Self Driving Cars?

[GrayHare]

I can't count the number of times people flash their lights at me, fly around me, flipping me off, practically taking off the front of my car as they come back over. Then I get stuck behind them at the next light. Stupid. Was it really worth all that? What, will you get to work 2 minutes earlier?

I've seen that too, most recently the nut sailed around me so fast he could not slow enough to avoid rear ending the car ahead of me. I'm glad the driver of the car ahead of me was not hurt. That's one manuever if done by an autonomous vehicle would lead to a lawsuit against the manufacturer.

But it did make me wonder, when speed-limit-observing autonomous vehicles are common, will the aggressive drivers become even more aggessive believing the autonomous vehicles will be able to get out of their way?

 

[Midpack]

Self driving cars still don't make a lot of sense to me if we're all going to have one. You'll still have tons of parking lots, traffic congestion, etc. Yeah, they may theoretically be safer some day, but there are plenty of safer alternatives to driving now.

Not sure what parking lots have to do with autonomous cars, that has to do with car/ride sharing, not self driving cars themselves so much.

But there won't be as much congestion, or the engineers will have failed completely.

• If there are 90% fewer accidents, there will be less congestion.
• A self driving car won't change lanes arbitrarily, those maneuvers will be more systematic. Cars will probably "talk to each other" sooner or later - already R&D underway.
• Self driving cars won't drive all different speeds, I suspect there will be default speeds for different traffic zones, and a decision tree to vary according to conditions - or something like that.
• Self driving cars won't be talking, reading or texting, under the influence, putting on makeup, half asleep, reaching in the backseat, dealing with kids, won't care about rain, snow or dust, etc.

There will be less congestion to be sure.

 

[Midpack]

Alternatively, a system like the current self-driving car test driver, where you sit in the driver seat ready to take control at a moment's notice is guaranteed to fail. Very few non-drivers will remain attentive enough to be aware of a problem situation, much less be prepared to take the correct action if such a situation arises.

As I noted in the paragraph immediately after the excerpt you quoted.

 

[Midpack]

And as cars take over more and more of the driving, the human drivers (especially the younger ones who will have known nothing but self-driving cars) will become less and less skilled. And situations that require manual intervention are generally the trickiest. So the skills will be lacking just when they're needed the most.

Fortunately experienced licensed drivers will be very common at first, and for at least a generation. But it does bring up an interesting question re: newly minted teenaged drivers. Licensing will have to continue well until cars are truly totally autonomous if not well into that period?

Another thought is traffic that contains a mix or manually-driven and self-driven cars is much more difficult to manage than traffic that contains only self-driving cars. The self-driving cars will be more predictable and in cases where the "other (self-driving) guy" does something wrong due to a software glitch or whatnot, it can be alerted by the self-driving car that is about to become a victim.

True enough. Early development was about finding lanes, making turns and navigating from point A to B.

Now development is mostly about coordinating with other cars, virtually all of them manually driven. They're deliberately throwing as many potential accident situations at the development cars as they can think up - pop up pedestrians, cars blowing red lights thru intersections, construction and unmapped obstacles, and on and on. Some programs have driven millions of miles now, and learned some hard lessons, among manually operated cars. So that's being developed. They all know they have to clear a very high bar, and there will be regulations as needed. By the time you and I can buy an actual self driving car, dealing with manually operated cars will be largely sorted out. And we'll have steering, pedals, etc.

 

[swakyaby]

My car currently has some auto-driving features, not just safety features, that are fun to have, but I admit I use less than half the time. There is auto-steering which I almost never use--I like to steer the car myself. However, on long trips on congested highways, I like to use the adaptive cruise control which speed and slows depending on traffic. It's much more relaxing than having to control the accelerator and brake all the time.

Perhaps the most fun feature is the auto-park, where it will parallel park for you on a city street (perfectly!) or will back into a perpendicular parking space at the mall. Watching the steering wheel spin and the accelerator pedal depress as it positions itself perfectly between 2 parked cars is very impressive.

