MIT Study Shows Stop Lights Won't Be Necessary In The Future (computerworld.com)

Lucas123 writes: An MIT [Senseable City Lab] study based on mathematical modeling demonstrated a likely scenario in which high-tech vehicles, using sensors to remain at a safe distance from each other as they move through a four-way intersection, can eliminate the need for traffic lights in the future. By removing the waits caused by traffic lights, these so-called Slot-based Intersections speed-up traffic flow.The study claims this kind of traffic-light-free transportation design, if it ever arrives, could allow twice as much traffic to use existing roads.

Comments

If something does go wrong

[ickleberry]

And statistically, it will since autonomous cars are not mythical crash-proof transport machines then you're going to end up with an almighty bang

Re:If something does go wrong

[dgatwood]

Humans suck at driving. The problem is that 99.999% of the time, you can be borderline incompetent at driving and you'll still get there safely, because things only go wrong on rare occasions. Most of the time, at city street speeds, you could glance at the road for two seconds out of every ten, and you wouldn't crash, because there just isn't much happening. There are situations, however, in which humans are physically incapable of being good drivers. For example:

• Fatigue/falling asleep at the wheel (computers don't get tired)
• Distractions (both driver-initiated and external, e.g. rubbernecking at a wreck site; computers don't get distracted)
• Sudden, unexpected traffic stops in front of you (human reaction time is a large fraction of a second, versus microseconds)
• Backing out of a parking place into traffic (limited human vision versus ability to simultaneously monitor cameras pointed in every direction)
• Intoxication
• Cardiac arrest/seizure/narcolepsy/other medical issues or events

And in some cases, each of those situations can result in a crash with a human driver, depending mostly on luck. Computers, by contrast, won't exhibit any of those physical failings, and thus won't crash in any of those situations, typically.

So the key question is whether they will crash more often in other situations where a human wouldn't (e.g. when nothing is going wrong). As long as that answer is no, then they will likely be safer than human drivers.

Regardless, automation can kinda sorta mostly run trains and other vehicles on fixed guideways in a pristine environment most of the time without failing. That's about the state of things and it's not changing all that fast despite the constant droning from the "futurists" who have been wrong about absolutely everything ever.

That's just not true at all. Google's self-driving cars have clocked over a million miles on the roads, with basically no at-fault crashes. That's a far cry from barely being able to work in a pristine environment on a fixed guideway. It has some ability to recognize pedestrian behavior, avoid obstacles in the road, handle traffic lights (as long as it knows to look for them at a particular intersection), etc. It does require a lot of pre-mapping of the terrain so that it knows where to watch for traffic lights, roughly where lanes are, etc., but still, they've gone way beyond a subway system on a fixed track as you imply.

Re:If something does go wrong

[pixelpusher220]

The issue right now is basically what you explained for why human drivers suck. Automated cars are GREAT at the 99%. It's the edge cases, in bad weather, with mechanical failures...that will be the true test.

Has google been testing their cars in blizzards? Mostly they seem to be in fairly nice safe weather conditions. This link details Ford's ventures into this issue and it seems like good progress. But when the car simply loses traction on a snow covered street, who does it try to protect? The driver? or the pedestrian walking in the road because the sidewalk isn't plowed? How well does this system cope with a build up of ice across a vehicle? It's easy to cherry pick these questions obviously, but giving full access requires either the ability to deal with everything as well or better than humans. (which might be possible in most cases)

Re:If something does go wrong

[Dunbal]

So far autonomous cars have only been crashed into by people.

Re:If something does go wrong

[ShanghaiBill]

Automated cars are GREAT at the 99%. It's the edge cases, in bad weather, with mechanical failures...that will be the true test.

Automated cars are better than humans in bad weather. They have many types of sensors, while humans rely on vision. Tesla has recommended that owners engage Autopilot during snowstorms, rather than trying to drive. The sensors can detect other cars in fog or heavy rain, avoiding fatal pileups. As for mechanical failures, those cause a very small percentage of crashes, but I see no reason to assume that humans can handle them better than a computer. At the very least, the computer would have a faster reaction time.

Re:If something does go wrong

[witherstaff]

A google car hit a bus recently. The car's fault as it had to move to another lane to get around an obstacle http://www.latimes.com/local/l...

Re:If something does go wrong

[Anonymous Coward]

Looks like it actually was maneuvering inside a quite wide lane as the bus squeezes past...https://www.youtube.com/watch?v=I9T6LkNm-5w Not USian so not sure how you legally define lanes etc in such a situation though

Re:If something does go wrong

[khallow]

Humans suck at driving.

