Self Driving Cars?

[Midpack]

Lots of interesting articles on the transition and eventually outlawing today's human driven vehicles, relegated to 'amusement parks.' Not projected in our lifetimes, but we may experience some impact.

https://www.yahoo.com/news/humans-still-drive-autonomous-future-172010996.html

As a result, I am absolutely convinced that human driving as we know it will be outlawed, beginning in major urban centers in the first world before spidering out across major arteries to form regional and national autonomous transportation networks linked with multi-modal nodes.

Eventually, the professional wasteland of racing/driving schools will meet Disneyland in the middle. Racetracks will become mechanical petting zoos and amusement parks, with attendees signing waivers before they risk their lives in quaint machines under semi-controlled circumstances. Imagine Ferrari World was the only place you could drive a Ferrari.

 

[NW-Bound]

But the person associated with the licence plate (the registered owner) may not necessarily be the person driving the vehicle, committing the offences. You can't charge someone with a criminal offense because their car was seen being driven dangerously...

Yes, I know about this. So, when travelover said that the cars have Web access and can report to the police to intercept and apprehend bad drivers, I then said that we would need the data logs from the reporting cars to be admissible evidence in court.

About using dash cams and such, that is only usable when someone hits you and both vehicles and drivers are seen there on the spot by the police. We were talking about futuristic prevention measures to stop bad drivers before they cause an accident.

 

[JoeWras]

But the person associated with the licence plate (the registered owner) may not necessarily be the person driving the vehicle, committing the offences. You can't charge someone with a criminal offense because their car was seen being driven dangerously.

"But... the photos!" I hear you say. A valid point, but as you noted, it would require changes to the law. But the required changes are much more substantial than you think. Privacy laws would have to change. Currently, the interior of your vehicle is considered to be private property, and it's illegal to photograph someone inside their private property without their consent. So change the law, right? OK, so now the inside of their car is community property, and photographs are admissible in court.

But wait - now you've change drug laws, too. Currently, when police search a vehicle and find contraband, and none of the passengers confess to owning it, they can charge the driver with it, because it was found on their "private property." But if the interior of our vehicles is suddenly deemed to be community property, then that contraband might as well have been laying in the parking lot. So go ahead and throw out all those drug/gun convictions. Is that a worthwhile tradeoff for jailing a few aggressive drivers who haven't even caused an accident? (Because if they've caused an accident, you don't need camera evidence - a cop will show up and charge them with reckless/impaired/distracted driving).

It's a tricky landscape to navigate, and be careful of unintended consequences. Sure, aggressive drivers are annoying, but let's keep things in perspective. If safety is your priority, then just let them go and keep your distance from them.

Yes, there's going to have to be some lawmaking and regulating. This is going to slow things down. I think the tech will be ahead.

And don't forget international agreements. There will have to be a certain degree of international consensus.

And finally, infrastructure. For true level 5, I assert infrastructure will have to be improved. This could take time.

It will be a process. For those of us who live a few more decades, we'll start seeing the process. Level 5 nirvana of having a car pick up an infirm person is probably beyond my lifetime.

But Uber exists today...

 

[Midpack]

It will be a process. For those of us who live a few more decades, we'll start seeing the process. Level 5 nirvana of having a car pick up an infirm person is probably beyond my lifetime.

Who knows, but there are some more credible industry sources projecting this will move (much) faster than the public may expect. Again, I love classic cars, but I am enjoying watching this evolve even though I fully expect classic cars will be outlawed on public roads sometime in the distant future.

Waymo is conspicously absent from the chart below. You could make the case Waymo could get to fully autonomous before any automaker of they want to. And all indications are they plan to partner with an auto maker, FCA first?

Chris Gerdes, Chief Innovation Officer for the DOT, has said he thinks 35 percent of the cars on the road in the United States will be self-driving in ten years.

 

[JoeWras]

Chris Gerdes, Chief Innovation Officer for the DOT, has said he thinks 35 percent of the cars on the road in the United States will be self-driving in ten years.

I want whatever he is taking.

We were just talking about the turn-over of the car buying public, and how this induces lag into the switch. To get to 35% in 10, people are going to have to start buying these exclusively in about 4 years.* That would require huge incentives.

