Author Topic: Full-self-driving needs external infrastructure  (Read 6236 times)

0 Members and 1 Guest are viewing this topic.

Offline AlbertLTopic starter

  • Regular Contributor
  • *
  • Posts: 219
  • Country: us
Full-self-driving needs external infrastructure
« on: March 05, 2023, 06:12:21 am »
Some of the recently-reported Tesla FSD "fails" have reinforced my opinion that the vehicles need external supporting infrastructure to operate safely - they can't do it by relying solely on their own sensors.

Specifically, I mean (1) a database accessible by the vehicle, maintained in real-time, of temporary conditions such as work zones and accident scenes that affect driving through an area, and (2) vehicle-readable optical/radio beacons, passive markers and other devices at critical locations such as intersections, that convey static and dynamic information needed for navigation at that point.

We already have external infrastructure, in the form of GPS, which eliminates the need for a vehicle to read directional signs.  These ideas would extend the concept to other key aspects of driving.

For example, consider the problem of a FSD vehicle failing to react to a person directing traffic.  Instead of trying to teach the vehicle to visually recognize those situations, imagine if the construction worker or first responder could just pull out a tablet, call up a map, and draw a box around the area to be protected.  The restriction would immediately go into a database and be relayed to the affected vehicles.  Those nearby would be ordered to slow down and change lanes, while those further away would be directed to alternate routes.               

 
 
The following users thanked this post: Red Squirrel, RJSV

Online ataradov

  • Super Contributor
  • ***
  • Posts: 11905
  • Country: us
    • Personal site
Re: Full-self-driving needs external infrastructure
« Reply #1 on: March 05, 2023, 06:36:28 am »
Yes, for sure. But there is zero chance of that happening. US government stopped investing into public infrastructure (partially because they get sued by companies for that) and commercial companies will only have one goal - screw the competition and make as much money as possible. Then abandon it couple years later when they move on to the next fad.

Ideally there would also be some sort of a standard for machine-readable road marking and signs, but who is going to come up with this standard when half the states are trying to ban EVs. LOL.
« Last Edit: March 05, 2023, 06:38:31 am by ataradov »
Alex
 

Offline pcprogrammer

  • Super Contributor
  • ***
  • Posts: 4670
  • Country: nl
Re: Full-self-driving needs external infrastructure
« Reply #2 on: March 05, 2023, 06:47:35 am »
This was already known more then 25 years ago.

If you want something that is safer then a human driver you need a system that does not solely rely on it's own sensory. With GPS there is indeed extra information given to the system, but it is not reliable enough and does not work in tunnels. The driving system needs more data about the circumstances in it's direct presence. Like for instant the road can tell if it is icy, or that there is an obstruction around the corner.

A system I worked on in 1995-1996 for container transportation was designed around using road side beacons. Using IR data was send from one beacon to the next. Blocking a beam would then lead to the beacons next in line there is an obstruction ahead, and the vehicles, also picking up the IR communication, would know it has to stop. Here the intent was to have a private road with railroad tracks just for this system.

With better technology for data communication available it should be possible to create something more safe, but it takes a lot of investment.

Offline nctnico

  • Super Contributor
  • ***
  • Posts: 28429
  • Country: nl
    • NCT Developments
Re: Full-self-driving needs external infrastructure
« Reply #3 on: March 05, 2023, 07:49:42 pm »
We already have external infrastructure, in the form of GPS, which eliminates the need for a vehicle to read directional signs. 
Let me stop you right there. GPS is not suitable for self driving cars -period-. The accuracy is way too low for that. I've seen GPS (using a really good, industrial receiver + antenna) get the wrong location by +/-150 meters when used in a car. So where is the car is actually driving? It could be several streets away. Take a wrong exit on purpose and look for how long your navigation system thinks you are still on the road you should have been on. It gets worse at big intersections.

On top of that GPS is way too fragile for this purpose as well as it can easely be spoofed using a $15 SDR dongle in a laptop. Think about the amount of mayhem a terrorist can do by spoofing GPS.

However there are efforts to develop positioning systems that are suitable for self driving cars. For example: https://www.tudelft.nl/citg/over-faculteit/afdelingen/geoscience-remote-sensing/research/projects/supergps
« Last Edit: March 05, 2023, 07:57:54 pm by nctnico »
There are small lies, big lies and then there is what is on the screen of your oscilloscope.
 

