I like how we're at, "A lot of drivers are bad, so therefore we should use a computer driving system that is still bad but less bad," as though we've just completely given up on the idea that maybe human drivers should not be bad.
Well we've been trying for human drivers to not be bad for a century, so I don't know when you think we should give up.
Longer. People were crashing into things with horses or even just on their own two feet.
I believe that with sufficient regular training, that it MAY be possible to teach a human to walk safely without crashing into other humans!
There is a long way to go in the US. Other countries have way lower auto fatality rates - e.g. the US rate is like 2X the UK rate.
So this idea that you can't make things any better is bunk. We just choose not to.
That because the UK has a strict and rigorous program of regular vehicle inspection which checks vehicles for functionality and safety every year. This program effectively continuously removes old or poorly maintained vehicles from the road, and makes sure every vehicle has proper tire tread and good brakes every year. The UK also has roads which are better designed for safety, using roundabouts rather than stop signs etc.
Better road safety is a function of better road layout, typically shorter journeys and safer cars.
Yes, hence why I said "we choose not to" do any of those things, like design roads for better safety, design our built environment so people can make shorter trips or more trips without a car, etc.
But I never said you couldn't make things better. I said you can't make humans better at driving. You can absolutely make driving safer and easier for humans
This is objectively false. A trained driver is a better driver an untrained driver. A trained driver with ten years of experience will generally be superior to one with ten months of experience.
Better driver training, more frequent testing during license renewals, and a stronger driving culture will all lead to better drivers on the roads.
To a degree, yes, but there is a limit to how good a driver human can be, on average.
There will always be exceptionally good, and exceptionally bad, drivers, but average human will remain as just that.
A better trained driver will always be better than an untrained one, yes. However we are now approaching a possible reality where a computer will be able to drive better than either. At which point the question becomes why even bother for most people unless there’s a particular need.
I like how we're at, "A lot of drivers are bad, so therefore we should use a computer driving system that is still bad but less bad," as though we've just completely given up on the idea that maybe human drivers should not be bad.
Well we've been trying for human drivers to not be bad for a century, so I don't know when you think we should give up.
Longer. People were crashing into things with horses or even just on their own two feet.
I believe that with sufficient regular training, that it MAY be possible to teach a human to walk safely without crashing into other humans!
There is a long way to go in the US. Other countries have way lower auto fatality rates - e.g. the US rate is like 2X the UK rate.
So this idea that you can't make things any better is bunk. We just choose not to.
That because the UK has a strict and rigorous program of regular vehicle inspection which checks vehicles for functionality and safety every year. This program effectively continuously removes old or poorly maintained vehicles from the road, and makes sure every vehicle has proper tire tread and good brakes every year. The UK also has roads which are better designed for safety, using roundabouts rather than stop signs etc.
Better road safety is a function of better road layout, typically shorter journeys and safer cars.
Yes, hence why I said "we choose not to" do any of those things, like design roads for better safety, design our built environment so people can make shorter trips or more trips without a car, etc.
But I never said you couldn't make things better. I said you can't make humans better at driving. You can absolutely make driving safer and easier for humans
This is objectively false. A trained driver is a better driver an untrained driver. A trained driver with ten years of experience will generally be superior to one with ten months of experience.
Better driver training, more frequent testing during license renewals, and a stronger driving culture will all lead to better drivers on the roads.
Yes but ALL drivers become better with time, an individual driver will change their abilities as they gain more experience, changing from being a teen driver, to a young adult, and so on. This transition happens to everyone and doesn't affect the behavior of the driving population.
Places with higher road safety have better roads and better maintained vehicles (the UK), not better trained drivers. People don't use the techniques we DO teach during driver training, let alone the ones taught during advanced training.
Here's a good article on the subject. Advanced driver training makes people WORSE drivers
The evidence is that the only way to make people safer drivers is not to teach them how to drive in difficult conditions, or to teach them special techniques for safety. But to make them more concerned and aware of those difficult conditions and make them NOT drive in those situations. Peripheral road awareness (which is pretty much all the stuff you are talking about with looking at shadows and brakelights and considering upcoming curves) can't be trained, its a thing your teenagers don't have and that driving on the road will give them. For example, when you talk about looking at shadows and through windscreens and whatnot to see upcoming cars, you aren't actually doing that. It's just your brain being trained by 1000s of hours of driving experience noticing some weird 'flock' behavior in the cars ahead. You see, without really knowing why, that the cars ahead are odd and you start slowing down to avoid the oddness.
I mean, drivers can be trained a bit. The transition from 0 hours experience to 40 hours experience is something which can absolutely be helped by training. There might be something which could be done in early driving life to make sure people who are slow to gain this experience don't come onto the roads (teens, espescially boys shouldn't drive alone till 18, and driving before that should be done with their parents as part of supervised 'experience aquisition'. You could, I suppose make people retest every 12 months like aviation pilots, that might be enough to make a difference. But that test should focus overwhelmingly on safe and cautious driving techniques.
Slower
Fewer lane changes
Larger spacing
None of these things require any 'skill' just the attitude that every crash which happens near you is partially your fault.
"That is cool" - Abraham Lincoln
0
Options
CambiataCommander ShepardThe likes of which even GAWD has never seenRegistered Userregular
You mean, "British safety firm demonstrates why Autopilot behaves exactly as the user manual describes and is still safer than driving yourself in the situation they demonstrated"
I can assure you 100% that if there is a vehicle stopped in the road in front of you, and the vehicle ahead of you simply swerves around it without reducing speed or indicating and you haven't been lucky enough to get a good shadow indication of the stopped vehicle then the BEST human driver in the world is crashing into that vehicle, or swerving into traffic in lanes to the side. I would have crashed there. You would have crashed there. EVERYONE would have crashed there. When that driver in front moved aside at high speed at the last minute without deceleration he may as well have got out and shot the following driver in the gut.
Stopped vehicle (or large thing) in high speed free flowing traffic on a freeway, with a large vehicle in front of you, which hasn't yet transitioned to slow flow upstream of it is literally the most dangerous thing that you can run into on the road. Other than a stopped thing which might start moving perpendicular to you.
Uh-uh, nope. There is something very crucial you guys are missing when throwing up your hands and saying "a human would have crashed too!"
Spoiler for large image:
I borrowed this from a driver safety website, of all places. What do you notice about the green station wagon?
It has a window you can see through. There's a white vehicle moving with the flow of traffic a few hundred yards ahead of it.
If your method of driving is to stare at the bumper of the car in front of you and react to whatever it's doing, you are a horrible driver. You know what they pound into you over and over and over again if you go to any sort of performance driver training? Look ahead. No, look ahead. No, further than that. As far as you can possibly see. Your peripheral vision will take care of what's directly in front of the car, you should be looking as far down the road as you can. If you're coming up to a 90 degree turn, your head should be pointed at the side window, not the windshield. Look ahead.
A competent human driver would not have crashed. They would have noticed the stopped car from a long ways back.
And if you're behind a van or truck and can't see through it? Then you're following much too closely. A police officer at this accident would write the driver a ticket for failure to maintain a clear stopping distance.
Quote from the driver ed website I cribbed the photo from:
Highway traffic can decelerate from 65mph to 15mph in a very short time period. To avoid being caught off-guard by these quick speed changes, teach your teen to look for the brake lights of cars in the distance. By looking far down the highway (up to a quarter mile), they can get a heads up on speed changes.
As soon as your teen sees brakes lights coming on up ahead, they should immediately cover the brake. This will give them a little extra cushion in case they need to apply brake pressure.
I'm appalled there is pushback on this. Looking at the brakelights and behaviour of traffic as a whole is like one of the most basic driving skills people learn. You can absolutely see cars around the car in front of you except under certain circumstances (and that mostly comes down to SUVs being abominations that should be removed from the roads). Through the windows, over the top, around the sides because people don't drive in perfectly rigid straight lines, etc, etc.
If you are only watching the car in front of you, you are gonna rear-end people constantly because nobody actually follows at a distance where they can brake safely just looking at the car in front of them.
I think I am feeling skepticism because I drive a sedan in Texas, where nearly eveyone else on the road is driving a truck. I certainly try to look through the car ahead of me's window, but the opportunities to do so are rare.
