October 06, 2015
Toyota: Autonomous Highway Driving By 2020
Check out what Toyota is doing with autonomous vehicle development.
By contrast, Tesla is just weeks away from autonomous highway driving. Elon Musk expects Tesla to achieve full autonomy in 3 years but with regulatory delays the deployments will vary between jurisdictions. Looks like the 2017 Cadillac CT6 will be the second car to hit the market with autonomous highway driving capability.
Recently Daimler tested an autonomous truck in real highway traffic.
Future cars are going to be autonomous and electric. General Motors says their cost of batteries has fallen to $145/kwh. To put that in perspective, that's a drop from $1000 in 2008 and $485 in in 2013.
These autonomous electric vehicles will be powered by solar energy.
Now if only biomedical technology could advance as fast as computers, batteries, and solar panels. Then I'd be thrilled about the future.
Randall Parker, 2015 October 06 08:22 PM
You can't automate what we don't understand [physiology]
Automated highway driving only helps when almost ALL cars have it. So still 20 years.
Plus, what about those hybrid 45 mph expressways with traffic lights? What about those?
@TTT No, not all vehicles have to be autonomous for benefits. Safety will be much enhanced as the fleet gradually shifts, not to mention the individual benefits of watching the game on your phone while "driving".
I think at least 50% fleet shift is needed to see a real drop in accidents and traffic gridlock.
If automated cars have cheaper insurance and other immediate benefits, the first 50% may be faster than the expected 8 years or so.
But holdouts who are financially strapped will be a problem. There needs to be a tax credit for automated cars (one of the few tax credits I would approve of).
Keep in mind that the average car on the road is 11 years old in the US, despite the major technological improvements that have already been seen in 2015 models vs. 2004 and earlier models.
If there is a kit that can retrofit existing cars, that would be something..
The age of the cars in the fleet isn't so much of a factor as the age of the cars on the road. The average US LDV travels half its lifetime mileage in its first 6 years. Even if you have lots of 15-yr-old cars still registered, they are going to be a much smaller fraction of the vehicles encountered.
You are wrong for multiple reasons:
1) Single car accidents. By definition they only involve a single car. All autonomous cars will be drastically less likely to get involved in single car accidents. I did some quick googling and found In Pennsylvania in 2012 52% of fatal accidents are single car accidents.
2) An autonomous car can react far more quickly and optimally to a bad human driver's actions than can the average human driver. So autonomous vehicles will reduce the consequences of bad choices by both their owners and other drivers.
3) As E-P quite correctly points out, most miles driven in a car are driven in its first 6 years on the road.
The wild card: What kinds of people will be the first to buy autonomous cars? I'm thinking the early autonomous cars will be deluxe models: Cadillac, Mercedes, Lexus, BMW, Audi and at the higher end of their offerings. So more upper class buyers. That trends somewhat older (and new car buyers are generally older than the average adult person). On the one hand that reduces danger from old people who are becoming more dangerous. On the other hand, its not addressing the dangerous teen driver problem. However, some affluent parents will certainly rush to get autonomous vehicles for their kids because, hey, the kids are dangerous.
I expect the benefits of reduced deaths will more heavily flow to the more affluent for this reason. However, taxis will go autonomous to save labor costs. This will make taxis cheaper and move all social classes somewhat out of their own cars into using taxi rides instead. Per car the benefits of reduced accidents will be greatest for taxis because they are driven many more miles per year than the average car.
I don't agree.
Stanford demonstrated self-parking and self-driving cars in 2007, and the progress since then has been glacial, with only a little burst recently.
We are still 20 years away from the full effects of mass adoption, such as low accidents, longer commutes (hence a deflation of expensive real estate in some areas), and saving of money from road widening.
It is true that half of all accidents are single-car, but those are usually not the same people who will adopt expensive self-driving cars at first. Those single-car fatalities are often people in their 20s who go out clubbing, etc.
You just moved the goal posts. "full effects" does not equal 50%.
Stanford 2007 vs today: A huge step forward. Self driving in 2007 was crude.
Pretty scary to me, what happens when someone figures out how to hack one of these?
Pyora, I guess it's much the same as what happens when someone hacks the traffic lights. People's lives are endangered for no practical purpose. Fortunately self-driving cars won't be as easy to hack as human beings, who can be distracted from what they are supposed to be doing just by calling their smart phone.