Offline james_s

  • Super Contributor
  • ***
  • Posts: 21611
  • Country: us
Re: Full-self-driving needs external infrastructure
« Reply #4 on: March 05, 2023, 07:59:06 pm »
We already have external infrastructure, in the form of GPS, which eliminates the need for a vehicle to read directional signs.  These ideas would extend the concept to other key aspects of driving.

Have you ever tried using GPS in an urban environment where you are surrounded by tall buildings? It confuses the hell out of it.

I think it is obvious that Tesla will never achieve true FSD with the hardware they have. There are just too many edge cases, snow on the road, leaves, flooding, construction, worn markings, dirt/spray accumulation on the cameras, etc. There's just no way it will ever work consistently under less than ideal situations and for it to be practical it MUST work in all these cases. A human driver cannot be expected to take over at a moment's notice, that is not realistic at all.
 
The following users thanked this post: Stray Electron

Offline TomKatt

  • Frequent Contributor
  • **
  • Posts: 529
  • Country: us
Re: Full-self-driving needs external infrastructure
« Reply #5 on: March 06, 2023, 01:04:16 pm »
I think it is obvious that Tesla will never achieve true FSD with the hardware they have.
It's probably obvious to Musk as well.

But you'll never hear him admit that.   FSD is always just around the corner.
Several Species of Small Furry Animals Gathered Together in a Cave and Grooving with a PICt
 

Offline pcprogrammer

  • Super Contributor
  • ***
  • Posts: 4670
  • Country: nl
Re: Full-self-driving needs external infrastructure
« Reply #6 on: March 06, 2023, 01:21:49 pm »
But you'll never hear him admit that.   FSD is always just around the corner.

It was 25 years ago and even though they made some progress it will stay there, unless they change the approach to having a roadside support system like suggested in the original post. And then still things can go wrong.

A human is not perfect at driving and also causes accidents, but can be held accountable for actions taken. These systems are also not perfect but the problem is accountability. Whom to blame in case of an accident.

And even if they succeed in making a standalone system that performs better then a human driver, without the additional input it still won't be able to detect an obstacle around a corner quickly enough to avoid an accident.

Offline TomKatt

  • Frequent Contributor
  • **
  • Posts: 529
  • Country: us
Re: Full-self-driving needs external infrastructure
« Reply #7 on: March 06, 2023, 01:36:17 pm »
Whom to blame in case of an accident.
The legal test case hasn't arrived yet.  But it seems inevitable an incident will occur that decides that.

I suspect that manufacturers must feel they have some legal protection, otherwise they're putting themselves at risk.

Edit - or, possibly the manufacturers feel the profits are worth the risk.  I doubt that Ford was the only car company to calculate the profit/loss calculations for human lives lost due to an engineering design flaw...
« Last Edit: March 06, 2023, 01:38:11 pm by TomKatt »
Several Species of Small Furry Animals Gathered Together in a Cave and Grooving with a PICt
 

Offline nctnico

  • Super Contributor
  • ***
  • Posts: 28429
  • Country: nl
    • NCT Developments
Re: Full-self-driving needs external infrastructure
« Reply #8 on: March 06, 2023, 02:22:34 pm »
A human is not perfect at driving and also causes accidents, but can be held accountable for actions taken. These systems are also not perfect but the problem is accountability. Whom to blame in case of an accident.
That has already been sorted and put into laws in various countries. Bottom line: the manufacturer takes the blame.
There are small lies, big lies and then there is what is on the screen of your oscilloscope.
 

Offline tom66

  • Super Contributor
  • ***
  • Posts: 7336
  • Country: gb
  • Electronics Hobbyist & FPGA/Embedded Systems EE
Re: Full-self-driving needs external infrastructure
« Reply #9 on: March 06, 2023, 04:31:57 pm »
Why should it strictly be necessary to have external sensors?  The eventual goal might be zero fatalities and injuries, but at this point, anything better than humans should be good enough.