Also, any time leave a safe distance between me and the car ahead, the traffic around me fills it. Fucking Dallas traffic.
"If you divide the whole world into just enemies and friends, you'll end up destroying everything" --Nausicaa of the Valley of Wind
I mean, statistics like this are the nail in the 'driver training can help' coffin...
"Evidence shows that in the USA the highest skilled drivers (registered race and rally car drivers) have a much higher crash rate than the average driver (Naatanen and Summala, 1976)."
And so on.
All the studies show that the only thing people can do to lower accident risks is to drive carefully and slowly.
You mean, "British safety firm demonstrates why Autopilot behaves exactly as the user manual describes and is still safer than driving yourself in the situation they demonstrated"
I can assure you 100% that if there is a vehicle stopped in the road in front of you, and the vehicle ahead of you simply swerves around it without reducing speed or indicating and you haven't been lucky enough to get a good shadow indication of the stopped vehicle then the BEST human driver in the world is crashing into that vehicle, or swerving into traffic in lanes to the side. I would have crashed there. You would have crashed there. EVERYONE would have crashed there. When that driver in front moved aside at high speed at the last minute without deceleration he may as well have got out and shot the following driver in the gut.
Stopped vehicle (or large thing) in high speed free flowing traffic on a freeway, with a large vehicle in front of you, which hasn't yet transitioned to slow flow upstream of it is literally the most dangerous thing that you can run into on the road. Other than a stopped thing which might start moving perpendicular to you.
Uh-uh, nope. There is something very crucial you guys are missing when throwing up your hands and saying "a human would have crashed too!"
Spoiler for large image:
I borrowed this from a driver safety website, of all places. What do you notice about the green station wagon?
It has a window you can see through. There's a white vehicle moving with the flow of traffic a few hundred yards ahead of it.
If your method of driving is to stare at the bumper of the car in front of you and react to whatever it's doing, you are a horrible driver. You know what they pound into you over and over and over again if you go to any sort of performance driver training? Look ahead. No, look ahead. No, further than that. As far as you can possibly see. Your peripheral vision will take care of what's directly in front of the car, you should be looking as far down the road as you can. If you're coming up to a 90 degree turn, your head should be pointed at the side window, not the windshield. Look ahead.
A competent human driver would not have crashed. They would have noticed the stopped car from a long ways back.
And if you're behind a van or truck and can't see through it? Then you're following much too closely. A police officer at this accident would write the driver a ticket for failure to maintain a clear stopping distance.
Quote from the driver ed website I cribbed the photo from:
Highway traffic can decelerate from 65mph to 15mph in a very short time period. To avoid being caught off-guard by these quick speed changes, teach your teen to look for the brake lights of cars in the distance. By looking far down the highway (up to a quarter mile), they can get a heads up on speed changes.
As soon as your teen sees brakes lights coming on up ahead, they should immediately cover the brake. This will give them a little extra cushion in case they need to apply brake pressure.
I'm appalled there is pushback on this. Looking at the brakelights and behaviour of traffic as a whole is like one of the most basic driving skills people learn. You can absolutely see cars around the car in front of you except under certain circumstances (and that mostly comes down to SUVs being abominations that should be removed from the roads). Through the windows, over the top, around the sides because people don't drive in perfectly rigid straight lines, etc, etc.
If you are only watching the car in front of you, you are gonna rear-end people constantly because nobody actually follows at a distance where they can brake safely just looking at the car in front of them.
I think I am feeling skepticism because I drive a sedan in Texas, where nearly eveyone else on the road is driving a truck. I certainly try to look through the car ahead of me's window, but the opportunities to do so are rare.
Also, any time leave a safe distance between me and the car ahead, the traffic around me fills it. Fucking Dallas traffic.
Yeah, I was driving home yesterday and at least 50% of the vehicles in the road are pretty much completely vision blocking. Ever tried to look through the rear window of a Prius?
The crash of an Uber self-driving car that killed an Arizona woman in March was “entirely avoidable,” according to police reports released by the Tempe Police Department. Cellphone data obtained by police suggests that the Uber operator was also streaming an episode of reality show The Voice at the time of the fatal incident.
The documents, released to Gizmodo in response to a public records request, show that Tempe police found that the operator of the Uber autonomous vehicle could likely have avoided the fatal crash, had she been paying attention—but instead she was likely watching a video on her phone. Police also noted that Uber’s vehicles apparently did not alert operators to take over the vehicle during incidents.
So, if they had been distracted by having to monitor the software as previously speculated I would have said that the operator was not particularly at fault but if this is true they should also be charged.
I wonder, for those of you who say that human beings are terrible drivers and they're never going to get any better, so we need autonomous vehicles to make the roads safer, how do you incorporate within that thesis the fact that the above incident and others are going to happen, that the safer we make drivers feel, the more leeway they're going to take, essentially creating an equilibrium dynamic with a resting point between "safety" and "freedom"? It seems likely to me that some sort of psychological reversal is going to occur, as we see with speed limits and with autopilot and a variety of other examples in tech-related domains, wherein people respond to gains from technological innovation by simply raising their expectations/lowering their effort levels?
Do you think stuff like this won't happen? Do you think stuff like this already happens (see: drivers on cell phones) and it won't get any worse? Do you think it'll still be a net gain? Etc?
That premise presumes a *fully* autonomous vehicle such that it doesn't matter how distracted the passenger is.
Most automous vehicle advocates say semi automous vehicles are a bad idea for exactly the reason this crash showed. You can't expect a human to react in time if they aren't actively driving.
The crash of an Uber self-driving car that killed an Arizona woman in March was “entirely avoidable,” according to police reports released by the Tempe Police Department. Cellphone data obtained by police suggests that the Uber operator was also streaming an episode of reality show The Voice at the time of the fatal incident.
The documents, released to Gizmodo in response to a public records request, show that Tempe police found that the operator of the Uber autonomous vehicle could likely have avoided the fatal crash, had she been paying attention—but instead she was likely watching a video on her phone. Police also noted that Uber’s vehicles apparently did not alert operators to take over the vehicle during incidents.
So, if they had been distracted by having to monitor the software as previously speculated I would have said that the operator was not particularly at fault but if this is true they should also be charged.
I wonder, for those of you who say that human beings are terrible drivers and they're never going to get any better, so we need autonomous vehicles to make the roads safer, how do you incorporate within that thesis the fact that the above incident and others are going to happen, that the safer we make drivers feel, the more leeway they're going to take, essentially creating an equilibrium dynamic with a resting point between "safety" and "freedom"? It seems likely to me that some sort of psychological reversal is going to occur, as we see with speed limits and with autopilot and a variety of other examples in tech-related domains, wherein people respond to gains from technological innovation by simply raising their expectations/lowering their effort levels?
Do you think stuff like this won't happen? Do you think stuff like this already happens (see: drivers on cell phones) and it won't get any worse? Do you think it'll still be a net gain? Etc?
That premise presumes a *fully* autonomous vehicle such that it doesn't matter how distracted the passenger is.
Most automous vehicle advocates say semi automous vehicles are a bad idea for exactly the reason this crash showed. You can't expect a human to react in time if they aren't actively driving.
I think that the intial level of automation we are seeing now (which is actually on hundreds of thousands of vehicles, in the form of adaptive cruise control and lane departure assistance and warnings) will help with many accidents, as many people are already massively distracted, exhausted, or drunk. The adaptive cruise control and warnings are already useful at either just automatically braking or preventing a lane transition if the driver is asleep or rooting around in the glove box for a bagel.
'Fatal' highway accidents are most likely to be those where there is no response from the driver, and they just strike a stopped car (say) at 70 mph. With adaptive cruise control they may be more likely to reach for that bagel, but the car will respond even if they don't and cut the speed by 80% before the collision. So I would say that we might see more incidents by number, but that they would decrease by severity. Incidents by number may also decrease however if more and more people use driver assistance aids, as they will drive at smoother speeds and change lanes less.