Actually, they don't. That's why driving hasn't already been replaced, unlike say, computing a FFT by hand or screwing on a million bottle caps. For example, in the US we're down to about 11 deaths and 1850 crashes per billion vehicle miles.

Re: If something does go wrong

[pixelpusher220]

As my link indicates, the cars actually can't do these things...and are programmed to simply pull over and stop. That isn't reasonable for an entire roadway. Sensors can get blocked by snow and road spray and now the car is making decisions on faulty inputs. If the lane markings are covered by snow a camera isn't going to help stay in the lane it can't see.

Re: If something does go wrong

[hackwrench]

I'm having trouble expressing what I want here. As far as I can tell, the answer used as to what to do when a pedestrian is involved is along the lines of protecting nether or to do things that involve giving a pedestrian a chance.to react as long as they do something reasonable and don't act like they went to the Prometheus school of large moving object avoidance.

Re:If something does go wrong

[naughtynaughty]

The article it came from points out that the dramatic decline in deaths per 100 million miles is almost entirely due to safer vehicles. It isn't that drivers don't suck, it's that we've gotten a lot better at protecting drivers when they do get in an accident.

Automated driving is more difficult than programming a computer to perform an FFT or designing a machine to screw on bottle caps. The reason it hasn't already occurred is because the technology required hasn't been available at a reasonable cost in the past.

Re: If something does go wrong

[pixelpusher220]

the trouble expressing what you want is very understandable. Precisely because it will change depending on the specific circumstance encountered.

Lets take to the more extreme - you're on an icy road with a cliff on one side and walking along the road is a class of elementary school kids. Now your car loses control - Should your car drive you off the cliff to save the kids? or try to save you but possibly push the kids off the cliff or just run over them...

If the car will choose the kids over the driver...who in their right mind would ever BUY a car designed to kill you?

There is no right answer in all likelihood...and this is the type of societal values/questions/ramifications thing automation is going to need to handle when it starts becoming significant in its integration into the real world.

Another concept - when emergency stopping tech is standard in cars (2022 by recent agreement) - and in 2030 many cars still on the road won't have....is a new driver who's never known a car without it liable if they drive an older car without it and expect the car to save them?

Re: If something does go wrong

[ShanghaiBill]

As my link indicates, the cars actually can't do these things...

No, your link indicates that Ford cars can't actually do these things. Plenty of other cars, including tens of thousands of Teslas, can and are doing them. My wife owns a Tesla, and it auto-drives just fine in rain, fog, and snow. Ford was a late starter in automated driving, so I am not surprised that they are behind others.

Re: If something does go wrong

[pixelpusher220]

Not talking about Tesla, this was about Google and fully autonomous cars. Tesla doesn't have this yet.

Most emergency braking systems are radar driven...which is rendered useless with a coating of ice or snow over the sensors. Visual rules (for lane following) break down when the entire roadway is white or when the camera is blocked...

Physics is a harsh mistress.

Re:If something does go wrong

[dgatwood]

Actually, they don't. That's why driving hasn't already been replaced, unlike say, computing a FFT by hand or screwing on a million bottle caps. For example, in the US we're down to about 11 deaths [caranddriver.com] and 1850 crashes per billion vehicle miles.

Just because humans suck at something, that doesn't necessarily imply that computers are any better at it. Most humans suck at singing. Computers don't do it very well, either.

We've only just recently gotten to the point where computers might have a prayer of matching human drivers. If somebody had tried to create self-driving cars in 1978, they'd have been using an 8086 at .33 MIPS. Now, even a Raspberry Pi will give you 4,744 MIPS, and an eight-core i7 will net you up to 238,310 MIPS. So computers are almost a million times faster than they were in 1978. That makes a lot of things possible that were laughable even ten years ago.

Re:If something does go wrong

[delt0r]

How the fuck are humans better with mechanical failures or bad weather? Some magic pixie human brain shit? Some magical midiclorians or something. Cus a machine is going to see better than a human in *all* weather, adjust quicker than any human to changing conditions and more accurately detect these conditions.

Humans are not magic, why do you think it will be harder for a machine to these things right when they are going to be the easiest things to do better than humans. ALL humans as well, not the just the average human.

Re:If something does go wrong

[pixelpusher220]

Humans are better with mechanical failures because we know a failure has happened. An automatic car might not would continue applying the gas.

Computers definitely react faster than humans. The modern traction control systems and anti-lock brakes attest to that daily. It's an entirely different ballgame when we're talking about deducing the entire environment and planning ahead.

Computers are only as good as their sensors and they can be blocked very easily by road debris or ice and snow. Radar doesn't work through obstructions like ice build up or a randomly placed bit of road tar.