What level? I suspect he is including level 2 and 3 in these numbers. I can believe that number if those are included.

* - Source. Wikipedia https://en.wikipedia.org/wiki/Passenger_vehicles_in_the_United_States

The study found that of vehicles in operation in the US, 38.3% were older than ten years, 22.3% were between seven and ten years old, 25.8% were between three and six years old and 13.5% were less than two years old. According to this study the majority of vehicles, 60.6%, of vehicles were older than seven years in 2001.[7]

 

[NW-Bound]

What level? I suspect he is including level 2 and 3 in these numbers. I can believe that number if those are included.

+1

Waymo has no plans to build cars. I would imagine they will partner with multiple car makers. If they insist on Level 4, they will be out to the market later than other makers.

 

[Midpack]

I want whatever he is taking.

We were just talking about the turn-over of the car buying public, and how this induces lag into the switch. To get to 35% in 10, people are going to have to start buying these exclusively in about 4 years.* That would require huge incentives.

What level? I suspect he is including level 2 and 3 in these numbers. I can believe that number if those are included.

* - Source. Wikipedia https://en.wikipedia.org/wiki/Passenger_vehicles_in_the_United_States

Some of that was discussed in the article, link in #476, I didn't show the link again a few posts later.

There are people who assume this will take much longer than most industry projections. My point was simply there are estimates, from industry people who know (or should) more than any of us. There are plenty of others, Gerdes may be the most optimistic. Many innovations take us by surprise, timing and rate of adoption. We'll all choose whichever view makes sense to us. I won't defend or dispute here, none of us know better...and many of us are repeating ourselves a lot already (self included).

 

[JoeWras]

Some of that was discussed in the article, link in #476, I didn't show the link again a few posts later.

There are people who assume this will take much longer than most industry projections. My point was simply there are estimates, from industry people who know (or should) more than any of us. There are plenty of others, Gerdes may be the most optimistic. Many innovations take us by surprise, timing and rate of adoption. We'll all choose whichever view makes sense to us. I won't defend or dispute here, none of us know better...and many of us are repeating ourselves a lot already (self included).

Yes, I agree with what I bolded in your statement. Especially for items in the lower price ranges (phones, TVs, for example).

The problem with cars is the cost of entry. Even though people may want one, they may not be able to afford or qualify for a loan for one. This induces some lag into the uptake system.

That, and of course, the ramp up of the sensor manufacturers, etc.

There is clearly a desire by a large population of people for these cars. One only need look at the Tesla 3 deposits.

But actually laying down the cash or getting the loan is another matter.

 

[Midpack]

Yes, I agree with what I bolded in your statement. Especially for items in the lower price ranges (phones, TVs, for example).

The problem with cars is the cost of entry. Even though people may want one, they may not be able to afford or qualify for a loan for one. This induces some lag into the uptake system.

That, and of course, the ramp up of the sensor manufacturers, etc.

There is clearly a desire by a large population of people for these cars. One only need look at the Tesla 3 deposits.

But actually laying down the cash or getting the loan is another matter.

Could be. If your assumption is fewer buyers will qualify because self driving cars will be relatively expensive, you're probably right. Or that declining median incomes are just making new car ownership generally more difficult, probably a factor. But if it's financing, the auto industry has never had a problem finding a way to hook buyers, longer terms, leasing, etc.

And millenials aren't showing a fascination with cars like our generation did. Whether that's because of no money, or less interest...probably both?

 

[Texas Proud]

Except that's not what I've said. Again, "if level 4-5 cars actually reduce accidents by 80-90% vs today's accident/fatality standards as some industry folks estimate."

So it's not that anything 'skyrockets,' level 4-5 cars drop precipitously while level 0-2 cars remain the same - so the latter cars become relatively unsafe.

It's widely reported that 90% of accidents today are attributed to human error, I'm not sure that will change.

You could even make the argument level 0-2 drivers could become more complacent/inattentive assuming automous cars will avoid them (probably true). But that's also hypothetical. Another interesting dynamic we simply can't predict yet.

OK... I misread your post... I thought you were saying that the cost of insurance would be prohibitive... I think it will either remain the same or go down for the 0-2 driver... and if you can afford it now you can keep driving whatever you want...