For example, while 'distracted driving' is the illegal cause of many accidents, the legal cause (as in, the manover they made was a legal one) of a larger number of these is people deciding to change lanes. Distracted drivers then either smash into the back of them, or see them and swerve. As more people use driver aids, I would expect them to change lanes less. Driver aids also encourage use of proper lane spacing, which makes the driver aids work better (because they need less 'peripheral information') and may also cut down on incidents.
'Next generation' driver aids, such as that present on many luxury vehicles include those that will assist with steering and effectively manage all your freeway driving in normal conditions. These are the sort of driver aids which until recently seemed nearly impossible, and only recently have become practical with modern machine vision. They still have their limitations, and there is indeed a concern that they would be vastly safer for an attentive driver, but perhaps would make an inattentive driver even more prone to distractions. However, it is almost always the case in terms of safety that the majority of issues are caused by the simplest and most common incident type. Which (in this case) would be...
Innatentive driver approaches vehicle ahead. Vehicle ahead is travelling massively slower, but is clearly visible in the road. Driver takes no useful action. High speed collision.
With a next generation driver aid, and a driver who was prone to distraction, then this situation would be rendered far safer. A slow vehicle would be spotted directly ahead in the roadway, the driver assist car would slow down, and there would be no collision.
So I think that while our current best technology may render poor (a large fraction) drivers at greater risk of experiencing 'obscure' collisions (such as the one in the video shown earlier, where the lead vehicle swerves at the last minute to avoid a stationary vehicle ahead of it, and the driver assist vehicle hits it) it would minimize their risk of more simple (but equally lethal and dangerous) and far more common collisions.
Regarding the future, I expect that we will soon reach the end of 'on vehicle' vision as a basic concept. On vehicle vision will remain, but next generation systems will combine their on vehicle vision with imaging and guidance from towers and cameras on bridges to produce safer vehicles. By moving to a remote camera better safety can be achieved even with current systems by getting a better angle on the road.
Regarding the future, I expect that we will soon reach the end of 'on vehicle' vision as a basic concept. On vehicle vision will remain, but next generation systems will combine their on vehicle vision with imaging and guidance from towers and cameras on bridges to produce safer vehicles. By moving to a remote camera better safety can be achieved even with current systems by getting a better angle on the road.
You are definitely expecting a level on infrastructure investment that flat out does not exist. Here's a picture I took of one of the main roads on my way to work:
Regarding the future, I expect that we will soon reach the end of 'on vehicle' vision as a basic concept. On vehicle vision will remain, but next generation systems will combine their on vehicle vision with imaging and guidance from towers and cameras on bridges to produce safer vehicles. By moving to a remote camera better safety can be achieved even with current systems by getting a better angle on the road.
You are definitely expecting a level on infrastructure investment that flat out does not exist. Here's a picture I took of one of the main roads on my way to work:
I'm expecting this infrastructure investment to be made by the car companies themselves. So there will be a 'BMW/Honda/Nissan alliance' or whatever who puts up towers on roadways with cameras and whatnot. Then anyone who buys one of those cars can use their smart vision systems in the areas where they are on. Eventually, the systems will unify (like ATM's did eventually) and everyone can use all the various cameras which exist.
Also, cameras are MASSIVELY cheaper than potholes and they are much more 'immediate' in their gratification to more people. The US' absurd number of potholes is caused by the insanely challenging climate of much of the northern part of the country, and some of the southern parts. Stopping them requires large expenditures of funds while the roads still just have little cracks, or minor damage. Those challenges don't apply to cameras on poles, which also don't even need to be ON the roadway.
People LOVE driver assistance features. Its a massive reason people choose specific cars. People care about them, and will pay money for them every time. They care about them whenever they turn them on, and feel good about them every time it saves them even a little bit of effort. There's a massive market incentive for companies to make this camera network. You COULD get good results eventually with really good on vehicle machine vision, but I think you could actually (if you are say, Toyota and you've got 10 million vehicles to sell in the US) get it done cheaper by putting up cameras and sensors on roads and a few cameras and sensors on each vehicle than by putting a massively complex sensor suite on each vehicle. Remote distributed vision is just so much more effective than local vision.
People LOVE driver assistance features. Its a massive reason people choose specific cars. People care about them, and will pay money for them every time. They care about them whenever they turn them on, and feel good about them every time it saves them even a little bit of effort. There's a massive market incentive for companies to make this camera network. You COULD get good results eventually with really good on vehicle machine vision, but I think you could actually (if you are say, Toyota and you've got 10 million vehicles to sell in the US) get it done cheaper by putting up cameras and sensors on roads and a few cameras and sensors on each vehicle than by putting a massively complex sensor suite on each vehicle. Remote distributed vision is just so much more effective than local vision.
Massive citation needed. I spend a fair amount of time around these systems in development cars, and have a few in my personal car.
Adaptive cruise works fine, for the most part, but it's super wasteful. It's great if you want to goof off with your phone and let the car brake to avoid hitting the car in front, and then race off again when it sees a gap. It's awful if you actually want to match freeway speed smoothly and try to get decent gas mileage.
Forward collision warning I've left on so far, but I'm very close to turning it off completely. I routinely get false alarms from weather conditions and even occasional phantom alerts on a clear day. I get why the system has to be as sensitive as it is - any less sensitive and you wouldn't have sufficient time to stop. That doesn't make it any less obnoxious when it freaks out because of a car ahead that I already knew about and can see turning at an intersection.
Lane departure warning is a pain in the ass. I've tried turning it on a few times and always turn it back off in disgust. In my car, all it does is beep at me. In more "advanced" cars, it actively fights me. Part of my commute is a 2 lane road that's shared with bikers, and I need to be able to move over to avoid them. It's infuriating to have to fight the wheel to move over for them.
None of these systems add anything to your drive if you're actually paying attention to the road. All it does is give the jackoffs playing with their phones a false sense of security.
EDIT: Oh, I have parallel park assist too! Also completely useless, but a fun party trick.
There’s always going to be accidents. I trust computers to make fewer of them than humans.
My new car detects when I start departing the lane and starts tapping me back in to my lane. Which has taught me that I apparently did this all the time. Now I do it far less. More safety features are going to make people safer in aggregate, not less safe.
Pretty much the only thing making people safer at this point is improvements in vehicle technology. Humanity isn’t getting better at handling long stretches of boredom or avoiding distractions any time soon.
Our choices are as follows:
Status Quo
Private entities kill thousands and spend millions in order to create a system that will make the present logistical problems of transit worse.
OR End sprawl, build trains, and tax cars.
Edith Upwards on
0
Options
KetarCome on upstairswe're having a partyRegistered Userregular
People LOVE driver assistance features. Its a massive reason people choose specific cars. People care about them, and will pay money for them every time. They care about them whenever they turn them on, and feel good about them every time it saves them even a little bit of effort. There's a massive market incentive for companies to make this camera network. You COULD get good results eventually with really good on vehicle machine vision, but I think you could actually (if you are say, Toyota and you've got 10 million vehicles to sell in the US) get it done cheaper by putting up cameras and sensors on roads and a few cameras and sensors on each vehicle than by putting a massively complex sensor suite on each vehicle. Remote distributed vision is just so much more effective than local vision.
Massive citation needed. I spend a fair amount of time around these systems in development cars, and have a few in my personal car.
Adaptive cruise works fine, for the most part, but it's super wasteful. It's great if you want to goof off with your phone and let the car brake to avoid hitting the car in front, and then race off again when it sees a gap. It's awful if you actually want to match freeway speed smoothly and try to get decent gas mileage.
Forward collision warning I've left on so far, but I'm very close to turning it off completely. I routinely get false alarms from weather conditions and even occasional phantom alerts on a clear day. I get why the system has to be as sensitive as it is - any less sensitive and you wouldn't have sufficient time to stop. That doesn't make it any less obnoxious when it freaks out because of a car ahead that I already knew about and can see turning at an intersection.
Lane departure warning is a pain in the ass. I've tried turning it on a few times and always turn it back off in disgust. In my car, all it does is beep at me. In more "advanced" cars, it actively fights me. Part of my commute is a 2 lane road that's shared with bikers, and I need to be able to move over to avoid them. It's infuriating to have to fight the wheel to move over for them.
None of these systems add anything to your drive if you're actually paying attention to the road. All it does is give the jackoffs playing with their phones a false sense of security.
