My understanding is at least a half dozen companies are going to start releasing self driving cars around 2020 for commercial purchase. Tesla, GM, Volkswagon, Benz, Volvo, Nissan, Audi, Google, etc.
So the cars hit the market in 2020. How long until the majority of new cars sold are self driving from that point? How many years until the majority of cars on the road are self driving?
I think the average lifespan of a car is about 20 years, so it’ll be years after the majority of new cars are self driving before the majority of all cars on the road are self driving. Plus it’ll take years to get the tech into the mainstream.
Is 2050 a realistic time for the majority of cars to be self driving (30 years from the release of self driving cars to the point where most cars on the road are self driving)? Or will it be a niche market like hybrids and electric cars? A handful of models have this option, but most do not.
I believe I’ve heard a stat that says the average car lasts 7 years, so once they are offered to mainstream buyers it won’t be much longer than that before they make up a majority. If self-driving cars end up saving the buyer a lot of money in insurance fees they’ll become mainstream sooner than you think. My completely in-expert opinion says by 2030 they’ll make up a majority.
If they prove to be as safe as they seem likely to be, they’ll take over very, very fast. Google’s self-driving cars have driven over a million miles, and have been involved in just 11 accidents, all of minor, all caused by cars driven by human beings.
However, look for ANY serious accident caused by a self-driving car to be hailed as the absolute Worst Thing Ever and the harbinger of the Robot Car Holocaust.
I’m 59 and I won’t live to see it. Of course I was a computer science major in college in 1982 and didn’t think reliable voice-to-text would happen any time soon so I’ll be happily surprised if such cars become a majority.
Depends strongly upon their price. Hybrids caught on much faster in California thanks to our high gas prices. They’d catch on in more places if they were the same price as non-hybrids.
The only way to accelerate the use of self-driving cars is if the government gave huge incentives to reduce the social costs of collisions and drunk driving and the like.
The stat I heard was they last about 17 years, and that was about 10-20 years ago in an article about climate change. They were discussing how long it would take to transition the vehicle fleet in a developed nation. The average car in the US is about 11 years old nowadays. Of course cars can last a lot longer than 15-20 years, but after 10ish years the cost of repairing them is higher than the cost of buying something better.
Saving money on insurance is a good point. I did not consider that.
What I wonder is that self driving cars are already safer, and should continue to get safer than human drivers. Sooner or later we will have a situation where the chances of an accident by a manually driven car are much higher than a car driven on auto (we may already be there, google says all their accidents were due to humans not the AI). At that point do they start making it illegal to manually drive the same way we make other forms of unsafe and impaired driving illegal?
If the scenario you cite comes to pass, as seems very likely, we SHOULD make it illegal. That does not mean we will. We tolerate thousands of deaths every year from guns because of social factors that make it difficult to ban them.
I think it’s likely that gas mileage will go way up, too.
In selecting a car, right now “fun to drive” is more important that “great gas mileage.” If google’s driving it, and I’m just reading the SDMB on my tablet in the backseat, I don’t give a shit about fun to drive, give me 50 MPG.
I would imagine that if (when?) self-driving cars actually come to pass you’d see the transition be a bit more gradual. I could even see eventually a few lanes reserved for human-driven cars on city highways, at least for awhile. I could also imagine “fun” highways being reserved for recreational driving.
Imagine horse trails, for example.
ETA: I’d also add that if self-driving cars as (a) ubiquitous and (b) safer than human-driven cars, insurance for human-controlled driving would become prohibitively expensive. Think sports cars vs. Toyota Camrys now, but orders of magnitude worse, I’d think (due to so many fewer drivers paying the premiums).
To the OP, I’ve often quipped that I’m not 100% sure my children (now 3 and 8 mos) will ever learn to drive a car. That’s probably optimistic, but I’m reasonably confident that THEIR children never will.
I was thinking about this the other day when one roared past me: what do we do with motorcycles? They probably aren’t nearly as dangerous as other vehicles and I would think it less likely they have self-driving motorcycles.
Keep in mind that all data to date is severely cherry-picked.
We’ve driven 1,000,000 Miles! (on roads and tracks with no other vehicles).
If the story that Google is mapping every stop sign and speed limit so the car needs little real-time interaction is true, those tests are of limited value.