Recognizing a human being in the world is a non-trivial computer problem...one they are starting to work out in 'good' conditions. We don't drive only in 'good' conditions.

Re: If something does go wrong

[Cyberax]

All modern autopilot systems monitor car's response to control input and adjust accordingly. Just like humans do. The current crop of self-driving cars probably doesn't have enough smarts to deal with black ice and tire blowouts, but there's absolutely no reason future cars can't be adapted to do that.

Re:If something does go wrong

[naughtynaughty]

A more relevant statistic:
U.S. drivers crash approximately 5.6 million times per year

Re:If something does go wrong

[lsatenstein]

Humans suck at driving. The problem is that 99.999% of the time, you can be borderline incompetent at driving and you'll still get there safely, because things only go wrong on rare occasions. Most of the time, at city street speeds, you could glance at the road for two seconds out of every ten, and you wouldn't crash, because there just isn't much happening. There are situations, however, in which humans are physically incapable of being good drivers. For example:

• Fatigue/falling asleep at the wheel (computers don't get tired)
• Distractions (both driver-initiated and external, e.g. rubbernecking at a wreck site; computers don't get distracted)
• Sudden, unexpected traffic stops in front of you (human reaction time is a large fraction of a second, versus microseconds)
• Backing out of a parking place into traffic (limited human vision versus ability to simultaneously monitor cameras pointed in every direction)
• Intoxication
• Cardiac arrest/seizure/narcolepsy/other medical issues or events

And in some cases, each of those situations can result in a crash with a human driver, depending mostly on luck. Computers, by contrast, won't exhibit any of those physical failings, and thus won't crash in any of those situations, typically.

So the key question is whether they will crash more often in other situations where a human wouldn't (e.g. when nothing is going wrong). As long as that answer is no, then they will likely be safer than human drivers.

Regardless, automation can kinda sorta mostly run trains and other vehicles on fixed guideways in a pristine environment most of the time without failing. That's about the state of things and it's not changing all that fast despite the constant droning from the "futurists" who have been wrong about absolutely everything ever.

That's just not true at all. Google's self-driving cars have clocked over a million miles on the roads, with basically no at-fault crashes. That's a far cry from barely being able to work in a pristine environment on a fixed guideway. It has some ability to recognize pedestrian behavior, avoid obstacles in the road, handle traffic lights (as long as it knows to look for them at a particular intersection), etc. It does require a lot of pre-mapping of the terrain so that it knows where to watch for traffic lights, roughly where lanes are, etc., but still, they've gone way beyond a subway system on a fixed track as you imply.

In Quebec Canada, when we have a winter storm, the sensors will get covered with ice. After some driving, the road salt dissolves and replaces the ice. The sensor lenses will be completely layered with roadsalt. And of course, we have a more slippery road surface. Slippery from ice/snow, but after that is dissolved, slippery from the road salt that has become a powdered layer.

Humans will be needed, particularly if following a taxi driver. No, automation will require more human interventions.

Re:If something does go wrong

[khallow]

What's more relevant about that statistic? What does that tell us about how competent US drivers are?

Re:If something does go wrong

[blackanvil]

A big problem with the perception of cars unable to deal with snow/ice is that a lot of people leave all their traction control gizmos on even in slick conditions, and the car just won't move. This is because at least elements of traction control systems need to be turned off in some situations, something an auto-driving car is well aware of and presumably just bypasses when its driving in such conditions. I've seen cars abandoned in the middle of a 4-lane highway on a gentle slope because they didn't know that all they had to do was hit the button to turn off that feature.

Re: If something does go wrong

[pixelpusher220]

"presumably". That's the rub. They don't until proven to do so. Then once proven now move go edge cases and proven they always react correctly....which may be not doing anything. Google cars apparently pull over and stop. Not a bad choice but when a roadway worth does this its not a realistic solution.

Re: If something does go wrong

[delt0r]

And humans do? I have seen humans "deal with black ice", it is called crashing.

Re:If something does go wrong

[toddestan]

If it was so easy to make a traction control system that would automatically disengage when needed, then why don't the cars we have today already have such system to do just that instead of relying on the driver to push a button? I'm thinking the problem is a bit harder than you make it out to be.

Re:If something does go wrong

[blackanvil]

They do, Landrover and some other off-road systems do exactly that -- switch from one mode to another until they find one that works. Costly, though, and often invisible to the end user, as it should be.

Re:If something does go wrong

[khallow]

Just because humans suck at something, that doesn't necessarily imply that computers are any better at it.

The low accident rate per billion miles tells us otherwise. There are many people out there who drive a vast amount and have very few accidents.

Most humans suck at singing.