I guess my point being that they do not have to outlaw older cars because they will become a non-event.... IOW, how many 1960s cars have seen in an accident lately Time will take care of this...

 

[explanade]

Maybe you can recoup some of the cost of a level 4/5 car by renting it out when you're not using it as a SDC Uber.

 

[eroscott]

Recommended test to see how it feels if Tesla AP1 (1st gen HW 2nd gen SW) occasionally fails.

A good test would be to drive down a highway on normal cruise control and with you having a light grip on the steering wheel and focusing on staying in the center of the lane. Let your trusted passengers keep their hand on your steering wheel while you are doing this and within a 1-2 minute timeframe (enough for your to think you are in control), have them minorly adjust the steering wheel WITHOUT them giving you any warning. Your eyes will be watching the road straight ahead all this time so you will sense or feel the steering wheel move WITHOUT your input.

You will realize it is happening VERY VERY quickly and react!!!

Thanks, eroscott for the post. If I interpret what you were saying, it's that when the car "loses lock", if you will, it drifts gently and not violently as I feared.

There's a big difference between the above and a sudden twitch, like someone jerking the steering wheel. I have more concern of a more violent upset, like when my RV was hit broadside by a gust of wind. I was able to handle that, else would not be posting. But it is very tiring and uncomfortable, even though they have road signs warning of gusty areas ahead for me to be prepared.

What I occasionally get is the same as the test below would demonstrate to you. Try it. Rarely does Tesla AP1 flat out abort and when it does ther are loud noises and your driver info center changes to a large red steering wheel with large hands on it and it telling you to take full control immediately. It does not jerk. This is pretty rare and I haven't seen it for months. The error now are a lot more sublime like can be demonstrated. Plenty of time to take over and you feel it immediately. These don't even raise my heart rate or concern me.

Here is an actual example of when AutoPilot gave me the loud sounds to take over and large red steering wheel w/hands on it. I've taken a route to get my son from college several times. The first time autopilot "failed" when going through this (picture below). Imagine the sensor readings! It did not jerk and since my hands were on the wheel I just continued to drive straight.

WHAT is very interesting is that it LEARNED that this is easily passable and NOW AutoPilot goes through it normally!!! It just did it correctly a couple days ago. It LEARNED this was an OK location. Sure the ultrasonic side sensors light up bright because it is close to large cement structures.

KEEP IN MIND that I've driven AutoPilot likely 70-80% of my 27K miles. I use it a LOT. Remember it reads 2 cars ahead even when I can't see the second car. This happened TWICE last night! I could not see around the first car but it was showing the car ahead of it on my display.

Tesla AutoPilot works amazingly well in NUMEROUS situations...
Autopilot does not need perfect highway lines or weather (note to ERD50). I've driven on modest rural 2 lane roads, urban 45mph 4 lane roads (a lot) with patches, mix of good bad lines, and in very very bad rain storms where I even had a hard time seeing (I slowed down and had wipers on max speed but AutoPilot continued to work and I thought it would brake faster with its radar seeing better than me!).

It works on wet roads where lines are hard to see, it works when the sun is in your eyes and reflecting on the roads and lines which makes them hard to see, it works with curbs and no lines, it looks for differences in material and color. It estimates the width of the roads based on paint showing up sometimes, it recognizes lanes and path trajectory based on cars traveling in front of and around it. It takes NUMEROUS things into account to help it understand the environment and lanes.

 

[Midpack]

OK... I misread your post... I thought you were saying that the cost of insurance would be prohibitive... I think it will either remain the same or go down for the 0-2 driver... and if you can afford it now you can keep driving whatever you want...

I guess my point being that they do not have to outlaw older cars because they will become a non-event.... IOW, how many 1960s cars have seen in an accident lately Time will take care of this...

We agree to disagree on both issues, no biggie.

 

[NW-Bound]

...What I occasionally get is the same as the test below would demonstrate to you. Try it. Rarely does Tesla AP1 flat out abort and when it does ther are loud noises and your driver info center changes to a large red steering wheel with large hands on it and it telling you to take full control immediately. It does not jerk. This is pretty rare and I haven't seen it for months. The error now are a lot more sublime like can be demonstrated. Plenty of time to take over and you feel it immediately. These don't even raise my heart rate or concern me.