Waymo manage to drive taxis around Phoenix, AZ and SF bay area with no external sensors, and have had zero fatal or SDC-caused accidents in their history (the only accidents reported so far were due to other drivers colliding with Waymo vehicles.) (Source)

Humans manage to drive with eyes on what could more or less be termed a limited-rotation gimbal and with relatively advanced (for the animal kingdom) image processing and whatever neural network you could classify the brain as.  And in the US, these drivers kill about 40,000 per year, and between 10-20x that many serious injuries (depending on definition).  So any system that could achieve at least that level of safety would be better.  Why couldn't cameras or LIDAR with appropriate software reach at least that level of safety? It could likely be better, because it would never get distracted or drunk.

There isn't yet enough data to definitively conclude that Waymo or other SDCs are safer in terms of fatalities (they would need to drive well over 100 million miles to start to get to statistically significant data) but the early evidence suggests they do have far fewer crashes per mile, which would imply they are already at least as safe as human drivers, if not safer.
 

Online bdunham7

  • Super Contributor
  • ***
  • Posts: 8178
  • Country: us
Re: Full-self-driving needs external infrastructure
« Reply #10 on: March 06, 2023, 04:45:24 pm »
  And in the US, these drivers kill about 40,000 per year, and between 10-20x that many serious injuries (depending on definition).  So any system that could achieve at least that level of safety would be better.  Why couldn't cameras or LIDAR with appropriate software reach at least that level of safety? It could likely be better, because it would never get distracted or drunk.

The problem is that your statistics on human drivers includes drunk drivers, distracted drivers, drivers who flagrantly violate traffic laws, plainly incompetent drivers that haven't lost their licences as of yet and so forth.  I don't think that an automated system that does just slightly better than that is acceptable at all.  In fact I see absolutely no justification for self-driving unless it can do significantly better than an experienced, competent, wide-awake, undistracted and cautious human driver.  IMO, that would only be practical at current technology levels if you at least used some additional sensing (LIDAR, etc), massive self-updating databases and perhaps external guide systems or artifacts.  The sad part about Telsa FSD is that I think it probaby is possible, but they aren't doing it.
A 3.5 digit 4.5 digit 5 digit 5.5 digit 6.5 digit 7.5 digit DMM is good enough for most people.
 

Offline nctnico

  • Super Contributor
  • ***
  • Posts: 28429
  • Country: nl
    • NCT Developments
Re: Full-self-driving needs external infrastructure
« Reply #11 on: March 06, 2023, 05:05:32 pm »
  And in the US, these drivers kill about 40,000 per year, and between 10-20x that many serious injuries (depending on definition).  So any system that could achieve at least that level of safety would be better.  Why couldn't cameras or LIDAR with appropriate software reach at least that level of safety? It could likely be better, because it would never get distracted or drunk.

The problem is that your statistics on human drivers includes drunk drivers, distracted drivers, drivers who flagrantly violate traffic laws, plainly incompetent drivers that haven't lost their licences as of yet and so forth.  I don't think that an automated system that does just slightly better than that is acceptable at all.  In fact I see absolutely no justification for self-driving unless it can do significantly better than an experienced, competent, wide-awake, undistracted and cautious human driver.
Such drivers don't exist...

The truth between your and Tom66's posting is somewhere in the middle: on one hand doing self driving is very hard to do due to the interaction between drivers. Driving is way more than simple colission avoidance. OTOH a self driving system doesn't need to be perfect. It needs to be better than the average human driver (including drunks, retards and old people).
There are small lies, big lies and then there is what is on the screen of your oscilloscope.
 

Offline james_s

  • Super Contributor
  • ***
  • Posts: 21611
  • Country: us
Re: Full-self-driving needs external infrastructure
« Reply #12 on: March 06, 2023, 05:50:07 pm »
Such drivers don't exist...

The truth between your and Tom66's posting is somewhere in the middle: on one hand doing self driving is very hard to do due to the interaction between drivers. Driving is way more than simple colission avoidance. OTOH a self driving system doesn't need to be perfect. It needs to be better than the average human driver (including drunks, retards and old people).

I remember being taught way back in drivers ed that driving is primarily a social interaction rather than a mechanical task, and social interactions are something computers are not good at. Any automated driving system has to be substantially better than average drivers for it to be acceptable. If it turns out that manufactures are liable, then the obvious outcome is that they will dedicate huge resources to legal teams to crush anyone involved in an accident with one of these self driving cars that tries to claim the manufacture is at fault. Even in cases where it is blatantly the fault of the self driving car, the companies will be able to bleed people dry in the courts and coerce them into settling. Maybe it's different in other parts of the world but here it is all but impossible for a regular individual to take on a large corporate entity, an entity with enormous resources can game the court system until the little guy runs out of money and can't afford to fight further, or in many cases they know this will happen and do whatever they can to cut their losses. It will be disastrous.
 