EDIT: Oh, I have parallel park assist too! Also completely useless, but a fun party trick.
I like adaptive cruise control quite a bit, and I don't ever touch my phone while driving. It's sometimes useful when I have a lot on my mind or am upset about something and would tend toward driving a bit faster than I actually intended.
Forward collision warning has only been useful maybe twice so far, but both times were when I was distracted by my children screaming in the back about something and I appreciated the alert. The only false alarms I ever get while driving at speed are when the car believes I should be braking harder than I already am. They are fairly frequent when parking, but that doesn't bother me at all.
I've never used the parallel park assist that we have because I've had to parallel park in tight spaces for decades, but it never ceases to surprise me how many people in the suburbs haven't parallel parked a car since driver's ed and would rather keep driving and looking for parking than do it. Those are the drivers that parallel park assist is there for, and there are more of them than you'd want to believe.
Nah, false dichotomy. Increased public transit and better designed roads and neighborhoods is definitely better. Safer, autonomous cars for private transit is also better. One does not preclude the other.
There’s always going to be accidents. I trust computers to make fewer of them than humans.
My new car detects when I start departing the lane and starts tapping me back in to my lane. Which has taught me that I apparently did this all the time. Now I do it far less. More safety features are going to make people safer in aggregate, not less safe.
I mean no offense, but in my mind, this would make you a "bad" driver, and I would want you to have more remedial training than just a machine to tell you when you're drifting out of your lane in the specific situations where it can detect that happening. My dad does something similar, because he doesn't have a good appreciation for how large his vehicle is and he always edges to the right because he sees oncoming traffic on his left but can't "feel" the right-hand side of his car. But he and you are going to hit curves, in particular S-curves, or have multi-lane turns, or whatnot, and I'd be worried that you're also going to drift in those situations in which the computer can't nudge you back on the road.
I grant that such training might never be possible, and I guess you've mostly been driving solo so you haven't had any passengers to point out that you drift a lot, so in those cases the computer at least providing you that feedback is good. It's just... I'd want more, ideally.
People LOVE driver assistance features. Its a massive reason people choose specific cars. People care about them, and will pay money for them every time. They care about them whenever they turn them on, and feel good about them every time it saves them even a little bit of effort. There's a massive market incentive for companies to make this camera network. You COULD get good results eventually with really good on vehicle machine vision, but I think you could actually (if you are say, Toyota and you've got 10 million vehicles to sell in the US) get it done cheaper by putting up cameras and sensors on roads and a few cameras and sensors on each vehicle than by putting a massively complex sensor suite on each vehicle. Remote distributed vision is just so much more effective than local vision.
Massive citation needed. I spend a fair amount of time around these systems in development cars, and have a few in my personal car.
Adaptive cruise works fine, for the most part, but it's super wasteful. It's great if you want to goof off with your phone and let the car brake to avoid hitting the car in front, and then race off again when it sees a gap. It's awful if you actually want to match freeway speed smoothly and try to get decent gas mileage.
Forward collision warning I've left on so far, but I'm very close to turning it off completely. I routinely get false alarms from weather conditions and even occasional phantom alerts on a clear day. I get why the system has to be as sensitive as it is - any less sensitive and you wouldn't have sufficient time to stop. That doesn't make it any less obnoxious when it freaks out because of a car ahead that I already knew about and can see turning at an intersection.
Lane departure warning is a pain in the ass. I've tried turning it on a few times and always turn it back off in disgust. In my car, all it does is beep at me. In more "advanced" cars, it actively fights me. Part of my commute is a 2 lane road that's shared with bikers, and I need to be able to move over to avoid them. It's infuriating to have to fight the wheel to move over for them.
None of these systems add anything to your drive if you're actually paying attention to the road. All it does is give the jackoffs playing with their phones a false sense of security.
EDIT: Oh, I have parallel park assist too! Also completely useless, but a fun party trick.
The issue with my using your judgment on this, is that you have already established that you are a 'trained' driver who is likely passionate about driving. Drivers who believe themselves to be high skill are those who reject assistance systems and view them as obtrusive or degrading to their driving. I would imagine you also sometimes complain about power steering, or about automatic transmission gear selection times.
Both of my adaptive cruise control systems brake and accelerate smoothly, and maintain an excellent following distance. I would prefer if braking occurred earlier (attempt to achieve never coming to a complete stop), and acceleration was slower, but the manufacturers have to try and cater to aggressive drivers. Lane departure alerts are informative and clear. When I use them both my vehicles achieve a more uniform speed and better gas mileage. I find my commute CONSIDERABLY less exhausting when I have them on. If you drive in heavy traffic with a 'full stop capable' system, then the effect is trans-formative. I would no more buy a vehicle without such a system than I would buy a manual vehicle, or one where I had to manually prime the engine. I would not even take one if it were offered to me for free if I actually had to drive it to work. Hell, I think if you offered me a brand new vehicle without it, for free, and said you would pay me $5000 to drive it to work for 12 months, I would say no.
Sales information for most vehicles reveal that those which have them are the more popular models despite their increased cost. Companies continue to introduce them on more and more vehicles to help sales.
Regarding parallel park assist. My wife and mother in law literally CANNOT park without it. Before they had it, if a space required them to parallel park either I would have to take over driving and do it, or we would drive around until we found a space.
"That is cool" - Abraham Lincoln
+2
Options
KetarCome on upstairswe're having a partyRegistered Userregular
edited June 2018
It's almost one year to the day since I got rear-ended by a teenager looking down at her phone instead of watching the road. Would forward collision warning have gotten her to look up soon enough to not hit us at all? I don't know. But it probably would have allowed her to slow down enough that I wouldn't have ended up with a herniation of a disc in my neck and two months of physical therapy and $8,000 in medical bills that took a while to get her insurance to cover. Not to mention not traumatizing the shit out of my kids.
It's almost one year to the day since I got rear-ended by a teenager looking down at her phone instead of watching the road. Would forward collision warning have gotten her to look up soon enough to not hit us at all? I don't know. But it probably would have allowed her to slow down enough that I wouldn't have ended up with a herniation of a disc in my neck and two months of physical therapy and $8,000 in medical bills that took a while to get her insurance to cover. Not to mention not traumatizing the shit out of my kids.
Honestly, she should get the fuck off the road. 10-year driving ban, minimum. Hurgle burgle old man ranting about making technology that enables shitty youth behaviour. (This is basically where I'm at now. I get that this is coming and happening, but it feels like papering over real problems with drivers being terribad and getting terriworse, and possibly even enabling them getting even more terriworse as time goes on, while I yell impotently at the sky about it. Personal responsibility blah dee blah!)
People LOVE driver assistance features. Its a massive reason people choose specific cars. People care about them, and will pay money for them every time. They care about them whenever they turn them on, and feel good about them every time it saves them even a little bit of effort. There's a massive market incentive for companies to make this camera network. You COULD get good results eventually with really good on vehicle machine vision, but I think you could actually (if you are say, Toyota and you've got 10 million vehicles to sell in the US) get it done cheaper by putting up cameras and sensors on roads and a few cameras and sensors on each vehicle than by putting a massively complex sensor suite on each vehicle. Remote distributed vision is just so much more effective than local vision.
Massive citation needed. I spend a fair amount of time around these systems in development cars, and have a few in my personal car.
Adaptive cruise works fine, for the most part, but it's super wasteful. It's great if you want to goof off with your phone and let the car brake to avoid hitting the car in front, and then race off again when it sees a gap. It's awful if you actually want to match freeway speed smoothly and try to get decent gas mileage.
Forward collision warning I've left on so far, but I'm very close to turning it off completely. I routinely get false alarms from weather conditions and even occasional phantom alerts on a clear day. I get why the system has to be as sensitive as it is - any less sensitive and you wouldn't have sufficient time to stop. That doesn't make it any less obnoxious when it freaks out because of a car ahead that I already knew about and can see turning at an intersection.
Lane departure warning is a pain in the ass. I've tried turning it on a few times and always turn it back off in disgust. In my car, all it does is beep at me. In more "advanced" cars, it actively fights me. Part of my commute is a 2 lane road that's shared with bikers, and I need to be able to move over to avoid them. It's infuriating to have to fight the wheel to move over for them.