You still can’t map the kid darting into traffic, or a semi jackknifing or emergency road closure.
You can get some great clinical results for your new drug if you get to pick the patients you treat.
You hear a LOT more about “Promising new cancer drug results” than FDA approvals of new “Breakthrough treatments for cancer”.
I see no reason to expect the folks pushing autonomous cars to behave any better than the pill-pushers…
Well, it’s not that cherry-picked. It’s actual street miles. But as I’ve said in other threads, that’s impressive, but not that impressive. I made it through my first 3/4 of a million miles without an accident that was my fault, and that includes the maniacal teen years. No one who had observed my driving during most of those miles would have called it “safe”.
I can see the current level 3 cars becoming more common in 10-15 years, easily, but I doubt they’ll become mandatory in my lifetime. Having aids to keep a driver from crashing is great, but the cost of an automated vehicle just about doubles the price.
I still think that the level 4 car (completely driver-less) is a fantasy that we won’t realize any time soon.
I have heard that while self driving cars are doing well they fail badly in snow and parking lots. I assume that will get solved but till it does they are not viable.
Cities make BIG bank on fines for driving infractions. Self-driving cars will tank city income from speeding tickets and parking fees.
Taxis, trains, buses and airplanes will not be thrilled.
Short version is no matter how cool or awesome self driving cars are don’t expect to see them anytime soon. I suspect they will creep in slowly and slowly overcome resistance but it’ll take decades.
I doubt anyone older than 20 today will see it happen in their lifetime.
It’s more than likely to be something that’s graduated, and/or similar to carpool lanes.
We can’t raise the fuel tax in state because it’s “regressive”; forcing the poor to give up their 20 year old Grand Ams isn’t going to be very popular. Nor will forcing the leftover Boomers and GenXers to give up their classic cars.
Graduation is likely to start with dedicated lanes, and maybe dedicated roadways. Obviously none of us have any data to predict how this will actually play out in the future, but I feel it could be 50 to 100 years before self-driven cars disappear entirely.
Something else that I predict (and terrifies me as I’m in the car industry) is that we will approach 100% resource utilization. Maybe not get there 100% (some people will still insist on ownership), but get close. Why own a car if you only use it for 9% of its availability? It’s much better to use a car-hire service. You’ll never have to worry about an inappropriate vehicle for a task because you can hire what’s needed. Granted fleet owners have to worry about 100% utilization, and so this might not really be achievable without societal changes for working hours, i.e., instead of everyone commuting at 9 and 5, we’d have to better stagger ourselves to support the resource utilization (many of us do this now, e.g., going to work early to avoid traffic).
no they operate on city streets alongside normal drivers…how else would they be hit 11 times.
In science it’s called a control. that mapping is so they can compare sensor input and car behavior to actual road conditions not to spoon feed it.
Actually the cars sensors see far more than you realize and you are fundamentally factually wrong. I have posted videos on several occasions showing how the cars see the environment around them. If a little kid runs out in front of it, assuming sufficient stopping distance. They do stop.
What about stricter licensing requirements (more difficult tests)? Do you think that will ever happen? If so, in what timeframe? I could see it (at least a decade after the cars become available to the masses) because the argument can be made there are more alternatives available (SDCs) than there were in 2015. Also because it’d only apply to new drivers (at least, I don’t know of anywhere in the US that makes you retest later) and people just don’t seem to mind new restrictions rolling out as long as they only affect other people. Especially those under 21.
I think they are likely to become commonplace in and around cities and urban areas in the next 50 years. However, I don’t see them taking over entirely. You will still have people reluctant to adopt the latest technology in every generation. And the poor will likely never be able to afford an autonomous car, even second-hand, as long as regular cars are still around. Perhaps they will be as common as hybrids and electric cars today within the next 30 years, but again, only in urbanized areas.
I cannot see them catching-on in rural areas, or in areas where terrain or roads will never be upgraded to accommodate the needs of these cars. There will always be a market for people-driven cars somewhere.
Also, in checking out the Wiki article on this, I got stuck on this obstacle: Self-driving cars could potentially be loaded with explosives and used as bombs. In our age of security hysteria, which seems to be getting worse, I can see that some cities, or parts of cities, could be “no-autonomous-car-zones”. No matter how good the technology is, it may never be able to overcome some of our fears.