Here is an actual example of when AutoPilot gave me the loud sounds to take over and large red steering wheel w/hands on it. I've taken a route to get my son from college several times. The first time autopilot "failed" when going through this (picture below). Imagine the sensor readings! It did not jerk and since my hands were on the wheel I just continued to drive straight.

WHAT is very interesting is that it LEARNED that this is easily passable and NOW AutoPilot goes through it normally!!! It just did it correctly a couple days ago. It LEARNED this was an OK location. Sure the ultrasonic side sensors light up bright because it is close to large cement structures.

KEEP IN MIND that I've driven AutoPilot likely 70-80% of my 27K miles. I use it a LOT. Remember it reads 2 cars ahead even when I can't see the second car. This happened TWICE last night! I could not see around the first car but it was showing the car ahead of it on my display.

Tesla AutoPilot works amazingly well in NUMEROUS situations...
Autopilot does not need perfect highway lines or weather (note to ERD50). I've driven on modest rural 2 lane roads, urban 45mph 4 lane roads (a lot) with patches, mix of good bad lines, and in very very bad rain storms where I even had a hard time seeing (I slowed down and had wipers on max speed but AutoPilot continued to work and I thought it would brake faster with its radar seeing better than me!).

It works on wet roads where lines are hard to see, it works when the sun is in your eyes and reflecting on the roads and lines which makes them hard to see, it works with curbs and no lines, it looks for differences in material and color. It estimates the width of the roads based on paint showing up sometimes, it recognizes lanes and path trajectory based on cars traveling in front of and around it. It takes NUMEROUS things into account to help it understand the environment and lanes.

Very interesting to know.

Now, perhaps Tesla knew that its first clientele was mostly geeky types and enthusiastic early male adopters who don't mind experiment and provide feedback for them to fine tune it. And it seems to work out OK.

And then, there's Ford and GM who, if they released the same initially limited capabilities to somebody like my 84-year old mother, they would do a lot of harm.

So, both sides are right.

I am glad that you contribute to this thread.

 

[NW-Bound]

Waymo has no plans to build cars. I would imagine they will partner with multiple car makers. If they insist on Level 4, they will be out to the market later than other makers.

The chart that Midpack posted did not show Waymo. However, just a few months ago, Honda announced to be in talk with Waymo. And the chart shows Honda as introducing driverless cars on highway in 2020. This is a limited-mode level-4 I think.

Ford goes full Level-4 in 2021. And Ford has a joint venture with Waymo!

Wow! I am impressed. What I am waiting to see is how much these cars cost. I would not be surprised that the car makers lose money at first, but the price may be still quite high so that they will be low-volume loss leaders. That allows them to maintain brand visibility with the public, while they learn to make them cheaper.

I have to admit they are a lot more aggressive than I expected. I stand corrected.

 

[explanade]

Would be interesting to see how insurance price these cars.

On the one hand, they're suppose to be less susceptible to accidents.

OTOH, they carry a lot more gear which could be costly to repair/replace should you be in an accident.

Would a normal body shop be able to replace the parts of the body around where the sensors are housed?

 

[NW-Bound]

The logistics of these cars is something new. Maybe they will be leased only.

I recall a similar situation with GM EV1. My, that's 20 years ago. They brought a demo car to my megacorp. The car was packed full of lead-acid batteries. Horrible weight, and lousy range. These cars were leased only, after which they were recalled by GM and destroyed.

 

[eroscott]

The logistics of these cars is something new. Maybe they will be leased only.

I recall a similar situation with GM EV1. My, that's 20 years ago. They brought a demo car to my megacorp. The car was packed full of lead-acid batteries. Horrible weight, and lousy range. These cars were leased only, after which they were recalled by GM and destroyed.

If you sold them there are laws about having to support (parts) for them for a decade. The EV1 was very expensive and was tech was lousy for sure. Experiment that CA CARB forced.