Offline AVGresponding

  • Super Contributor
  • ***
  • Posts: 4931
  • Country: england
  • Exploring Rabbit Holes Since The 1970s
Re: Full-self-driving needs external infrastructure
« Reply #13 on: March 06, 2023, 05:59:33 pm »
Such drivers don't exist...

The truth between your and Tom66's posting is somewhere in the middle: on one hand doing self driving is very hard to do due to the interaction between drivers. Driving is way more than simple colission avoidance. OTOH a self driving system doesn't need to be perfect. It needs to be better than the average human driver (including drunks, retards and old people).

I remember being taught way back in drivers ed that driving is primarily a social interaction rather than a mechanical task, and social interactions are something computers are not good at. Any automated driving system has to be substantially better than average drivers for it to be acceptable. If it turns out that manufactures are liable, then the obvious outcome is that they will dedicate huge resources to legal teams to crush anyone involved in an accident with one of these self driving cars that tries to claim the manufacture is at fault. Even in cases where it is blatantly the fault of the self driving car, the companies will be able to bleed people dry in the courts and coerce them into settling. Maybe it's different in other parts of the world but here it is all but impossible for a regular individual to take on a large corporate entity, an entity with enormous resources can game the court system until the little guy runs out of money and can't afford to fight further, or in many cases they know this will happen and do whatever they can to cut their losses. It will be disastrous.

Hence the class action.
nuqDaq yuch Dapol?
Addiction count: Agilent-AVO-BlackStar-Brymen-Chauvin Arnoux-Fluke-GenRad-Hameg-HP-Keithley-IsoTech-Mastech-Megger-Metrix-Micronta-Racal-RFL-Siglent-Solartron-Tektronix-Thurlby-Time Electronics-TTi-UniT
 

Offline james_s

  • Super Contributor
  • ***
  • Posts: 21611
  • Country: us
Re: Full-self-driving needs external infrastructure
« Reply #14 on: March 06, 2023, 06:03:16 pm »
Hence the class action.

Ah yes, the windfall for the lawyers while giving the peasants a few pennies. A class action lawsuit is not going to help the random person that gets creamed by a self driving car, receives a debilitating injury and gets crushed in court by the manufacture of the car that caused the accident.

I think if they do manage to hold the manufactures liable they will go out of business. A self driving car that is slightly better than the average driver will cause hundreds if not thousands of accidents a day around the world. Many of those will be serious and damages will be many billions of dollars a year. It could put them out of business.
 

Online bdunham7

  • Super Contributor
  • ***
  • Posts: 8178
  • Country: us
Re: Full-self-driving needs external infrastructure
« Reply #15 on: March 06, 2023, 06:51:24 pm »
Such drivers don't exist...

I don't think that's true at all. 

Quote
It needs to be better than the average human driver (including drunks, retards and old people).

And this is the is the precise point of unresolvable disagreement as I already pointed out.  As long as videos and stories abound of self-driving cars making serious errors that I or any competent, awake driver wouldn't make--or even ones that humans imagine they wouldn't make--there isn't going to be acceptance of these systems, at least by me and most people my age.  From conversations I've had, the younger crowd seems to be more willing to go along with this madness.  That's fine for them, I suppose, as long as we pass laws making both the human and FSD fully and jointly liable for all damages.

As an additional point, current Tesla self-driving systems are a bit more like aviation autopilot systems in one disturbing way, one that I think should be disqualifying--they only work under conditions that they can handle and when they are over their head, they shut off and hand the controls back to the human.  To me this is a pretty clear admission that the human is the superior driver.  FSD is claiming all of the safe miles driven but then it only works in situations where humans would be fairly safe as well as long as they are sober and attentive.  Under more challenging conditions, it quits.  Humans have to deal with 100% of driving situations without cherry-picking.  When someone develops a system that works the other way around and has a button I can press so the car takes over when I'm in over my head in a low-visibility or other emergency situation, then I'll be impressed.  Volvo's 'City Safety' system is an example of a system that goes in that direction as it intercedes when it detects a driver error, but it isn't FSD and certainly isn't marketed as such.