None of these systems add anything to your drive if you're actually paying attention to the road. All it does is give the jackoffs playing with their phones a false sense of security.
EDIT: Oh, I have parallel park assist too! Also completely useless, but a fun party trick.
There is also a bigger picture I think you are missing. I dream of adaptive cruise control not just for highway use, but all the time if I can. I deliberately choose routes that allow for long stretches or steady speeds because it helps me when driving. The reptitive motion of braking and accelerating combine with trying to hold my foot in the exact right position with little support can be incredibly painful. I basically can't drive in stop and go traffic regularly unless I want to have a lot of trouble walking.
Driver assist isn't just for people who are bad at driving, but also for people who just struggle with certain aspects. The more we can accommodate people who are capable of driving but have issues the better. I think we do need better training for drivers for sure, but don't throw the baby out with the bathwater.
People LOVE driver assistance features. Its a massive reason people choose specific cars. People care about them, and will pay money for them every time. They care about them whenever they turn them on, and feel good about them every time it saves them even a little bit of effort. There's a massive market incentive for companies to make this camera network. You COULD get good results eventually with really good on vehicle machine vision, but I think you could actually (if you are say, Toyota and you've got 10 million vehicles to sell in the US) get it done cheaper by putting up cameras and sensors on roads and a few cameras and sensors on each vehicle than by putting a massively complex sensor suite on each vehicle. Remote distributed vision is just so much more effective than local vision.
Massive citation needed. I spend a fair amount of time around these systems in development cars, and have a few in my personal car.
Adaptive cruise works fine, for the most part, but it's super wasteful. It's great if you want to goof off with your phone and let the car brake to avoid hitting the car in front, and then race off again when it sees a gap. It's awful if you actually want to match freeway speed smoothly and try to get decent gas mileage.
Forward collision warning I've left on so far, but I'm very close to turning it off completely. I routinely get false alarms from weather conditions and even occasional phantom alerts on a clear day. I get why the system has to be as sensitive as it is - any less sensitive and you wouldn't have sufficient time to stop. That doesn't make it any less obnoxious when it freaks out because of a car ahead that I already knew about and can see turning at an intersection.
Lane departure warning is a pain in the ass. I've tried turning it on a few times and always turn it back off in disgust. In my car, all it does is beep at me. In more "advanced" cars, it actively fights me. Part of my commute is a 2 lane road that's shared with bikers, and I need to be able to move over to avoid them. It's infuriating to have to fight the wheel to move over for them.
None of these systems add anything to your drive if you're actually paying attention to the road. All it does is give the jackoffs playing with their phones a false sense of security.
EDIT: Oh, I have parallel park assist too! Also completely useless, but a fun party trick.
The issue with my using your judgment on this, is that you have already established that you are a 'trained' driver who is likely passionate about driving. Drivers who believe themselves to be high skill are those who reject assistance systems and view them as obtrusive or degrading to their driving. I would imagine you also sometimes complain about power steering, or about automatic transmission gear selection times.
Both of my adaptive cruise control systems brake and accelerate smoothly, and maintain an excellent following distance. I would prefer if braking occurred earlier (attempt to achieve never coming to a complete stop), and acceleration was slower, but the manufacturers have to try and cater to aggressive drivers. Lane departure alerts are informative and clear. When I use them both my vehicles achieve a more uniform speed and better gas mileage. I find my commute CONSIDERABLY less exhausting when I have them on. If you drive in heavy traffic with a 'full stop capable' system, then the effect is trans-formative. I would no more buy a vehicle without such a system than I would buy a manual vehicle, or one where I had to manually prime the engine. I would not even take one if it were offered to me for free if I actually had to drive it to work. Hell, I think if you offered me a brand new vehicle without it, for free, and said you would pay me $5000 to drive it to work for 12 months, I would say no.
Sales information for most vehicles reveal that those which have them are the more popular models despite their increased cost. Companies continue to introduce them on more and more vehicles to help sales.
Regarding parallel park assist. My wife and mother in law literally CANNOT park without it. Before they had it, if a space required them to parallel park either I would have to take over driving and do it, or we would drive around until we found a space.
TLDR:
blah blah blah
"I am clearly wrong about this as, I am about all things down to my very core. I am a miserable human being incapable of feeling joy."
blah blah blah
People LOVE driver assistance features. Its a massive reason people choose specific cars. People care about them, and will pay money for them every time. They care about them whenever they turn them on, and feel good about them every time it saves them even a little bit of effort. There's a massive market incentive for companies to make this camera network. You COULD get good results eventually with really good on vehicle machine vision, but I think you could actually (if you are say, Toyota and you've got 10 million vehicles to sell in the US) get it done cheaper by putting up cameras and sensors on roads and a few cameras and sensors on each vehicle than by putting a massively complex sensor suite on each vehicle. Remote distributed vision is just so much more effective than local vision.
Massive citation needed. I spend a fair amount of time around these systems in development cars, and have a few in my personal car.
Adaptive cruise works fine, for the most part, but it's super wasteful. It's great if you want to goof off with your phone and let the car brake to avoid hitting the car in front, and then race off again when it sees a gap. It's awful if you actually want to match freeway speed smoothly and try to get decent gas mileage.
Forward collision warning I've left on so far, but I'm very close to turning it off completely. I routinely get false alarms from weather conditions and even occasional phantom alerts on a clear day. I get why the system has to be as sensitive as it is - any less sensitive and you wouldn't have sufficient time to stop. That doesn't make it any less obnoxious when it freaks out because of a car ahead that I already knew about and can see turning at an intersection.
Lane departure warning is a pain in the ass. I've tried turning it on a few times and always turn it back off in disgust. In my car, all it does is beep at me. In more "advanced" cars, it actively fights me. Part of my commute is a 2 lane road that's shared with bikers, and I need to be able to move over to avoid them. It's infuriating to have to fight the wheel to move over for them.
None of these systems add anything to your drive if you're actually paying attention to the road. All it does is give the jackoffs playing with their phones a false sense of security.
EDIT: Oh, I have parallel park assist too! Also completely useless, but a fun party trick.
The issue with my using your judgment on this, is that you have already established that you are a 'trained' driver who is likely passionate about driving. Drivers who believe themselves to be high skill are those who reject assistance systems and view them as obtrusive or degrading to their driving. I would imagine you also sometimes complain about power steering, or about automatic transmission gear selection times.
Both of my adaptive cruise control systems brake and accelerate smoothly, and maintain an excellent following distance. I would prefer if braking occurred earlier (attempt to achieve never coming to a complete stop), and acceleration was slower, but the manufacturers have to try and cater to aggressive drivers. Lane departure alerts are informative and clear. When I use them both my vehicles achieve a more uniform speed and better gas mileage. I find my commute CONSIDERABLY less exhausting when I have them on. If you drive in heavy traffic with a 'full stop capable' system, then the effect is trans-formative. I would no more buy a vehicle without such a system than I would buy a manual vehicle, or one where I had to manually prime the engine. I would not even take one if it were offered to me for free if I actually had to drive it to work. Hell, I think if you offered me a brand new vehicle without it, for free, and said you would pay me $5000 to drive it to work for 12 months, I would say no.
Sales information for most vehicles reveal that those which have them are the more popular models despite their increased cost. Companies continue to introduce them on more and more vehicles to help sales.
Regarding parallel park assist. My wife and mother in law literally CANNOT park without it. Before they had it, if a space required them to parallel park either I would have to take over driving and do it, or we would drive around until we found a space.
Okay, that's a somewhat fair assessment, though I'm actually in favor of a lot of driver assistance systems. Power steering is great. Antilock brakes are great. I love stability control - I turn it off when I want to play in a controlled environment, and turn it right back on when I'm back on public roads. Blind zone monitoring is a fantastic feature. But it's true that these systems all exist to give me better control as a driver; I'm not a fan of systems that take it away.