I never use the self-driving features to become distracted with texting, video-watching, eating, sleeping (!), or other stupid distractions.

 

[Midpack]

I can't wait and have been thinking of reserving a spot for a Tesla model 3. I'm figuring that self-driving cars will arrive just in time for DW and I.

So are we...though not so much for any self driving feature. I suspect like any new tech, the first truly self driving cars will be premium priced. There are already cars (mostly premium) with many of the early systems that'll be required for self driving vehicles, that helps us get gradually get used to the idea and gives manufacturers confirmation testing (after thorough development, DON'T think 'guinea pigs').

 

[6miths]

Yes I think that is a good question of what will become of speed limits as we transition to autonomous vehicles. Prevailing speed on most controlled access highways in our area is 20% above posted and on the local toll way is 30-40% above posted.

 

[gerntz]

Agreed, but accidents caused by machines are perceived differently than accidents caused by people. Computers and robots are almost always more reliable than humans, but the mistakes that machines make tend to be different from the mistakes that people make. And there's the expectation that people will make mistakes, but that machines should not. When a self-driving car causes an accident that a human could have prevented, it will be perceived as a much bigger negative than the reverse.

Get over yourself - which is the problems you describe say you're not.

 

[gerntz]

As I get older it's more and more obvious that there will come a time when I can no longer drive myself. My parents have been dependent on others to drive them places for years and they deeply miss the independence.

They could set up a retainer situation with a cab company or Uber/Lyft to cover this.

 

[gerntz]

I drove in rush hour traffic for many years, and even today the vast majority of accidents are caused by human error.

Which ones aren't?

 

[gerntz]

Alternatively, a system like the current self-driving car test driver, where you sit in the driver seat ready to take control at a moment's notice is guaranteed to fail,

On "a moment's notice", yes. But for many of those, humans fail today, so no difference. On less extreme situations, the vehicle could start slowing to a stop till the mere human takes over.

 

[gerntz]

Self-driving cars will be like computers, cell phones, or just about any new technology. The older generation who is used to the old way of doing things will find the adjustment somewhat difficult, but the kids will have never known anything different, so it will be natural for them.

Excellent.

 

[gerntz]

And as cars take over more and more of the driving, the human drivers (especially the younger ones who will have known nothing but self-driving cars) will become less and less skilled. And situations that require manual intervention are generally the trickiest. So the skills will be lacking just when they're needed the most.

Humans can't outperform machines in that situation now. So you're still better off.

 

[gerntz]

I can't wait and have been thinking of reserving a spot for a Tesla model 3. I'm figuring that self-driving cars will arrive just in time for DW and I.

Don't see what Tesla has to do with this, but OK.

 

[Animorph]

I've got a Tesla Model X on order, with all the automated driving features, most of which are yet to come. That includes "full self driving". This is a car we'll be driving on cross-country trips. the more driving it can do the easier that will be.

I've always driven Porsche sports cars until now. Yes I like to drive, I like to manual shift, I like to modify my suspensions, and I've driven a couple of my cars in race track events with the Porsche club. But long distance freeway driving? Not so fun.

I'm already a worse driver than I used to be, though I'm only 62. My concentration is no longer solid. My brain is pretty close to 100% occupied with driving while navigation and conversation tends to be neglected at times. I have a harder time turning my head to look behind me. I don't suppose that's going to get any better in the future. If my car can start driving me around when I want it to, great!

I do expect this to be my last car purchase. I assume autonomous car sharing or taxis will be more economical for occasional use in the future. So I'll be renting or subscribing or something different then.

 

[harley]

One of the possible improvements that people often mention with self driving cars is the "on demand" aspect where there won't be individually owned cars in every garage. I was reading an article on a technology risks newsletter I get that mentioned how that could backfire. If there's a hurricane or some other catastrophe and there are a limited number of vehicles, it will make evacuating much more difficult. Although perhaps the lack of traffic will allow the cars to make multiple trips. I'm not sure how that will play out.