Remember this was 20 years ago ... 1996.
https://en.wikipedia.org/wiki/General_Motors_EV1

Costs[edit]
GM based the lease payments for the EV1 on an initial vehicle price of US$33,995.[2] Lease payments ranged from around $299 to $574 per month, depending on the availability of state rebates.[citation needed] Since GM did not offer consumers the option to purchase at the end of the lease, the car's residual value was never established, making it impossible to determine the actual full purchase price or replacement value. One industry official said that each EV1 cost the company about US$80,000, including research, development and other associated costs;[64] other estimates placed the vehicle's actual cost as high as $100,000.[2] Bob Lutz, GM Vice Chairman responsible for the Chevrolet Volt, in November 2011 stated the EV1 cost $250,000 each and leased for just $300 per month.[65] GM stated the cost of the EV1 program at slightly less than $500 million before marketing and sales costs, and over $1 billion in total, although a portion of this cost was defrayed by the Clinton Administration's $1.25 billion Partnership for a New Generation of Vehicles (PNGV) program.[66][67][68] In addition, all manufacturers seeking to produce electric cars for market consumption also benefited from matching government funds committed to the United States Advanced Battery Consortium.

 

[NW-Bound]

Yes, I can see the same thing happening here with these first-batch level-4 cars. The true cost of them may not be known even to the makers until the end of the lease.

They will not know what it costs to repair, and to maintain them until they have done it.

 

[ERD50]

If you sold them there are laws about having to support (parts) for them for a decade. The EV1 was very expensive and was tech was lousy for sure. Experiment that CA CARB forced.

Remember this was 20 years ago ... 1996.
https://en.wikipedia.org/wiki/General_Motors_EV1

It gets to be a little bit of 'fuzzy math' when the development costs are included in a short run product, that really wasn't even intended to turn a profit. The CA CARB law 'experiment' said GM had to sell 1% (or some number?) of cars of that type in order to sell in CA at all. So it could be a loss leader for GM, just a cost of doing business. GM couldn't give up the entire CA market.

My personal take (and I have not found confirmation of this) on why GM leased rather than sold them is that if sold them at a price that enough people would buy them at to meet the CARB limit, the price of the car would be less than the cost of the components, and people would strip them (maybe put in a V8?) and sell them for parts! By leasing them, GM retained ownership.

Anyhow, it was a twisted market, and doomed to fail. When people claim that "GM killed the Electric Car", I say "No, it was the stillborn bastard child of CA CARB and GM".


-ERD50

 

[Midpack]

Lots already written re: insurance and self-driving cars. Just a few:

Self-Driving Cars and Insurance | III

Who Will Insure Self-Driving Cars?

How driverless cars could negatively affect the insurance industry - Business Insider

If you are an executive of an auto insurance company, pay attention — you may not have a business in 20 years. You can blame the fundamental shifts in auto safety and data mining that connected car and autonomous vehicle technologies will bring. Robot drivers will outnumber humans behind the wheel. The remaining human drivers will be safer, thanks to collision-preventing sensors and analytics on board. Insurance claims will be rare, losses will be reduced, premiums will decline, and insurance companies will probably lose control of the data that makes their pricing models possible. Car owners might no longer purchase insurance directly. Instead, automakers would bundle insurance into each new car purchase, much as they do satellite radio and roadside service contracts today.

These types of structural business model changes don’t happen often, especially in a regulated sector like auto insurance. And although the shift isn’t imminent, it is practically inevitable. Already, driverless cars have moved from low-speed and test environments to limited use on public roads and highways. Safety and production issues need to be addressed before widespread adoption can occur, but the auto industry and the businesses that support it are pushing hard for change.

Global auto insurance is a US$700 billion market that represents 42 percent of global aggregate property and casualty insurance, according to Swiss Re. It is an extremely competitive market today, and it will be still more competitive by 2030. Research conducted by our DeNovo strategy consulting platform suggests that auto insurance companies will find themselves squeezed out of conventional auto insurance altogether. They would then have to find new sources of revenue: expanding into non-automotive fields such as commercial liability and cybersecurity coverage or, perhaps, partnering with car manufacturers and ride-sharing services.

 

[NW-Bound]

Would be interesting to see how insurance price these cars.

On the one hand, they're suppose to be less susceptible to accidents.

OTOH, they carry a lot more gear which could be costly to repair/replace should you be in an accident...

They don't cause crashes, but jerks still crash into them. The insurance may have no ideas how much to price the premium. The car makers may have to insure the cars themselves, or to work out a deal with the insurers.