It occurs to me that's I've been overwhelmingly negative in this thread and that's not my intention. @tbloxham, I actually agree with you that autonomous driving will be a net good. It's true that most humans are shit drivers and we will have safer and more efficient roadways if they weren't driving. @Gnizmo also raised a very good point about making transportation more accessible. I think where we disagree is what the systems are capable of and should be allowed to do today, in the gray area between manual driving and fully autonomous driving. I see a lot of pie-in-the-sky ideas or claims in this thread and I want to bring them back down to reality, because we are still a ways off from where these systems can be trusted today.
It seems like there's definitely a risk where vehicles that are almost autonomous will lead to drivers not paying attention and at least potentially causing more accidents than they would have in "manual" vehicles.
+14
Options
zepherinRussian warship, go fuck yourselfRegistered Userregular
It seems like there's definitely a risk where vehicles that are almost autonomous will lead to drivers not paying attention and at least potentially causing more accidents than they would have in "manual" vehicles.
It's already at that point, I had a super who was in a car with lane departure warning and adaptive cruise control, and just set it and didn't touch the wheel while he drove through South Carolina, and that made me a bit concerned.
Autonomous vehicles will also drive increased enthusiasm for cars. Which, is not good. The United States does not need cars, we need an adequate rail network capable of transporting passengers and cargo to large portions of the country under conditions of increasing fuel and energy scarcity.
Autonomous vehicles will also drive increased enthusiasm for cars. Which, is not good. The United States does not need cars, we need an adequate rail network capable of transporting passengers and cargo to large portions of the country under conditions of increasing fuel and energy scarcity.
Nope, we need both improved public transit AND improved vehicles. A robot driving a car vs a human driving a car could probably achieve 20-30% better gas mileage.
Autonomous vehicles will also drive increased enthusiasm for cars. Which, is not good. The United States does not need cars, we need an adequate rail network capable of transporting passengers and cargo to large portions of the country under conditions of increasing fuel and energy scarcity.
Baring some massive redesign of our cities then yes we pretty much do need cars unfortunately. I really don't see the will/desire or investment needed to undo suburban sprawl enough to be able to ditch cars any time soon.
Humans are actually pretty darn good at driving over all.
Laugh if you want, but its true.
At least in America (and i know UK drivers are statistically better) while on your way home consider the sheer AMOUNT of drivers on the road at any time, the lack of any communication between them, and the simple road signage we use, its absolutely STAGGERING that there is so few accidents relatively speaking.
Especially considering the immense vehicles we drive.
This isnt to say, of course, that there isn't monumental stupid drivers sometimes. But for every 2 or 3 i see and get exasperated at, there is usually -at any given time- HUNDREDS of other drivers around me doing well enough that i never give it a passing thought.
Now, we can ALWAYS be better. But trying to claim that humans cant drive is, on the whole, very incorrect.
Humans are actually pretty darn good at driving over all.
Laugh if you want, but its true.
At least in America (and i know UK drivers are statistically better) while on your way home consider the sheer AMOUNT of drivers on the road at any time, the lack of any communication between them, and the simple road signage we use, its absolutely STAGGERING that there is so few accidents relatively speaking.
Especially considering the immense vehicles we drive.
This isnt to say, of course, that there isn't monumental stupid drivers sometimes. But for every 2 or 3 i see and get exasperated at, there is usually -at any given time- HUNDREDS of other drivers around me doing well enough that i never give it a passing thought.
Now, we can ALWAYS be better. But trying to claim that humans cant drive is, on the whole, very incorrect.
Pretty darn good is actually pretty darned awful, considering that traffic flow and road safety are overwhelmingly governed by the behavior of the worst human drivers.
Humans suck at driving, but we deal with it by agreeing on rules on how to operate, but what happen when someone, or something, deviates from the script?
Main reason autonomic car will, eventually, be better than a human, is that they can effectively keep track of the surrounding area and react in a timely fashion to deviations to the script we all operate with while on the road (and they'll reduce the amount of times when something does go off script).
Sometimes I watch those compilation videos of road rage, brake checks, etc.. To me those are one of the best arguments for more automation. Maybe people will stop getting so ridiculously pissed off about perceived slights when they think it could just be a software glitch.
+2
Options
JuliusCaptain of Serenityon my shipRegistered Userregular
So far there’s been way more progress in making cars better than making drivers better.
in the US.
edit: like, I get the general argument for the superiority of computers in doing driving, but there seems to be this assumption here that the massive problems in drivers ed and road safety are somehow the fault of the people themselves. people can't drive well so it is clearly our flawed capabilities instead of shit education.
(i understand that utopian self driving cars in the future seems more realistic than true institutional change in the US, but it really isn't.)
So far there’s been way more progress in making cars better than making drivers better.
in the US.
edit: like, I get the general argument for the superiority of computers in doing driving, but there seems to be this assumption here that the massive problems in drivers ed and road safety are somehow the fault of the people themselves. people can't drive well so it is clearly our flawed capabilities instead of shit education.
(i understand that utopian self driving cars in the future seems more realistic than true institutional change in the US, but it really isn't.)
So far there’s been way more progress in making cars better than making drivers better.
in the US.
edit: like, I get the general argument for the superiority of computers in doing driving, but there seems to be this assumption here that the massive problems in drivers ed and road safety are somehow the fault of the people themselves. people can't drive well so it is clearly our flawed capabilities instead of shit education.
(i understand that utopian self driving cars in the future seems more realistic than true institutional change in the US, but it really isn't.)
Agreed. I've watched Canada's Worst Driver and been on the road in Poland and people are shit drivers there too. But there are roundabouts and lower speed limits and various other elements of road design that push safety.
While racing light mechs, your Urbanmech comes in second place, but only because it ran out of ammo.
So far there’s been way more progress in making cars better than making drivers better.
in the US.
edit: like, I get the general argument for the superiority of computers in doing driving, but there seems to be this assumption here that the massive problems in drivers ed and road safety are somehow the fault of the people themselves. people can't drive well so it is clearly our flawed capabilities instead of shit education.
(i understand that utopian self driving cars in the future seems more realistic than true institutional change in the US, but it really isn't.)
How long do you think it would take to undo a century of whatever cultural shortcomings led to people being bad at driving? Even if the government somehow managed to decide that this is an education problem and devotes the budget to fixing it (good luck), overhauling the driver's ed curriculum alone would take decades to bear fruit.
0
Options
JuliusCaptain of Serenityon my shipRegistered Userregular
So far there’s been way more progress in making cars better than making drivers better.
in the US.
edit: like, I get the general argument for the superiority of computers in doing driving, but there seems to be this assumption here that the massive problems in drivers ed and road safety are somehow the fault of the people themselves. people can't drive well so it is clearly our flawed capabilities instead of shit education.
(i understand that utopian self driving cars in the future seems more realistic than true institutional change in the US, but it really isn't.)
is that why I always hear about terrible road behavior that I very rarely see here? people get their vehicles inspected and they suddenly start driving better?
couldn't possibly be that the education is flawed. even though that is an accepted argument for literally every other problem in the US.
Sometimes I watch those compilation videos of road rage, brake checks, etc.. To me those are one of the best arguments for more automation. Maybe people will stop getting so ridiculously pissed off about perceived slights when they think it could just be a software glitch.
My computer not doing what I want or expect it to causes me to rage harder than literally anything else in my life. I don't think you understand what rage is when you sum it up as "...people will stop getting so ridiculously pissed off about perceived slights..." Rage isn't a logical reaction, and it isn't something that you consciously control.
Posts
To a degree, yes, but there is a limit to how good a driver human can be, on average.
There will always be exceptionally good, and exceptionally bad, drivers, but average human will remain as just that.
Yes but ALL drivers become better with time, an individual driver will change their abilities as they gain more experience, changing from being a teen driver, to a young adult, and so on. This transition happens to everyone and doesn't affect the behavior of the driving population.
Places with higher road safety have better roads and better maintained vehicles (the UK), not better trained drivers. People don't use the techniques we DO teach during driver training, let alone the ones taught during advanced training.