So what? We don't care whether most humans can sing or not. We do care if most humans can drive or not.

The problem here is two-fold. First, we have a very low accident rate in the US. There is not that much room for improvement. If you look at the proposed benefits of self-driving vehicles, you find that they speak of highway driving a lot. That is already among the safest sort of driving.

Further, both cargo and taxi driving are high priority targets. It would replace the safest human drivers, the ones who do it as a profession.

There could be a period of time when automated driving makes things worse due to replacing the safest human driving and human rather than the worst.

Roundabouts?

[Anonymous Coward]

What would be the comparative advantage with respect to a roundabout?

I know they are not very popular in the US, but they can be very efficient, and prevent the frustration of waiting at busy intersection (especially if going in the non-popular direction).

Re:Roundabouts?

[Intron]

They are definitely better, especially when more than 2 roads come into an intersection.

Re:Roundabouts?

[Sarten-X]

Space requirements. A high-traffic roundabout requires more land area than a cross intersection.

Re:Roundabouts?

[nukenerd]

Space requirements. A high-traffic roundabout requires more land area than a cross intersection.

I thought that the US had loads of space. Roundabouts have been back-fitted all over the place in the UK, including "mini-roundabouts" (which I don't believe achieve much), Having done that though, the trend in the UK is to install traffic lights at every intersection of the roundabout as well. There is a large roundabout near me (just off the M5 at Avonmouth) that has I reckon about fifty traffic light heads; I must count them one day.

Re:Roundabouts?

[Culture20]

Well sure! Let's just take some extra space from Montana and plunk it into an intersection in Manhattan. Now it fits a roundabout!

Re:Roundabouts?

[Intron]

https://en.wikipedia.org/wiki/...

Re:Roundabouts?

[Ichijo]

Space requirements. A high-traffic roundabout requires more land area than a cross intersection.

That's partially correct, sometimes:

A roundabout may need more property within the actual intersection, but often take up less space on the streets approaching the roundabout. Because roundabouts can handle greater volumes of traffic more efficiently than signals, where drivers may need to line up to wait for a green light, roundabouts usually require fewer lanes approaching the intersection.

Re:Roundabouts?

[gweihir]

They are efficient, pretty safe (no blowing through red lights, at worst low-speed collisions), have a high traffic capacity and are far cheaper to maintain. Sure, drivers tend to dislike them for a while, but that passes.

Re:Roundabouts?

[beelsebob]

Not true. What's true is that given an intersection with 2 lane roads coming into it, a roundabout will need more space than a light controlled intersection.

However, a roundabout causes the roads into it and out of it to be used more efficiently, as traffic flows down them constantly, rather than only coming in bursts roughly 1/4 of the time. The result is that said 2 lane light controlled junction gets replaced by one lane roads, and a roundabout that takes up the same space as the junction.

Overall result - a space saving, not a space cost.

Re:Roundabouts?

[beelsebob]

While you're right - roundabouts are much better than light controlled junctions, the thing I don't understand is why the US insists on using 4 way stops for small junctions, rather than just having a major/minor road configuration like in the UK. A 4 way stop results in everyone having to stop all the time, rather than allowing the 90% case (the more major road having cars on it) to work efficiently. In practice, because the more major road flows more smoothly, the minor roads get cars out quickly too.

Re:Roundabouts?

[beelsebob]

Right, we call those "mini roundabouts" in the UK. They work well in cases where there's not really one road that's more major than the other. Typically trucks just drive over them.

As I said - a major/minor junction is the default for an intersection in the UK - they always work too, and rather better than a 4 way stop.

Re:Roundabouts?

[dgatwood]

Must have been an American driver. :-)

With that said, I'm glad I went around the one at the Arc de Triomphe at 2 in the morning instead of during rush hour. When roundabouts even start to approach eight lanes, they get terrifying pretty quickly. They're awesome for two-lane roads, though, and serviceable for four, depending largely on traffic patterns.

Re:Roundabouts?

[dgatwood]

While you're right - roundabouts are much better than light controlled junctions, the thing I don't understand is why the US insists on using 4 way stops for small junctions, rather than just having a major/minor road configuration like in the UK. A 4 way stop results in everyone having to stop all the time, rather than allowing the 90% case (the more major road having cars on it) to work efficiently. In practice, because the more major road flows more smoothly, the minor roads get cars out quickly too.

Statistically, most U.S. road intersections with stop signs actually are two-way stops rather than four. Typically, every fourth or fifth intersection is a four-way stop to ensure that people who are afraid to pull out into traffic have at least one path that guarantees them the ability to get on the road. :-)

The more annoying thing to me is the overuse of traffic lights. There are many, many intersections near me where I wait for an average of a minute or more per trip and watch two cars pass. Now granted, a few of those intersections are busy enough at certain times of day to warrant a traffic light. Those lights should ideally go into a two-way-stop (flashing red), two-way-yield (flashing yellow) configuration except during rush hour.

