Jalopnik asks if self-driving cars will actually happen.
 

Please excuse me for starting yet another self-driving car thread, but I did not want to hijack someone else's discussion.

Jalopnik says that self-driving cars have only been a few years away for decades, much like flying cars and virtual reality--and presumably, self-flying cars with virtual reality.

I have a friend that insists that self-driving cars would be hacked.

I call her a know-it-all.

She says I'm a know-it-all.

She also says that the same bad drivers will be programming the cars.

I insist they will not be texting and coding. Plus, I am pretty sure that programming gets tested, while texts are not proofread. Messages still send with invalid syntax.

Jalopnik also brought up terrorist attacks. I recently asked if self-driving cars would be easier or more difficult to steal, but in theory, a terrorist would need to hack a vehicle, but if they could, that could be a pandemic.

http://ecomodder.com/forum/attachmen...1&d=1511885337
https://www.cdc.gov/mmwr/volumes/65/wr/mm6543a8.htm

Vehicular injuries are significantly down, vehicular homicide is down slightly, and suicide has roughly been constant. As long as programmers can prevent vehicles from being hacked, all of these go down, if not away, right?

In the comments, people discussed requiring continual driver training and testing, but someone asked how much that would cost, who would pay for it, and how it compared to self-driving vehicles.

Would the world be better or worse with automatic blinkers? :)

Jalopnik pointed out that traffic fatalities would halve if people just wore seatbelts.

"I don’t want an autonomous car. I want a robot butler who can drive a regular car."

"[T]here are cameras on the highways [in England] that monitor if you’re wearing a seatbelt. I’d imagine similar could be used to flag people on cell phones."

I understand that photo radar takes people's money without changing their behavior. In my experience, people slow down for them, but speed back up immediately.

"In 2013, at least 697 people were killed in an accident that involved running a red light while an estimated 127,000 were injured. Last year, some 3.7 million drivers received a violation for driving through a red light." https://www.forbes.com/sites/niallmc.../#5b3c8f73db27

3,700,000 people were cited for running red lights, but 700 or more people died anyway.

Perhaps the fines are not enough.

What do you fine and informed people think of these arguments?

"Super scared" seems like a drastic overstatement, thus suggesting this article is clickbait. https://jalopnik.com/americans-are-s...tin-1819136046

Uncomfortable ≠ super scared.

"It was a short article and nothing else stood out to me. In the comments: Americans are just super scared of everything right now. I have to travel overseas for my job and I do get asked why so many Americans seem so scared of everything, quite regularly. The only reply I can give them is `I don’t know, I’m Canadian', because I’m too scared to admit to being American right now."
"Your disguise is up, you didn’t apologize."

"To be fair, the auto industry hasn’t exactly nailed implementing security into existing CAN/infotainment systems, so they aren’t inspiring confidence that they’ll get it right for autonomous driving systems either."

Self driving cars as they exist today do things a human driver would not
These things can be dangerous in the presence of real human drivers.

These errors are a form of violating commonly known ediquette by driving in a way human drivers view as irrational (aka common sense)

I believe what we will find is that self driving cars won’t mix well with humans or even “open country roads”

Does this mean they won’t be put on the road anyway ?

Nope, but we probably aren’t going to put out cars quite ready for prime time, we will error proof by making everyone a test driver

As for yellow lights around here they flash like a strobe light,

Just increasing their duration would help,
Also stop and go lights on 45mph+ highways suck, putting a stop and go light on a 65mph highway is poor road design and causes accidents ,
If that practice were illegal and such intersections were properly designed we might have fewer accidents.

Mitch Bainwol, the president and CEO of the Alliance of Automobile Manufacturers, a trade group that represents automakers that produce 80 percent of cars on U.S. roads today, testified to the Senate that retail sales of fully-automated vehicles aren’t expected to begin until 2025 at the earliest.

"[M]ore than 20 percent of cars on the road today were produced before 2002 – [fully autonomous] vehicles [...] will likely not be a majority of the fleet for three more decades."

https://jalopnik.com/automakers-admi...tak-1796103407

Someone commented that autonomous vehicles should mean people stop owning cars, but somebody pointed out that we do not need to have our own homes, but we choose to anyway.