A 3.5 digit 4.5 digit 5 digit 5.5 digit 6.5 digit 7.5 digit DMM is good enough for most people.
 
The following users thanked this post: james_s

Offline pcprogrammer

  • Super Contributor
  • ***
  • Posts: 4670
  • Country: nl
Re: Full-self-driving needs external infrastructure
« Reply #16 on: March 06, 2023, 07:00:57 pm »
I wonder if a computer system can be programmed or taught to anticipate. Something humans can and in lots of situations will be the reason accidents are avoided. We can make judgement on how someone is driving and decide to not overtake because the driver is possibly drunk or aggressive and could do stupid things leading to an accident when being overtaken. Or make judgment on other situations to allow for safe responses.

Call it a sixth sense, possibly non catch-able with whatever sensor there is out there.

Driving is not a simple task of stay between the lines and uphold the speed limit, and just turn the steering wheel when needed. It is looking at the bigger picture of what is happening around the vehicle.

By the way go back some 25 years and the self driving systems being worked on had inter vehicle communication to improve on safety. This way the self driving cars were able to adapt to each other. So the idea of external input is not so strange, but probably abandoned due to needing adaptation of every vehicle on the road to make it work. Just like ataradov wrote in the first response the realization of such a system is far to expensive.


Offline tom66

  • Super Contributor
  • ***
  • Posts: 7336
  • Country: gb
  • Electronics Hobbyist & FPGA/Embedded Systems EE
Re: Full-self-driving needs external infrastructure
« Reply #17 on: March 06, 2023, 07:20:49 pm »
The problem is that your statistics on human drivers includes drunk drivers, distracted drivers, drivers who flagrantly violate traffic laws, plainly incompetent drivers that haven't lost their licences as of yet and so forth.  I don't think that an automated system that does just slightly better than that is acceptable at all.  In fact I see absolutely no justification for self-driving unless it can do significantly better than an experienced, competent, wide-awake, undistracted and cautious human driver.  IMO, that would only be practical at current technology levels if you at least used some additional sensing (LIDAR, etc), massive self-updating databases and perhaps external guide systems or artifacts.  The sad part about Telsa FSD is that I think it probaby is possible, but they aren't doing it.

The problem is, a significant number of people on the roads *are* bad drivers.  It's one thing to say we should only have good drivers on the road, but the fact is that we don't, given that in most countries they kill thousands of people per year, and this is considered a *good* amount compared to what it used to be.  And it's common for those drivers to be so-called professional drivers; truck, taxi and van drivers and the like.  Something about driving every day makes people less careful and attentive.

Compared to railways, it is crazy how much we accept the death toll on the roads;  in the UK the last major rail accident killed 31, that was 24 years ago, attributed ultimately to driver error, passing a signal at danger.  Since then, we've only had single digit fatalities in the odd accident here and there.  The railway network is about 10x safer per passenger-km than the roads.

So if the bad and indifferent car drivers could be automated out - those are the people who likely don't care about what or how they drive, they just use a vehicle to get from A to B - then road safety could surely be improved. 

To be clear, if all cars tomorrow were SDC's and we still had 40k deaths (in the US as an example) I'd be unhappy, because that's an unacceptably large figure.  But if it turned out SDC's could get that down to half, I think I'd be pretty happy with that outcome, knowing that there should be no fundamental reason incremental improvement cannot get to nearly zero fatalities eventually. 

Such drivers don't exist...

The truth between your and Tom66's posting is somewhere in the middle: on one hand doing self driving is very hard to do due to the interaction between drivers. Driving is way more than simple colission avoidance. OTOH a self driving system doesn't need to be perfect. It needs to be better than the average human driver (including drunks, retards and old people).

I remember being taught way back in drivers ed that driving is primarily a social interaction rather than a mechanical task, and social interactions are something computers are not good at. Any automated driving system has to be substantially better than average drivers for it to be acceptable. If it turns out that manufactures are liable, then the obvious outcome is that they will dedicate huge resources to legal teams to crush anyone involved in an accident with one of these self driving cars that tries to claim the manufacture is at fault. Even in cases where it is blatantly the fault of the self driving car, the companies will be able to bleed people dry in the courts and coerce them into settling. Maybe it's different in other parts of the world but here it is all but impossible for a regular individual to take on a large corporate entity, an entity with enormous resources can game the court system until the little guy runs out of money and can't afford to fight further, or in many cases they know this will happen and do whatever they can to cut their losses. It will be disastrous.