Despite the fact that I absolutely love driving (I should probably have been a long haul trucker), I can't wait until self driving cars are available. Just the ability to relax on our twice yearly migration between MD and FL would be wonderful. And as much as I love to drive I realize that eventually I'll be even more of a risk than I am now. It would be nice to still be self reliant even after they pull my license. I'm sure I'll be driving pretty badly by the time they do that, since my license is from FL.

 

[meierlde]

On "a moment's notice", yes. But for many of those, humans fail today, so no difference. On less extreme situations, the vehicle could start slowing to a stop till the mere human takes over.

I suspect that the need to take over will be signaled by a vibrating seat (as Chevy Tahoes have) loud chimes and heads up displays. (as well as muting the sound system, essentially making the need totally obvious).

 

[mh]

One of the possible improvements that people often mention with self driving cars is the "on demand" aspect where there won't be individually owned cars in every garage. I was reading an article on a technology risks newsletter I get that mentioned how that could backfire. If there's a hurricane or some other catastrophe and there are a limited number of vehicles, it will make evacuating much more difficult. Although perhaps the lack of traffic will allow the cars to make multiple trips. I'm not sure how that will play out.

Despite the fact that I absolutely love driving (I should probably have been a long haul trucker), I can't wait until self driving cars are available. Just the ability to relax on our twice yearly migration between MD and FL would be wonderful. And as much as I love to drive I realize that eventually I'll be even more of a risk than I am now. It would be nice to still be self reliant even after they pull my license. I'm sure I'll be driving pretty badly by the time they do that, since my license is from FL.

I'd be willing to bet that a coordinated self driving car pool/share evacuation would actually be more efficient and safe than a everyone panic and jump in your car at the same time scheme.

 

[Midpack]

I'd be willing to bet that a coordinated self driving car pool/share evacuation would actually be more efficient and safe than a everyone panic and jump in your car at the same time scheme.

I'd agree, but I think he meant there would just be fewer cars per capita. That does seem likely when few if any of us have our own cars.

 

[savory]

Not sure what parking lots have to do with autonomous cars, that has to do with car/ride sharing, not self driving cars themselves so much.

But there won't be as much congestion, or the engineers will have failed completely.

• If there are 90% fewer accidents, there will be less congestion.
• A self driving car won't change lanes arbitrarily, those maneuvers will be more systematic. Cars will probably "talk to each other" sooner or later - already R&D underway.
• Self driving cars won't drive all different speeds, I suspect there will be default speeds for different traffic zones, and a decision tree to vary according to conditions - or something like that.
• Self driving cars won't be talking, reading or texting, under the influence, putting on makeup, half asleep, reaching in the backseat, dealing with kids, won't care about rain, snow or dust, etc.

There will be less congestion to be sure.

Parking lots and spaces have a ton to do with cars, environment and life quality. If riders could be dropped off at their destination, a car could go park in a garage for example. The on street parking, mall parking lots, etc could be replaced by various types of parking alternatives. Think of your local mall being sorrounded by a park vs 100s of concrete parking spaces. The car park might be a half mile away.

BTW, your taxes pay for the on street parking. We'll still pay for the parking structures I am guessing but safe to say they will be more efficient and cost less.

Lots of online information on this. A popular subject among bicycle advocates.

 

[exnavynuke]

I'm looking forward to the day when I can tell a car to take me to an appointment and it deals with rush hour traffic while I read a book or take a nap. 12 hour road-trips with no one having to concentrate on the road the whole time would be much more enjoyable. However, I can't imagine I'll plan on getting rid of my ability to go drive for enjoyment sometimes even if/when the rest becomes reality.

 

[Totoro]

Self-driving is great for people who:

• Cannot drive: elderly, small children, sight or otherwise impaired persons, sick persons.
• Shouldn't drive: drunk, high or emotional
• Don't want to drive
• Like to replace driving time with other times (reading, calling, sleeping, ..)