People don't like change, and it comes slowly. They are more comfortable living in the dangerous world they know, then venturing out into the safer world they don't. If you think about it rationally, it should bother you that so many people are not paying attention to their driving and realize that autonomous systems never get distracted.

Autonomous cars have not just been around the corner for decades. It's recent cost reductions and improvements in sensors and computers that have allowed it to become financially practical. A LIDAR might have cost several thousand dollars in the past and been bulky, but is now less than $100. My middle school bought a 640x480 digital camera for $3,000 back in the day; now a $30 smart phone has 2 cameras superior to it.

Driver assist technologies are creeping in and will become standard equipment just as Bluetooth connectivity is. This will get the average person more comfortable with slowly giving up control.

Finally, the whole world is operating on computers, so the "what about hacking" argument can be applied to anything. It's a possibility; and it hasn't prevented us from developing technology in the past. Almost all of my cash and investments are held as binary bits of data, and that is of much greater concern to me than someone hacking my car.

I've been wrong about the adoption of home networks, and internet on your cell phone. I'm not betting against self-driving cars, which are likely to be among the biggest advances in safety and productivity in my lifetime.

1) Unfortunately TOO many people want "change" simply for the sake of change, without actually IMPROVING anything except their bank account.
2) First editions of most human-produced items tend to be filled with problems.
3) "Don't TELL me, SHOW me" has eons of foundation.
4) Brownian Movement is Mother Nature's way of letting things "move" without actually displacing...works for humans too.

What is the definition of "prime time"? If it's safer than human alone control and reduces collisions and injury, then it meets my definition of prime time. We're all test dummies in life, as no product is fully dummy proof.

I believe 2025 is about right, if not a few years optimistic. Once it starts though, the majority of the fleet will turn over much faster than 3 decades. He's even quoted as saying only 20% of vehicles on the road are 15 years or older. That's 1.5 decades, and autonomous driving will make regular cars obsolete faster than normal.

Homes and vehicles are very different. I am perfectly comfortable renting a car or allowing a taxi to take me somewhere, but most people aren't comfortable sharing their personal living space with strangers. Sure, some people will still own vehicles for similar reasons they own their home, but many will forgo the expense.

I don't disagree with any of those points. People wanting change for the sake of financial gain is not necessarily bad. It's what drives innovation, and is what allows a variety of products to exist since what is "better" is subjective. Is the iPhone a better phone than a Galaxy? Both are encouraging people to change for the sake of increasing profits.

No doubt, the 100th iteration will still be filled with problems. Incremental improvement is what keeps things interesting.

Commercial airplanes are heavily automated with the goal of increasing safety and efficiency while reducing personnel (where did the engineer go?). The same logic applies to vehicles. This is the show part of the talk.

Automation will be worse until it is better. Has it ever been any different?

I think there are several huge hurdles to overcome, before it has a chance of becoming a reality:

Technical complexity

Legal liability

Regulations

I think it will take a massive leap in computing power (quantum computing?) in order to overcome the massive job if first gathering ALL the data, and then figuring out what it is - and then it has to logically "decide" what to do. And then have a redundant back up for all the technical systems.

Who / what is responsible for what happens? How do we decide who / what is responsible?

We have not even gotten side video mirrors approved for use - how can we possibly approve fully autonomous cars?

"It's about the future, Madame Chancellor. Some people think the future means the end of history. Well, we haven't run out of history quite yet. Your father called the future - "the undiscovered country". People can be very frightened of change."--The Priceline Spokesman

Speaking of 640x480 cameras, one help drive an autonomous junk minivan across the country in 1995: https://jalopnik.com/they-drove-cros...tho-1696330141

"For gas and spending money, we sold trip T-shirts. I’m not kidding. They were $10 apiece and helped pay for food and hotels. Seriously."

"When we did the trip, the field was about discovery and expanding technical frontiers. I think it still is now, but unfortunately, it’s now also about patent fights, liability concerns, and state laws."

How about laws telling humans how to behave around automatomobiles?

Why would it mean people stop owning cars? Urbanites, probably: the same sort of people who find it convenient to call a taxi to get around in the city. But if you live outside the cities, calling a taxi/autonomous car might involve an hour or more response time, and double the amount of driving (and hence fuel consumption & traffic) for the autonomous car. And really, if I drive to a remote trailhead, I really don't want to depend on summoning someone's autonomous car to pick me up.