakers are also working on the concept. Competitor Volkswagen showed off an autonomous vehicle with no steering wheel and facing sofa-style seats at the Geneva auto show last month. Alphabet's Waymo, the former Google autonomous car project, is testing autonomous vehicles on public roads in the United States with test drivers aboard.
In Soviet Russia, a car you don't drive is called a "taxi."
(Score: 3, Informative) by LoRdTAW on Tuesday April 04 2017, @11:15PM
At first I thought that was a little high. Then after having a look at Nvidia's Drive Automation [nvidia.com] page I can see that they have up to two Pascal GPU's per Drive PX 2 in Autochauffeur configuration[1]. I could see power draws surpassing 1kW if they are using multiple GPU's for the vision and LIDAR systems. The Nvidia GTX 1080, based on Pascal, draws upward of 310W under heavy load[2]. I can't see sensors or cameras drawing more than a few watts. Even a LIDAR I looked up draws only 8 watts[3]. So there are either ten or more GPU's in that car or it was a test system with way more compute power than the car needed.
And here is something else I learned on Nvidia's site that isn't sitting well with me: http://www.nvidia.com/object/hd-mapping-system.html [nvidia.com]
Essentially their Drive Automation platform is an always connected cloud platform which collects real-time 3D mapping. That means real time 3D surveillance in the vicinity of an autonomous vehicle. Who wants to bet the NSA and other three letter agencies already have hooks into such a project? Millions of 3D scanners running around 24/7 watching every move you make in 3D space. Police could instantly call up 3D scans from cars in the vicinity of a crime scene to see what was in the area. They could also track and measure physical changes such as objects being placed, moved, or removed. And there is the possibility that the vision system could be tied into this enabling real-time video as well. A tyrants wet dream. And lets not get started on the hacking and terrorism angle. Spoofing 3D data to cause cars to suddenly think there is an object in the road causing it to crash or come to a sudden stop is one thought.
Though, there is good that can come from this. Would eliminate the need for dedicated "street-view" cars for online maps. Instant video/3D from accident scenes. Better 3D mapping of the roads for helping construction crews find and prioritize road repairs, road kill removal, and even detect downed trees or other dangerous foreign obstacles. It could be used for analyzing traffic patterns to help design better traffic flows and roads. And a whole lot more.
1. https://en.wikipedia.org/wiki/Drive_PX-series#Comparison [wikipedia.org]
2. http://www.tomshardware.com/reviews/nvidia-geforce-gtx-1080-pascal,4572-10.html [tomshardware.com]
3. http://velodynelidar.com/vlp-16.html [velodynelidar.com]