Re:Roundabouts?

[Drethon]

Timed traffic lights are better as people don't have to even slow down. Of course that is often somewhat of a pipe dream...

Re:Roundabouts?

[beelsebob]

Statistically, most U.S. road intersections with stop signs actually are two-way stops rather than four. Typically, every fourth or fifth intersection is a four-way stop to ensure that people who are afraid to pull out into traffic have at least one path that guarantees them the ability to get on the road. :-)

Interesting, I moved to CA about 3 years ago, and I'm not sure I've seen a 2 way stop since coming. If it were like you said, here, it would be entirely sane.

The more annoying thing to me is the overuse of traffic lights. There are many, many intersections near me where I wait for an average of a minute or more per trip and watch two cars pass. Now granted, a few of those intersections are busy enough at certain times of day to warrant a traffic light. Those lights should ideally go into a two-way-stop (flashing red), two-way-yield (flashing yellow) configuration except during rush hour.

Or, just make it a passively controlled junction - i.e. a roundabout.

Re:Roundabouts?

[bugs2squash]

No, the roundabout need only be a dot in the middle of the intersection. It is only there to set rules as to who has to yield to whom.

Re:Roundabouts?

[dryeo]

Of course most people have to slow down. Timed traffic lights are the perfect example of using a carrot to enforce the speed limit. Speed and all your lights are red. Go the speed limit and all your lights are green.
Won't work in corrupt jurisdictions as they depend on the stick to enforce speed limits and raise revenue with safety an excuse.
Seems that they used to be common around here, drive for miles in town with no red lights. Now I think they time it for traffic management/calming as every light seems to be red, which just leads to frustration rather then safety.

Re:Roundabouts?

[nukenerd]

Traffic lights on a roundabout? Sort of defeats the purpose

Yep, in the UK. Try the Google Street View of Avonmouth, coming off the M5, the short spur North West to a traffic-light roundabout. There are eight signal heads for that intersection alone.

I often pass here late at night when there is no other traffic, yet the lights turn red in my face all the way round. I believe it is deliberate, it is supposed to "calm" you; actually it is infuriating and when I see them changing to red I floor it and jump them sometimes, the visibility is very clear anyway. I didn't say that.

Re:Roundabouts?

[Layzej]

Rome has had lightless intersections for a while now. No need for high tech driverless cars. Works great: https://www.youtube.com/watch?...

Re:Roundabouts?

[Rockoon]

If the visibility is good and nobody else is on the road, stop at and then ignore the red light.

Re:Roundabouts?

[johnlcallaway]

So you are willing to donate your paycheck to get all of those stop lights updated then??? If someone has to pay for it, I'd rather the people whose lives are so important they can't stand the thought of waiting a minute or two at a light pay for it.

Re:Roundabouts?

[turbidostato]

"If the visibility is good and nobody else is on the road, stop at and then ignore the red light."

The visibility being good means the cameras also have no problem to grab your license plate (nor a better thing to do, since your are the only car in the roundabout) so no, better not to ignore the traffic lights.

Re: Roundabouts?

[turbidostato]

"IMHO problem is not only the timid drivers but also the aggressive drivers inside de roundabout."

Nailed.

While not difficult to grasp, roundabouts are quite demanding on the politeness (meaning in this case, "please follow the rules to the dot") of its users. Driving and politeness have come basically antonyms, so go figure.

Re:Roundabouts?

[arth1]

No, the roundabout need only be a dot in the middle of the intersection. It is only there to set rules as to who has to yield to whom.

It's often preferable that it's just a dot in the middle, because then you can see the full traffic picture. The typical big traffic circles with trees and coyote dens in the middle block the view.

In Europe, many inner city roundabouts are no more than the plinth that used to have a policeman standing on it, directing traffic before traffic lights came about. Now, instead of the policeman, there's a sign marking it as a roundabout. This works reasonably well.
Getting the policeman back would likely work even better.

Re:Roundabouts?

[russotto]

Roundabouts are only popular among people who don't like cars. They slow traffic to a crawl at all times because you have to look in several different directions at once to see if it is safe to enter and to figure out where you have to be to exit.

Re:Roundabouts?

[gweihir]

No, they are not. If done very carefully, they may get somewhat near, but that is it.

Re:Roundabouts?

[beelsebob]

Yes and no - mini roundabouts are great and all, but roundabouts do become more efficient when more than 2 cars can fit on them at once.