It isn't for people who:

• Want to drive themselves
• Live in remote areas with limited means: too expensive
• Like to own cars

If you think self-driving cars it's better to think of a universal bus system. Only with smaller buses and no driver. On-call, always available in most denser areas. With time delay or own-your-own in more remote areas. In fact, the bus system will become obsolete. Most regional trains too, but they'll probably stay since the infrastructure is built already.


Accidents will drop by a factor 10, the car manufacturer will most likely self-insure. Millions of lives will be saved, tens of millions injuries prevented. Broken cars or cars in trouble will call roadside assistance. Parking space inside cities will be freed up, possibly removed. Big quick-charge stations near power plants will likely be built if it's electric (and it will be). These stations will likely have a service center embedded. Many (most?) people will no longer own their car, but it will be a long while before a family with children gives up their station with all their family stuff in there. If the price is right, they'll still own it.


Regarding timeline and state of the art: still a ways out for full autonomy (no steering wheel), but the race is on. Most expect 2030 - 2040 to be the tipping point. Closed-off sites will be first: mining sites, predictable routes (intra-airport), .. Big problems are snow, rain etc .. expect the self-driving fleet to be temporarily unavailable when conditions are poor. Later it will be better.


Multi-tiered and application specific fleets will emerge: shared rides (think bus), luxury rides, school pickup/dropoff rides (with extra safety features). Just like today. I'm not counting on reduced congestion though. If a service becomes cheaper and easier to use, demand usually skyrockets ..

 

[tn3sport]

This is a healthy and interesting debate, but its fruitless. The autonomous car is coming whether we like it or not. There are big strong corps pushing this technology. (The auto makers and tech giants are all behind it.) That translates into big strong lobbying efforts. The lobbying effort will drive out self driving cars through regulation and expenses.

I've accepted its coming and loosely baked into my LTC plan for when I'm too old to safely drive. i.e. Get in the car and tell it to take me to the grocery, gym, doctor, etc.

 

[harley]

Self-driving is great for people who:

• Cannot drive: elderly, small children, sight or otherwise impaired persons, sick persons.
• Shouldn't drive: drunk, high or emotional
• Don't want to drive
• Like to replace driving time with other times (reading, calling, sleeping, ..)

It isn't for people who:

• Want to drive themselves
• Live in remote areas with limited means: too expensive
• Like to own cars

I would make few changes.

Shouldn't drive should include a lot more people than your list. Anyone without good situational awareness, which would include most people much of the time. My personal estimate is that 90% of accidents or near accidents are caused by people not paying proper attention (texting, conversing, listening to the radio or GPS, yelling at the kids, daydreaming, or just oblivious), so most of those shouldn't be driving. Most of the other 10% are caused by people over-estimating their abilities (speed, mostly), so they shouldn't be driving either.

And I don't see why living in a remote area would matter. If they own cars now, they could own one that does self driving. They don't all have to be garage based. That's just a way to make things more convenient for those that want it that way. If their means are too limited to own a car at all, then there's no difference.

And those that like to drive themselves? Fine. But maybe when they get into an accident, which would likely be their fault if most of the rest of the cars are self driving, then they get put into the "shouldn't drive" category, at least for a certain time. Better than jail, or a fine and then back on the road to do it again.

 

[Totoro]

Very good points Didn't want to come across as full-on fanboy. >99% of people would qualify with that amended list (and you are probably right).

Driving your own car will probably evolve to something akin to driving a horse: as a hobby done by a few enthusiasts, with special considerations if you want to take it on a public road.

Regarding remote living: it alters the economics quite a bit: self-driving equipment on a car is expected to cost roughly at least $10k, possibly $30k and more for a long while, so it at least doubles your expenses on the low end. And circumstances are a bit more variable, so it's harder to pull off 100% availability while you need it more.

So it will probably be cheaper and more convenient to own an old-fashioned one at first. But you are right, in the end it probably won't matter.