If you have an accident in the UK it is very common for the insurance company to try to identify immediately if you are at fault or not, and to settle the case.  For instance, if you as a driver run in to the rear of someone, barring a few particular mitigating circumstances, you are assumed to be liable.  If you merge onto a highway and collide with a vehicle on that highway, you are at fault.  If you kill someone, you are extremely likely to face court unless the facts of the matter are so clear that the police cannot see a reasonable chance that you would be convicted.

With the amount of datalogging these SDC's do, I am sure that the first fatality from a mainstream operational vehicle will be examined in great detail.  And the insurance company will probably pay out the costs to the victims and settle the case quickly.  Most insurance companies try to avoid going through the courts as it's very expensive, so there are industry agreed rates on costs for injuries, fatalities, etc.  No doubt there will be the odd court case and class action lawsuit but I don't see how this would be any different to how current motoring offences are handled.

 

Offline AndyC_772

  • Super Contributor
  • ***
  • Posts: 4315
  • Country: gb
  • Professional design engineer
    • Cawte Engineering | Reliable Electronics
Re: Full-self-driving needs external infrastructure
« Reply #18 on: March 06, 2023, 07:42:35 pm »
I wouldn't trust insurance companies to apportion blame in a way which is scientifically useful.

Some years ago I came off a motorcycle on a roundabout. It was covered in diesel - which, for the benefit of those unfamiliar with motorcycling - is extremely slippery and known to be one of the worst hazards out there for riders of two wheelers. Here in the UK, depositing anything on the highway which causes someone to have an accident is a criminal offence.

I called the insurance company to report the incident.

"Whose fault was it?" they asked.

"Whichever idiot over filled their fuel tank and didn't put the cap back on properly", I replied.

"I need a name"

"What difference does it make whether or not I can name the specific, individual person that dumped oil all over the road?"

"If you can't provide a name, then we'll blame you".

This, my friends, is how to lose a long term customer in the space of about two minutes.
 
The following users thanked this post: james_s

Offline TimFox

  • Super Contributor
  • ***
  • Posts: 9003
  • Country: us
  • Retired, now restoring antique test equipment
Re: Full-self-driving needs external infrastructure
« Reply #19 on: March 06, 2023, 10:03:42 pm »
Recent experience on railroads shows that external infrastructure for traffic control requires vigilant maintenance.
 
The following users thanked this post: bdunham7

Offline tom66

  • Super Contributor
  • ***
  • Posts: 7336
  • Country: gb
  • Electronics Hobbyist & FPGA/Embedded Systems EE
Re: Full-self-driving needs external infrastructure
« Reply #20 on: March 06, 2023, 10:20:29 pm »
I wouldn't trust insurance companies to apportion blame in a way which is scientifically useful.

Some years ago I came off a motorcycle on a roundabout. It was covered in diesel - which, for the benefit of those unfamiliar with motorcycling - is extremely slippery and known to be one of the worst hazards out there for riders of two wheelers. Here in the UK, depositing anything on the highway which causes someone to have an accident is a criminal offence.

I called the insurance company to report the incident.

"Whose fault was it?" they asked.

"Whichever idiot over filled their fuel tank and didn't put the cap back on properly", I replied.

"I need a name"

"What difference does it make whether or not I can name the specific, individual person that dumped oil all over the road?"

"If you can't provide a name, then we'll blame you".

This, my friends, is how to lose a long term customer in the space of about two minutes.

That's just down to how insurance works though.  An at-fault claim means your insurer had to pay out.  Since you came off your bike,  presumably damaging it in the process,  and there was no other person to take the blame for the accident then your insurer pays out.  Statistically, people who make at-fault claims are more likely to make more at-fault claims in the future.  Now, should there have been CCTV or some other way to prove who dropped the diesel, the claim could be put against them and might have been settled as a non-fault on your part. 

