I'm fairly certain that in the not too distant future, we will think back to these times and wonder why people were ever trusted to drive on their own.
Aside from the GPS and street map data inputs, such as used by the Google system, I think what's required before autonomous cars can become reality is also vehicle-to-vehicle and road-to-vehicle comms. These are the ingredients you need to accommodate all of those real-time unknowns.
I foresee another very useful function for the systems installed on autonomous vehicles. They would make great recorders for accidents and nearby, aggressive, dangerous fools on the road. I think the inputs from the video, ranging systems, etc. should be constantly recorded and analyzed for incidents that warrant a law enforcement official or agency to be notified. The data could be transmitted automatically and instantly. What a great way to catch the remaining "problems" on the road. If a system like this was in place (which I greatly hope for), the blogger who thinks the autonomous cars would move out of his way would suffer quick repercussions in the form of a fine and points on his driving record. He wouldn't be doing that for long. However, I'm already suspecting that the "invasion of privacy" whiners will do everything in their power to stop a system such as this from being put into place (the same ones who don't want cameras at intersections). We need to find a way to actually start enforcing the laws on the roads. We need to all come up with practical solutions like this for our roadways to be safe and usable in the future - with or without autonomous vehicles.
There's a great Wiki being run by Bryant Walker-Smith.
It has the full updates on the legal situation across the US.
Driverless Car HQ
I'm going to jump in and disagree on this one.
Many Driverless Car safety and privacy concerns revolve around V2V. There's significant resistance within the tech community to anything that can be used to invade privacy, and they are the people we need the most as the early adopters to make this technology viable. This is notwithstanding hacking worries.
While we envisage that V2V is an inevitability we don't see it as a necessity. Cars right now drive without difficulties and they don't have V2V. Driverless Cars have far better driving ability than your average joe driver so they should be ok, at least in my opinion.
Driverless Car HQ
Hi Matthew. It seems to me that the Google system has to be augmented by real-time information of road conditions where this autonomous car is driving. So for example, is there a huge pothole in th way? Is there a stalled vehicle in the way? Is there some other obstruction that Google maps and GPS can't possibly know about?
The concerns about privacy can be dealt with. This doesn't have to divulge personal info, it only has to divulge the driving environment. I'm not sure there is any way to achieve practical autonomous vehicles with what Google has done so far as the only ingredients.
This is all completely ridiculous; we've got here a bunch of very clever people and expensive technology solving a problem that doesn't exist. Get human drivers off the road. Then design a road system for machines. The problem then becomes a lot simpler. It is physically impossible to mix human drivers, working within a set of rules designed for humans by humans, with computer-controlled vehicles.
Autonomous vehicles are the key to real mass transist. The current system, involving steel rail tracks able to hold multi-ton railroad cars, is obsolete and cost prohibitive. Stations are too far away from most residences for people to walk up to, and trains are too expensive to make available for on-demand use - in otherwords, the passengers must wait for the train rather than vehicle waiting for the passenger.
A system built to handle vehicles weighing less than 2000 lbs gross would not need the massive infrastructure costs. Individual vehicles would be sufficiently inexpensive that a surplus can be available for on-demand use except during peak traffic periods, stations can be very close to individual residences, and the vehicles, which would only carry a handful of passengers at a time, would be able to selectively skip those stations where current passengers are not intending to get off.
NASA's Orion Flight Software Production Systems Manager Darrel G. Raines joins Planet Analog Editor Steve Taranovich and Embedded.com Editor Max Maxfield to talk about embedded flight software used in Orion Spacecraft, part of NASA's Mars mission. Live radio show and live chat. Get your questions ready.
Brought to you by