Here's a good article on the subject. Advanced driver training makes people WORSE drivers
http://driving.ca/auto-news/news/why-advanced-driver-training-makes-teens-worse-drivers
https://practicalmotoring.com.au/car-advice/why-advanced-driving-courses-dont-always-make-you-a-safer-driver/
The evidence is that the only way to make people safer drivers is not to teach them how to drive in difficult conditions, or to teach them special techniques for safety. But to make them more concerned and aware of those difficult conditions and make them NOT drive in those situations. Peripheral road awareness (which is pretty much all the stuff you are talking about with looking at shadows and brakelights and considering upcoming curves) can't be trained, its a thing your teenagers don't have and that driving on the road will give them. For example, when you talk about looking at shadows and through windscreens and whatnot to see upcoming cars, you aren't actually doing that. It's just your brain being trained by 1000s of hours of driving experience noticing some weird 'flock' behavior in the cars ahead. You see, without really knowing why, that the cars ahead are odd and you start slowing down to avoid the oddness.
I mean, drivers can be trained a bit. The transition from 0 hours experience to 40 hours experience is something which can absolutely be helped by training. There might be something which could be done in early driving life to make sure people who are slow to gain this experience don't come onto the roads (teens, espescially boys shouldn't drive alone till 18, and driving before that should be done with their parents as part of supervised 'experience aquisition'. You could, I suppose make people retest every 12 months like aviation pilots, that might be enough to make a difference. But that test should focus overwhelmingly on safe and cautious driving techniques.
Slower
Fewer lane changes
Larger spacing
None of these things require any 'skill' just the attitude that every crash which happens near you is partially your fault.
I think I am feeling skepticism because I drive a sedan in Texas, where nearly eveyone else on the road is driving a truck. I certainly try to look through the car ahead of me's window, but the opportunities to do so are rare.
Also, any time leave a safe distance between me and the car ahead, the traffic around me fills it. Fucking Dallas traffic.
"Evidence shows that in the USA the highest skilled drivers (registered race and rally car drivers) have a much higher crash rate than the average driver (Naatanen and Summala, 1976)."
And so on.
All the studies show that the only thing people can do to lower accident risks is to drive carefully and slowly.
Yeah, I was driving home yesterday and at least 50% of the vehicles in the road are pretty much completely vision blocking. Ever tried to look through the rear window of a Prius?
I wonder, for those of you who say that human beings are terrible drivers and they're never going to get any better, so we need autonomous vehicles to make the roads safer, how do you incorporate within that thesis the fact that the above incident and others are going to happen, that the safer we make drivers feel, the more leeway they're going to take, essentially creating an equilibrium dynamic with a resting point between "safety" and "freedom"? It seems likely to me that some sort of psychological reversal is going to occur, as we see with speed limits and with autopilot and a variety of other examples in tech-related domains, wherein people respond to gains from technological innovation by simply raising their expectations/lowering their effort levels?
Do you think stuff like this won't happen? Do you think stuff like this already happens (see: drivers on cell phones) and it won't get any worse? Do you think it'll still be a net gain? Etc?
I think that the intial level of automation we are seeing now (which is actually on hundreds of thousands of vehicles, in the form of adaptive cruise control and lane departure assistance and warnings) will help with many accidents, as many people are already massively distracted, exhausted, or drunk. The adaptive cruise control and warnings are already useful at either just automatically braking or preventing a lane transition if the driver is asleep or rooting around in the glove box for a bagel.
'Fatal' highway accidents are most likely to be those where there is no response from the driver, and they just strike a stopped car (say) at 70 mph. With adaptive cruise control they may be more likely to reach for that bagel, but the car will respond even if they don't and cut the speed by 80% before the collision. So I would say that we might see more incidents by number, but that they would decrease by severity. Incidents by number may also decrease however if more and more people use driver assistance aids, as they will drive at smoother speeds and change lanes less.
For example, while 'distracted driving' is the illegal cause of many accidents, the legal cause (as in, the manover they made was a legal one) of a larger number of these is people deciding to change lanes. Distracted drivers then either smash into the back of them, or see them and swerve. As more people use driver aids, I would expect them to change lanes less. Driver aids also encourage use of proper lane spacing, which makes the driver aids work better (because they need less 'peripheral information') and may also cut down on incidents.
'Next generation' driver aids, such as that present on many luxury vehicles include those that will assist with steering and effectively manage all your freeway driving in normal conditions. These are the sort of driver aids which until recently seemed nearly impossible, and only recently have become practical with modern machine vision. They still have their limitations, and there is indeed a concern that they would be vastly safer for an attentive driver, but perhaps would make an inattentive driver even more prone to distractions. However, it is almost always the case in terms of safety that the majority of issues are caused by the simplest and most common incident type. Which (in this case) would be...
Innatentive driver approaches vehicle ahead. Vehicle ahead is travelling massively slower, but is clearly visible in the road. Driver takes no useful action. High speed collision.
With a next generation driver aid, and a driver who was prone to distraction, then this situation would be rendered far safer. A slow vehicle would be spotted directly ahead in the roadway, the driver assist car would slow down, and there would be no collision.
So I think that while our current best technology may render poor (a large fraction) drivers at greater risk of experiencing 'obscure' collisions (such as the one in the video shown earlier, where the lead vehicle swerves at the last minute to avoid a stationary vehicle ahead of it, and the driver assist vehicle hits it) it would minimize their risk of more simple (but equally lethal and dangerous) and far more common collisions.
Regarding the future, I expect that we will soon reach the end of 'on vehicle' vision as a basic concept. On vehicle vision will remain, but next generation systems will combine their on vehicle vision with imaging and guidance from towers and cameras on bridges to produce safer vehicles. By moving to a remote camera better safety can be achieved even with current systems by getting a better angle on the road.
You are definitely expecting a level on infrastructure investment that flat out does not exist. Here's a picture I took of one of the main roads on my way to work:
You can't give someone a pirate ship in one game, and then take it back in the next game. It's rude.
I'm expecting this infrastructure investment to be made by the car companies themselves. So there will be a 'BMW/Honda/Nissan alliance' or whatever who puts up towers on roadways with cameras and whatnot. Then anyone who buys one of those cars can use their smart vision systems in the areas where they are on. Eventually, the systems will unify (like ATM's did eventually) and everyone can use all the various cameras which exist.
Also, cameras are MASSIVELY cheaper than potholes and they are much more 'immediate' in their gratification to more people. The US' absurd number of potholes is caused by the insanely challenging climate of much of the northern part of the country, and some of the southern parts. Stopping them requires large expenditures of funds while the roads still just have little cracks, or minor damage. Those challenges don't apply to cameras on poles, which also don't even need to be ON the roadway.
People LOVE driver assistance features. Its a massive reason people choose specific cars. People care about them, and will pay money for them every time. They care about them whenever they turn them on, and feel good about them every time it saves them even a little bit of effort. There's a massive market incentive for companies to make this camera network. You COULD get good results eventually with really good on vehicle machine vision, but I think you could actually (if you are say, Toyota and you've got 10 million vehicles to sell in the US) get it done cheaper by putting up cameras and sensors on roads and a few cameras and sensors on each vehicle than by putting a massively complex sensor suite on each vehicle. Remote distributed vision is just so much more effective than local vision.
Massive citation needed. I spend a fair amount of time around these systems in development cars, and have a few in my personal car.
Adaptive cruise works fine, for the most part, but it's super wasteful. It's great if you want to goof off with your phone and let the car brake to avoid hitting the car in front, and then race off again when it sees a gap. It's awful if you actually want to match freeway speed smoothly and try to get decent gas mileage.
Forward collision warning I've left on so far, but I'm very close to turning it off completely. I routinely get false alarms from weather conditions and even occasional phantom alerts on a clear day. I get why the system has to be as sensitive as it is - any less sensitive and you wouldn't have sufficient time to stop. That doesn't make it any less obnoxious when it freaks out because of a car ahead that I already knew about and can see turning at an intersection.
Lane departure warning is a pain in the ass. I've tried turning it on a few times and always turn it back off in disgust. In my car, all it does is beep at me. In more "advanced" cars, it actively fights me. Part of my commute is a 2 lane road that's shared with bikers, and I need to be able to move over to avoid them. It's infuriating to have to fight the wheel to move over for them.
None of these systems add anything to your drive if you're actually paying attention to the road. All it does is give the jackoffs playing with their phones a false sense of security.
EDIT: Oh, I have parallel park assist too! Also completely useless, but a fun party trick.