Every insurer will do this, so you're going to be disappointed should you need to make a similar complaint in the future.  If you didn't want the at-fault claim, you may elect to pay for the damages yourself if they're not too large.  This is what I did when I reversed into someone's 4x4 in a car park - oops - we traded details and I offered £200 to look the other way, which was accepted.  I had to notify my insurer I had done this, and had everything in writing with the other party, but it was not logged as an at-fault claim and come renewal time my premium didn't budge.

As for self-driving cars, I am sure the first few accidents will have significant regulatory oversight. Insurers may want to settle quickly but you can bet the NHTSA or equivalent will look into the incidents in considerable detail.  Remember the current insurance system is more built for efficiency and the balance of probabilities rather than finding the true answer every time - no insurance company would have the budget to investigate your diesel spill.  Ultimately it's just money, and civil claims in most common law countries are settled on the balance of probabilities rather than a criminal trial which is usually on beyond reasonable doubt.
 

Offline james_s

  • Super Contributor
  • ***
  • Posts: 21611
  • Country: us
Re: Full-self-driving needs external infrastructure
« Reply #21 on: March 06, 2023, 10:40:47 pm »
A friend of mine thinks that sooner or later there will be an incident where a self driving car kills a bunch of people and then the whole thing will be outlawed. I tend to agree, but I'll watch and see what happens. For a while at least a lot of companies were jumping in and trying to rush this tech out into the world, and there is a lot of hubris. All it's going to take is an incident where one of these cars gets confused and plows into a big group of cyclists, a parade route or event in a park and mows down a dozen people. It actually doesn't even matter if they're statistically safer because people fail at statistical analysis. People worry about EV battery fires even though ICE cars catch fire at far greater rates. They worry about shark attacks even though far more people die from drowning in the water than from sharks. They fear flying because the plane could crash and yet driving to the airport is far more likely to result in a fatal crash.
« Last Edit: March 06, 2023, 10:42:27 pm by james_s »
 

Offline AndyC_772

  • Super Contributor
  • ***
  • Posts: 4315
  • Country: gb
  • Professional design engineer
    • Cawte Engineering | Reliable Electronics
Re: Full-self-driving needs external infrastructure
« Reply #22 on: March 06, 2023, 10:43:41 pm »
That's just down to how insurance works though.  An at-fault claim means your insurer had to pay out.  Since you came off your bike,  presumably damaging it in the process,  and there was no other person to take the blame for the accident then your insurer pays out.

My point is, there's a really important subtlety to the terminology here.

Was I, in practical engineering terms, "at fault?"

As in, was there some identifiable, describable fault with my actions that caused the accident? And if there was, what exactly was it? What was the fault?

In this case, there really wasn't one. The fault was with whoever covered the road in diesel. Yet, as you point out, the way insurance works is they insist on this idea that whichever named person is "at fault" is the one who pays out - and in the absence of a name, as in this case, they fall back to an error handler which intentionally breaks the database.

The list of individuals labelled "at fault" no longer corresponds to the list of individuals whose driving exhibited a fault, and if information from the database were used to inform developers of self-driving equipment, somewhere there would be an engineer tasked with making a bike not fall over when it loses all grip at one of its wheels. Good luck with that.

Offline tom66

  • Super Contributor
  • ***
  • Posts: 7336
  • Country: gb
  • Electronics Hobbyist & FPGA/Embedded Systems EE
Re: Full-self-driving needs external infrastructure
« Reply #23 on: March 06, 2023, 11:09:48 pm »
That's just down to how insurance works though.  An at-fault claim means your insurer had to pay out.  Since you came off your bike,  presumably damaging it in the process,  and there was no other person to take the blame for the accident then your insurer pays out.

My point is, there's a really important subtlety to the terminology here.

Was I, in practical engineering terms, "at fault?"

No.  You were not, in the strictest sense, legal or otherwise, at fault.  But...

In this case, there really wasn't one. The fault was with whoever covered the road in diesel. Yet, as you point out, the way insurance works is they insist on this idea that whichever named person is "at fault" is the one who pays out - and in the absence of a name, as in this case, they fall back to an error handler which intentionally breaks the database.

The list of individuals labelled "at fault" no longer corresponds to the list of individuals whose driving exhibited a fault, and if information from the database were used to inform developers of self-driving equipment, somewhere there would be an engineer tasked with making a bike not fall over when it loses all grip at one of its wheels. Good luck with that.