 

[meierlde]

It would not surprise me if it takes 30 years before more than half the cars on the road are level 4-5. But that doesn't mean level 0-2 cars won't be outlawed on all or some public roads eventually - I doubt it will be left to drivers choice indefinitely. I'd think it will depend in large part on relative accident/fatality incidence rates. At the risk of repeating earlier discussion on this thread, all hypothetical:

• If level 4-5 cars actually reduce accidents by 80-90% vs today's accident/fatality standards as some industry folks estimate, insuring a level 0-2 car may slowly become prohibitively expensive.
• If so, that might effectively force some drivers to adopt car/ride sharing level 5 public or private transportation, assuming it's clearly more cost effective. Obviously much easier for urbanites than rural, so maybe urban drivers will lose their right to drive before rural.
• And if level 0-2 cars are that much more unsafe, eventually they could be legally barred from all or some public roads, urban roads presumably first. Manually driving a classic level 0 might be confined to private roads far in the future.
• Some day it might be much, much harder to get a license or registration for a level 0-2 car. Or eventually impossible - forcing classic cars onto private roads only.

Might take 60 years before manual cars are outlawed, probably not in our lifetimes, though we may live to see it in insurance rates. And it'll be telegraphed well in advance, maybe eliminating level 0, then 1, 2 and so on. It's going to be interesting.

Since 50% of todays fatalities are due to folks not wearing their seatbelt, (for 13% of drivers) that is the low hanging fruit. Note that todays forward crash mitigation systems tension the seat belts if a collision is expected.
So it appears that primary seat belt laws and moving to seat belt not fasten car does not move, are the current low hanging fruit for reducing fatalities.
I do wonder how many fatalities wearing seatbelts are due to vast mismatch in vehicle weights? If a small car takes on a semi it looses.
Now another insurance idea might be to equip all cars with recording dash cams. If you watch the retarded drivers videos on you tube you see a lot of crashes in Russia because traffic there appears to be wild west style, so you record in case of a crash. (Lots of running red lights, driving in the wrong lane etc)

 

[NW-Bound]

...Now another insurance idea might be to equip all cars with recording dash cams...

All self-driving cars, even level-2, have cameras. So, it's the matter of storing the last 10 seconds or so of pertinent sensor data.

Waymo cars will have 360-deg vision from the dome atop the roof. And as they track all moving objects around them, they "know" of all collisions happening in the vicinity. They can also store this data as a bystander to an accident. And that will be very useful as court evidence.

Eventually, when solid-state memory becomes cheaper, I can see these cars storing the last 24 hours, or 1 week, or 1 month, of what their vision cameras see, even if they detect no collision. Why?

So that the stored videos can be subpoenaed by the court if the cars happen to be where a crime is later determined to have occurred. This is good stuff. As the cars are Web connected, the court can query all cars to see which ones happen to drive by that particular location at that time. And when crime investigators track that suspect, who may be driving or be on foot, they can query for the next location he was at, and on, and on.

The stored videos make for a lot better evidence than human eyewitnesses right now, whose memory is often faulty, and biased as they recall a certain sequence of events.

 

[meierlde]

All self-driving cars, even level-2, have cameras. So, it's the matter of storing the last 10 seconds or so of pertinent sensor data.

Waymo cars will have 360-deg vision from the dome atop the roof. And as they track all moving objects around them, they "know" of all collisions happening in the vicinity. They can also store this data as a bystander to an accident. And that will be very useful as court evidence.

Eventually, when solid-state memory becomes cheaper, I can see these cars storing the last 24 hours, or 1 week, or 1 month, of their vision cameras, even if they see no collision. Why?

So that the stored videos can be subpoenaed by the court if the cars happen to be where a crime is later determined to have occurred. This is good stuff. As the cars are Web connected, the court can query all the cars to see which ones happen to drive by that particular location at that time. And when crime investigators track that suspect, who may be driving or be on foot, they can query for the next location he was at, and on, and on.

Note that modern cars have 30 second event recorders to monitor throttle brakes steering etc. The hold the last 30 seconds if air bag deployment happens. Of course there is a whole bunch of legal involving this that is still undecided, starting with who owns the data, and who can access it.