You can't give someone a pirate ship in one game, and then take it back in the next game. It's rude.
My new car detects when I start departing the lane and starts tapping me back in to my lane. Which has taught me that I apparently did this all the time. Now I do it far less. More safety features are going to make people safer in aggregate, not less safe.
Our choices are as follows:
Status Quo
I like adaptive cruise control quite a bit, and I don't ever touch my phone while driving. It's sometimes useful when I have a lot on my mind or am upset about something and would tend toward driving a bit faster than I actually intended.
Forward collision warning has only been useful maybe twice so far, but both times were when I was distracted by my children screaming in the back about something and I appreciated the alert. The only false alarms I ever get while driving at speed are when the car believes I should be braking harder than I already am. They are fairly frequent when parking, but that doesn't bother me at all.
I've never used the parallel park assist that we have because I've had to parallel park in tight spaces for decades, but it never ceases to surprise me how many people in the suburbs haven't parallel parked a car since driver's ed and would rather keep driving and looking for parking than do it. Those are the drivers that parallel park assist is there for, and there are more of them than you'd want to believe.
I mean no offense, but in my mind, this would make you a "bad" driver, and I would want you to have more remedial training than just a machine to tell you when you're drifting out of your lane in the specific situations where it can detect that happening. My dad does something similar, because he doesn't have a good appreciation for how large his vehicle is and he always edges to the right because he sees oncoming traffic on his left but can't "feel" the right-hand side of his car. But he and you are going to hit curves, in particular S-curves, or have multi-lane turns, or whatnot, and I'd be worried that you're also going to drift in those situations in which the computer can't nudge you back on the road.
I grant that such training might never be possible, and I guess you've mostly been driving solo so you haven't had any passengers to point out that you drift a lot, so in those cases the computer at least providing you that feedback is good. It's just... I'd want more, ideally.
The issue with my using your judgment on this, is that you have already established that you are a 'trained' driver who is likely passionate about driving. Drivers who believe themselves to be high skill are those who reject assistance systems and view them as obtrusive or degrading to their driving. I would imagine you also sometimes complain about power steering, or about automatic transmission gear selection times.
Both of my adaptive cruise control systems brake and accelerate smoothly, and maintain an excellent following distance. I would prefer if braking occurred earlier (attempt to achieve never coming to a complete stop), and acceleration was slower, but the manufacturers have to try and cater to aggressive drivers. Lane departure alerts are informative and clear. When I use them both my vehicles achieve a more uniform speed and better gas mileage. I find my commute CONSIDERABLY less exhausting when I have them on. If you drive in heavy traffic with a 'full stop capable' system, then the effect is trans-formative. I would no more buy a vehicle without such a system than I would buy a manual vehicle, or one where I had to manually prime the engine. I would not even take one if it were offered to me for free if I actually had to drive it to work. Hell, I think if you offered me a brand new vehicle without it, for free, and said you would pay me $5000 to drive it to work for 12 months, I would say no.
Sales information for most vehicles reveal that those which have them are the more popular models despite their increased cost. Companies continue to introduce them on more and more vehicles to help sales.
Regarding parallel park assist. My wife and mother in law literally CANNOT park without it. Before they had it, if a space required them to parallel park either I would have to take over driving and do it, or we would drive around until we found a space.
Honestly, she should get the fuck off the road. 10-year driving ban, minimum. Hurgle burgle old man ranting about making technology that enables shitty youth behaviour. (This is basically where I'm at now. I get that this is coming and happening, but it feels like papering over real problems with drivers being terribad and getting terriworse, and possibly even enabling them getting even more terriworse as time goes on, while I yell impotently at the sky about it. Personal responsibility blah dee blah!)
There is also a bigger picture I think you are missing. I dream of adaptive cruise control not just for highway use, but all the time if I can. I deliberately choose routes that allow for long stretches or steady speeds because it helps me when driving. The reptitive motion of braking and accelerating combine with trying to hold my foot in the exact right position with little support can be incredibly painful. I basically can't drive in stop and go traffic regularly unless I want to have a lot of trouble walking.
Driver assist isn't just for people who are bad at driving, but also for people who just struggle with certain aspects. The more we can accommodate people who are capable of driving but have issues the better. I think we do need better training for drivers for sure, but don't throw the baby out with the bathwater.
TLDR:
blah blah blah
"I am clearly wrong about this as, I am about all things down to my very core. I am a miserable human being incapable of feeling joy."
blah blah blah
Okay, that's a somewhat fair assessment, though I'm actually in favor of a lot of driver assistance systems. Power steering is great. Antilock brakes are great. I love stability control - I turn it off when I want to play in a controlled environment, and turn it right back on when I'm back on public roads. Blind zone monitoring is a fantastic feature. But it's true that these systems all exist to give me better control as a driver; I'm not a fan of systems that take it away.
It occurs to me that's I've been overwhelmingly negative in this thread and that's not my intention. @tbloxham, I actually agree with you that autonomous driving will be a net good. It's true that most humans are shit drivers and we will have safer and more efficient roadways if they weren't driving. @Gnizmo also raised a very good point about making transportation more accessible. I think where we disagree is what the systems are capable of and should be allowed to do today, in the gray area between manual driving and fully autonomous driving. I see a lot of pie-in-the-sky ideas or claims in this thread and I want to bring them back down to reality, because we are still a ways off from where these systems can be trusted today.
You can't give someone a pirate ship in one game, and then take it back in the next game. It's rude.
Nope, we need both improved public transit AND improved vehicles. A robot driving a car vs a human driving a car could probably achieve 20-30% better gas mileage.
Baring some massive redesign of our cities then yes we pretty much do need cars unfortunately. I really don't see the will/desire or investment needed to undo suburban sprawl enough to be able to ditch cars any time soon.
Laugh if you want, but its true.
At least in America (and i know UK drivers are statistically better) while on your way home consider the sheer AMOUNT of drivers on the road at any time, the lack of any communication between them, and the simple road signage we use, its absolutely STAGGERING that there is so few accidents relatively speaking.
Especially considering the immense vehicles we drive.
This isnt to say, of course, that there isn't monumental stupid drivers sometimes. But for every 2 or 3 i see and get exasperated at, there is usually -at any given time- HUNDREDS of other drivers around me doing well enough that i never give it a passing thought.
Now, we can ALWAYS be better. But trying to claim that humans cant drive is, on the whole, very incorrect.
Pretty darn good is actually pretty darned awful, considering that traffic flow and road safety are overwhelmingly governed by the behavior of the worst human drivers.
Main reason autonomic car will, eventually, be better than a human, is that they can effectively keep track of the surrounding area and react in a timely fashion to deviations to the script we all operate with while on the road (and they'll reduce the amount of times when something does go off script).
in the US.
edit: like, I get the general argument for the superiority of computers in doing driving, but there seems to be this assumption here that the massive problems in drivers ed and road safety are somehow the fault of the people themselves. people can't drive well so it is clearly our flawed capabilities instead of shit education.
(i understand that utopian self driving cars in the future seems more realistic than true institutional change in the US, but it really isn't.)
Countries with better road safety have...
Shorter commutes
Strict vehicle inspection standards
Shorter workdays (Less driving when it's dark)
Better road layouts
Driver training isn't a meaningful part.
Agreed. I've watched Canada's Worst Driver and been on the road in Poland and people are shit drivers there too. But there are roundabouts and lower speed limits and various other elements of road design that push safety.
How long do you think it would take to undo a century of whatever cultural shortcomings led to people being bad at driving? Even if the government somehow managed to decide that this is an education problem and devotes the budget to fixing it (good luck), overhauling the driver's ed curriculum alone would take decades to bear fruit.
is that why I always hear about terrible road behavior that I very rarely see here? people get their vehicles inspected and they suddenly start driving better?
couldn't possibly be that the education is flawed. even though that is an accepted argument for literally every other problem in the US.
(also shorter workdays??)
My computer not doing what I want or expect it to causes me to rage harder than literally anything else in my life. I don't think you understand what rage is when you sum it up as "...people will stop getting so ridiculously pissed off about perceived slights..." Rage isn't a logical reaction, and it isn't something that you consciously control.