... as far as the insurer is concerned, you cost them £x more that day than the day before.  They had to cut you a cheque, or pay for repairs to the bike.  Injuries could have occurred.  As far as the insurer is concerned, you are now a higher risk.

The granularity of this is poor, but look at it from the insurer's perspective (I know, hard to see insurers as reasonable, but work with me here.)  Before you have an accident, all an insurer knows is, your age, your vehicle, where you likely use that vehicle and maybe some ancillary details like your profession and employer.  They build a risk model with that information, make a bet that they'll only have to pay out, say, £500 on average that year, work out their admin cost and present you with a premium for, say, £550.   If you go and have an accident, then you change that risk model.  Even if that's an accident which isn't fundamentally your *actual* fault, they might argue if you were travelling 10 mph slower, or at a different time of day, or on different roads, you would be less likely to be involved, and not made the claim.  It would be nice if the model took more information in, but they've probably only just moved on from giant actuary tables for risk assessment.  Insurance companies move slowly!

If you want better granularity, you get things like black boxes which micro-analyse your driving.  This often results in premiums reducing for genuinely careful drivers, but suddenly those drivers who are 'worse' find their premiums going up because they were previously lying nicely in the middle of that distribution of risk benefiting from those who drive safer on average keeping their premiums low.  (Side note - the lack of transparency of these algorithms is one reason I will never get one of these boxes fitted.) 

A friend of mine thinks that sooner or later there will be an incident where a self driving car kills a bunch of people and then the whole thing will be outlawed. I tend to agree, but I'll watch and see what happens. For a while at least a lot of companies were jumping in and trying to rush this tech out into the world, and there is a lot of hubris. All it's going to take is an incident where one of these cars gets confused and plows into a big group of cyclists, a parade route or event in a park and mows down a dozen people. It actually doesn't even matter if they're statistically safer because people fail at statistical analysis. People worry about EV battery fires even though ICE cars catch fire at far greater rates. They worry about shark attacks even though far more people die from drowning in the water than from sharks. They fear flying because the plane could crash and yet driving to the airport is far more likely to result in a fatal crash.

I think your friend misunderstands just how cautious SDC's are, if anything they are too cautious compared to human drivers.  Tesla's for instance phantom brake sometimes when they perceive pedestrians on the road regardless of whether they are really there - the NN ends up being overly sensitive to these objects because it's hyper-trained on them. The Waymo car will pretty much not move if there is a pedestrian standing near to it.  I could actually foresee very effective protests against SDC's by say taxi drivers now out of a job, you merely need to get close to one to interfere with its operation and it will shut right down.  There have been some experiments with fake stop signs on billboards to interfere with SDC's, which will need some interesting workarounds.  I actually think we will get to the point where we have remote human drivers to 'unstick' vehicles that get stuck, but the number of these operators will be many fewer than cars on the road so it will still work out as a net win from a labour savings perspective. And that's really what SDC's are about from a commercial perspective, Waymo et al. don't really care that much about safety, other than not wanting to be embroiled in lawsuits, they want to run taxis where you don't need to employ people to run them.

I could see freak accidents like that Uber SDC which ran down the cyclist - Uber tried to pin the blame on the safety driver who was texting while driving but looking at the footage, the cyclist was wearing all black pushing a black bike at night with no lights, and it's hard to see how any driver could have seen that either - but those won't stop the rollout of SDC's. 

SDC's won't be outlawed because the US knows it needs to stay ahead in this field to remain competitive with the world.  But it will create a huge rift in the labour market and it's alarming that there seems to be no plan to cope with this sudden loss of employment.  I think the statistic is that in about 30 out of the 50 states, driving a truck is the most common job.

 

Offline TomKatt

  • Frequent Contributor
  • **
  • Posts: 529
  • Country: us
Re: Full-self-driving needs external infrastructure
« Reply #24 on: March 07, 2023, 11:26:07 am »
Slightly OT - FSD strikes me as a technology that manufacturers (or nerdy billionaires) want to create for consumers much more than consumers are clamoring for some new technology...

One way or another, we are all paying for the development of technologies like this, when I wonder how much people really want them.  Add to that possible requirements for investments into physical infrastructure and consumers are paying for things many of them may not even want.   
Several Species of Small Furry Animals Gathered Together in a Cave and Grooving with a PICt
 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf