Obviously, it will be purchased from somebody, hence the “little” (rather than “no”) “cost”.
An automated car-control system could include an emergency override allowing you to go somewhat faster than the usual limit (since the risk-benefit ratio would be a bit tilted relative to the usual situation for which the limits are set) and direct other drivers and traffic control signals to accommodate you. Of course, it would record the incident, and you’d better have a valid explanation afterward or else face suitable penalties for raising a false alarm.
I was thinking about this earlier. You could initiate an “emergency drive” (sort of like dialing 911) that would take you to one of the nearest ERs or police stations. I find it interesting that people keep focusing on simple problems and don’t see the huge benefits that could arise. 50 years from now people will look back on our times and marvel at how dangerous it was to drive.
Yes, good point. Essentially any car becomes an ambulance.
Ironically I was still seeing this from the perspective of the human driver world; a lot of things become possible when cars can communicate.
Remember, as has been said several times in several topics on this subject, this stuff won’t happen all at once. It will be gradual as the car manufacturers introduce stuff piece by piece because they are afraid of the liability. The cars aren’t ready now for full blown automation, but they will eventually, so for now, stuff is being introduced piecemeal.
That is good to hear, considering that road markings do get changed. How about road construction work, where we have to obey a worker holding a stop/slow sign?
It won’t be little. The ST replicator transforms energy into matter, for God’s sake. The energy costs would be enormous–far, far, FAR more than any conventional method of food storage and distribution. It could only be practical in a fantasy environment.
Ideally, the slow/stop sign would contain a 50 cent RFID that would let the car know to slow down. Short of that, though, the car can probably be programmed to recognize the appearance of the sign, same as we humans do. And short even of that, it’ll detect the other cars, workers, and other obstacles, and move in such a way as to avoid hitting them.
On Halloween, kids will change the RFID with one from a popup toaster so the car will become unbearably hot inside and bounce its way to the next stop sign!
Damned kids!
It’ll be great when all cars are driverless. Cost to individuals, revenue for governments, and worker displacements will all be worked out to everyone’s benefit. What I’m having trouble envisioning is the transition process from what we have now. Google is sinking a lot of money into R&D, so driverless cars will be very expensive as long as they hold the patents/copyrights. Has Google published any projections for the initial prices?
A lot of the advantages (no drunk drivers, reduced traffic congestion, lower cost from mass production) won’t kick in until all vehicles are driverless. For the first owners of driverless cars, any advantage I can think of could be had much cheaper by using a taxi. What incentive will consumers have to spend big bucks on the new technology?
For good or bad, the same incentives as they do now with seat belts, airbags and whatever else. Laws forcing their use.
Some of the same incentives that they might now have for taking a commuter train: They can spend their commute time snoozing, reading, eating their granola bars, or otherwise taking it easy.
Chronos, a technical-minded Doper, suggested RFID’s in the signs. I, Senegoid, a typical Doper Teeming Mass of no particular technical knowledge, have no particular idea how they might deal with this. Even the researchers at Google haven’t solved all these sorts of problems yet. However, from looking at the videos I’ve seen, I’m fairly satisfied that they really have thought of a whole lot more cases than all the rest of us combined are likely to, and they’re working on all that stuff.
They do have a good start on optical recognition of various signs, traffic signals, other vehicles, bicycles, pedestrians, and lots more. One outstanding problem noted in the videos was driving in rainy or snowy weather, when the camera lenses will get splattered with water or snow or mud. They are studying how badly the visibility could be degraded while still maintaining enough ability to recognize things.
Why do you say that? Cheat codes are a very useful tool for programmers when it comes to debugging. And history tells us that said cheat codes don’t always get disabled when the software is released to the public.
Powers &8^]
The ST replicator transforms energy into matter, but only after transforming matter into energy first, like the transporter does.
Powers &8^]
One important point about automated car technology: The engineers involved will often say things like “We’re working on solving problems A, B, and C”. But the fact that the engineers still have problems to solve does not mean that the technology isn’t good enough yet. In fact, many of the problems the engineers are working on solving are problems that human drivers have, and which have not been solved.
For instance, the problem of poor visibility in fog or dense precipitation: Yes, this can be mitigated with windshield wipers, and I’m sure the engineers have already implemented that. But even with wipers, it’s still a problem. With human-driven cars, we just have to shrug our shoulders and say “well, be extra careful out when the weather is bad”, and there are still a lot of crashes due to the poor visibility. But with self-driving cars, the engineers are working on radar, sonar, infrared, and other senses that aren’t blocked by the weather, and thus actually solving the problem. Those solutions aren’t available for human drivers, but they are for computers. Engineers aren’t trying to make a computer that’s as good a driver as a human is; they’re trying to make the computer as good as possible, which is far better than a human is.
Thanks for the responses, but you both missed my point. For Google to recoup their R&D costs, the first models available for sale will be way more expensive than conventional automobiles; maybe ten times as much. Nobody is going to pay $400k for a little more me time, and the government can’t mandate that kind of individual expense. Am I wrong about the R&D costs, or is there some other way to cover them?
On Halloween those same damn kids will muck with the traffic lights so that they show green in both directions. Imagine the accidents when two busy streets both have green lights!
Why do you expect the R&D for Google would be any more than the R&D by GM or Ford? If GM can do research on different technologies and still keep cars < $400k why can’t Google. I’m sure they have all sorts of ways to monetize their investment.
New technology for GM means a car that can parallel park itself. Google is developing technology to control every conceivable action that might be required in any possible situation that a car might encounter. IMHO that will cost considerably more.
It’s quite common to not recoup R&D expenses until a substantial number of a new invention have been built. Thus Google will be running significant losses on this product for a few years.