Re:Roundabouts?

[wwalker]

Didn't Mythbusters test roundabouts vs 4-way stops? I don't remember the exact throughput advantage of a roundabout, I think it depended on the amount of traffic and was around 50-100% increase. But roundabouts also require quite a bit more real estate, and are more expensive. They didn't test it vs a traffic light though.

Re:Roundabouts?

[mikael]

https://goo.gl/maps/gATLigGo3y...

https://goo.gl/maps/CMgjYCjVru...

Re:Roundabouts?

[dgatwood]

Interesting, I moved to CA about 3 years ago, and I'm not sure I've seen a 2 way stop since coming. If it were like you said, here, it would be entirely sane.

It might depend on the particular city. I can list a bunch of streets with 2-way stop in the South Bay. :-)

Or, just make it a passively controlled junction - i.e. a roundabout.

Space permitting, sure, except that many of those lights are there to guarantee that cars can actually get out onto roads when otherwise cars would be backed up through the intersection from the next red light.... :-)

Re:Roundabouts?

[dgatwood]

First, it isn't one light programmed stupidly. It is darn near every light around here. I'd estimate that I waste at least half an hour per day waiting at traffic lights, and that the vast majority of that time was wasted waiting for nobody. And yes, you bet your ass I'd pay to get them fixed. With that said, I'm already paying for those lights to be maintained, in the form of property tax, gasoline tax, and who knows what else. And so are fifty or sixty thousand other people in my city.

The most egregious lights I've seen cause an extra two or three minutes of delay for about 3-4,000 residents on a daily basis thanks to poor programming (unnecessary red left arrow when there are only two or three cars per hour coming in the opposite direction, not counting the people who drive straight, turn around in a parking lot, and come back because of the stupid red left arrow). Add that up, and that's 8,000 minutes of time per day, or more than 5 1/2 years per year.

Assuming an average of about one liter per 30 minutes of idling, that's an estimated 26,000 gallons of wasted gasoline every year just from one minor city street light. Fixing that one light would be like taking 66 cars off the road permanently. If even 5% of the 300,000+ traffic lights in this country have similar efficiency problems, then fixing them would be like taking two million cars off the roads, or almost a 10% reduction in automotive fuel consumption.

It's not just about my time. It's about the time spent by every single driver times the number of poorly programmed lights. It adds up to a huge economic cost, a huge environmental cost, and a huge time sink, on top of being annoying. I really think that this issue deserves more attention than it gets.

Re:Roundabouts?

[mjwx]

What would be the comparative advantage with respect to a roundabout?

I know they are not very popular in the US, but they can be very efficient, and prevent the frustration of waiting at busy intersection (especially if going in the non-popular direction).

Roundabouts are excellent with light to medium traffic, but when there is heavy traffic things tend to get backed up as one pair of entrances start to monopolise the roundabout. You really need traffic lights then.

What MIT are talking about are replacing traffic lights with computers in cars, which in reality will never work as you've got to ensure that every single car has the same technology and that no humans are taking control (or have modified the software). The last assumption is a stupid one as humans are selfish and if they can be faster by doing something dodgy, they will.

Re:Roundabouts?

[beelsebob]

Space permitting, sure,

Thankfully, roundabouts take up the same amount of space as traffic light controlled junctions, and allow you to shrink the inbound/outbound roads due to their greater efficiency.

[quote]except that many of those lights are there to guarantee that cars can actually get out onto roads when otherwise cars would be backed up through the intersection from the next red light.... :-)[/quote]
That's fine - roundabouts allow for this too. As soon as one person turns right or left, rather than going straight on, it opens up a small gap for someone on a side street to pop out. That in itself then causes a small hole for someone else to get out. It *is* possible to end up with a poor configuration for roundabouts where one entrance gets starved, but at that point you can put lights on the roundabout which alternate red/green in a circle. It's not as efficient as a plain roundabout, but it's much more efficient than a light controlled junction still.

Re:Roundabouts?

[Etcetera]

Interesting, I moved to CA about 3 years ago, and I'm not sure I've seen a 2 way stop since coming. If it were like you said, here, it would be entirely sane.

2-way and 3-way intersections are common in suburban areas of CA. Typically where there's a large discrepancy in size/traffic between the two roads, but not so large that the side street can never enter (in which case, it'd usually get updated to a traffic light).

The more annoying thing to me is the overuse of traffic lights. There are many, many intersections near me where I wait for an average of a minute or more per trip and watch two cars pass. Now granted, a few of those intersections are busy enough at certain times of day to warrant a traffic light. Those lights should ideally go into a two-way-stop (flashing red), two-way-yield (flashing yellow) configuration except during rush hour.

Or, just make it a passively controlled junction - i.e. a roundabout.

Within California, traffic light deployment varies dramatically by county. In SD County, there are a fair number of traffic lights, but outside of the downtown core, most of them are sensor-based. They default to the primary thoroughfare and will trigger for a side street when the induction coil sees a car (or a pedestrian button is pressed). There are virtually NONE that degenerate down to flashing red/yellow during off-hours, although I've seen those during long road trips around the west (and they were very confusing at first).

Additionally, most traffic lights here in San Diego have protected left turns (also sensor controlled), whereas even going up to LA they're much less common.

Pedestrians

[Anonymous Coward]

How are they supposed to cross? Without lights, there could just be continuous stream of them walking over the road and cars can't pass since they try to avoid hitting them thus causing even bigger jams in big city centers.

Re:Pedestrians

[aix tom]

Well, they just have to invent the high-tech autonomous pedestrian then to replace all the old models.

Re:Pedestrians

[Shadow99_1]

This was my question as well. Though here in the US everyone regularly forgets pedestrians. I've even lived in a city (Columbus OH), which often omits sidewalks in highly congested areas. So I know how badly they are already treated. This would just make it lots worse.

Re:Pedestrians

[gweihir]

Bridges and tunnels. This is not rocket-science.

Re:Pedestrians

[jeti]

Those are expensive to build and require additional space.

Re:Pedestrians

[jeti]

Also don't forget bicycles, motorbikes and tractors.

Re:Pedestrians

[shess]

Every intersection will have an autonomous vehicle to shuttle pedestrians across.

Or perhaps the police will just arrest them.

Re:Pedestrians

[istartedi]

Just step off the curb, and the cars will stop. To keep busy intersections from slowing to a crawl, you'd still have to have walk/don't walk signs. Impromptu street protests where nobody obeys the signs would automatically shut down all traffic in the area because of this. Jay-walking might become a more significant offense, especially if it was done as part of a mass un-permitted march.

Re:Pedestrians

[Solandri]

I've even lived in a city (Columbus OH), which often omits sidewalks in highly congested areas.

In other countries I've been in, if the city needs to expand its roads, it just seizes the land adjacent to the road and expands it.

In the U.S. the government is required to compensate landowners for the market value of the seized land. So if traffic was highly congested as you say, the city probably made an accounting decision that it was congested enough to eliminate the sidewalks for an extra lane, but not congested enough to use Eminent Domain to take the surrounding property and pay off the landowners to keep the sidewalks..

Re:Pedestrians

[cascadingstylesheet]

How are they supposed to cross? Without lights, there could just be continuous stream of them walking over the road and cars can't pass since they try to avoid hitting them thus causing even bigger jams in big city centers.

Well duh, we'll just have to have self-driving pedestrians too. Luddite!

Re:Pedestrians

[YukariHirai]

Yeah, there's a realistic and practical option. Rather than the current situation of hitting a button and waiting a minute or two for the lights when I need to quickly go somewhere just over the road, I can order an Uber, wait a while for it to show up, do a lap of the block in it because the traffic management system says that's the way we need to do it, and pay for the privilege. And repeat for the trip back.

Re:Pedestrians

[vtcodger]

Tomorrows computer controlled pedestrians will fit is seamlessly. (They better if they plan to survive).

Seriously. Eventually, fewer and less complex stop lights but still quite a few to handle pedestrians, synchronizing activity where bumper to bumper traffic streams cross and other special cases.

Oh, and, BTW, expect some monumental traffic jams when those Over The Air updates that folks regard as a solution to some problem or other cause every Honda in North America to decide not to talk to any GM car at 3:00 pm on a Friday afternoon..

Re:Pedestrians

[phantomfive]

building a pedestrian overpass is probably cheaper than building new roads

Re:Pedestrians

[kheldan]

Also don't forget bicycles, motorbikes and tractors.

THIS.
Some people seem to be in such a rush to get goddamned self-driving cars that they're completely ignoring people who don't want them, won't use them, or don't drive at all. You can't just outlaw bicycles, motorcycles, pedestrians or whatever! The Real World is what's going to kill self-driving cars. There are too many variables, and you can't just arbitrarily exclude things!

Re:Pedestrians

[michelcolman]

Automated Segways!

Re:Pedestrians

[Nyh]

Self driving cars will be a lot safer for bicycles, pedestrians and so on. The software in the car is always vigilant and will actually see them and stop for them. Their awareness of other road user might be their biggest problem. People just will take priority of self driving cars because they know the car will stop in time to prevent an accident.

Nyh

Re:Pedestrians

[MMC Monster]

Maybe with the cars seeing them and making a wide path around them?

I remember an Asimov story (one of the more forgetable ones) in which there was a room with robots all going in straight lines criss-crossing the room. The robots were in direct communication with each other so none of them ever hit another. When a person had to cross the room, he essentially had to close his eyes and keep walking, regardless of what he heard happening around him.

Not sure I would trust this in the real world, mind you. A lot more jerks would just stand in front of the traffic and taunt the cars.

Re:Pedestrians

[mikael]

It's like that on the A27 on the South of England. There are a couple of express buses X4/X5 that go between Gosport and Southampton. Naturally, they used opposite sides of the road to travel. Which leads to the challenging problem of getting the bus to go in the opposite direction when you go to or from anywhere. Traffic lights have been removed to improve traffic flow. So it's like playing a game of Frogger trying to get to the opposite side without being squished.

Re:Pedestrians

[mikael]

Can't they put in mini-elevators? Norway has that for some overpasses. They even have special underground lift systems for skateboards and wheelchairs to automatically move them up a hill.

Re:Pedestrians

[Shadow99_1]

The place in question had retail businesses on both sides of the road. Another nearby location had retail and housing complexes along it. Other areas exist as well, including sections that have nothing more than trees (but just happened to be underdeveloped within the city limits) with warehouses back behind them. The city didn't need to do anything to have sidewalks installed in most cases.

Where I live now requires the land owner to build a sidewalk as part of the building and zoning codes. Even residential houses are required to keep their sidewalks in 'good repair'. That said the city decided to pay to rebuild some of the sidewalks to make them easier to access for the disabled and to 'beautify' them in high traffic areas. So cities have a lot of options when it comes to developing transport for pedestrians.

Re:Pedestrians

[gweihir]

And where did you miss the part that roundabouts are a lot cheaper and require much less space in the roads leading up to them? If you selectively ignore data, you will always be able to justify any point-of-view, no matter how flawed.

Re:Pedestrians

[gweihir]

You forget that these are not buildings. No "access" here, and hence wrong act.

Re:Pedestrians

[kheldan]

because they know the car will stop in time to prevent an accident

If you actually do that you're nuts.

I find it strange that too many of you on Slashdot don't appear to be engineers or anyone who has been a maker-of-things, which must be the case because too many of you seem to think that nothing could ever possibly go wrong which is so far from reality that I'd almost have to laugh. Maybe in 50 years there will be some sort of 'self-driving' car that you can trust, but not any sooner than that, and there won't EVER be anything with no manual controls. Ever! Human lives matter. You don't just put people in a box that a human can't control directly and hope for the best! That's about as blind and dystopian as you can get! Don't you understand? A human being must be the final fail-safe backup system to any automated system when there is human safety at stake. It just can't be any other way, and you're nuts if you believe otherwise. Seriously, would you really put your children in some car driven by a computer and send them off to grandma's 100 miles away? No, you wouldn't.

Re:Pedestrians

[david_thornley]

Yeah, but it's still unfriendly to say "You're disabled, so you can't cross the street".

Pedestrian Traffic

[Anonymous Coward]

Did they also model pedestrian traffic crossing the road in their automated "hoverboards" of tomorrow? Vehicular traffic modeling is all well and good for non-pedestrian intersections, but not for dense urban areas with high volumes of pedestrian traffic.

Re:Pedestrian Traffic

[Dutch Gun]

This is the future. Pedestrians go in tubes, duh.

What about pedestrians?

[cmeans]

Seems like they still need to play a part in any scenario with or without traffic lights. Dropping traffic lights may be fine for autonomous vehicles, but that also means that people will just walk when they want to, which'll make cars stop to let them pass. Perfectly fine when there's few pedestrians, but when there's a lot, the automobile traffic will totally come to a half if no one is stopping people from walking into the street.

Re: What about pedestrians?

[phizi0n]

There are also tunnels which are more aesthetically pleasing above ground but afaik they are even more expensive and have the added downside of being dangerous since they are out of sight (ie. you walk down into one and someone robs/hurts you).

During the transition to driverless cars I think updating the existing push buttons at crosswalks to communicate pedestrians' desire to cross to the car network along with having the cars identify pedestrians appearing to be about to cross at intersections without lights will be needed, but eventually pedestrian crossing could be completely eliminated since anyone could just hop in a driverless car and go a couple blocks away instead of walking and if you wanted to go for a walk/run then you would take a driverless car to a nearby park or gym.

Re:What about pedestrians?

[Livius]

Not to mention bicycles.

And it's unlikely that every motorized vehicle can be replaced by an autonomous version. Most, perhaps, but there's bound to be exceptions.

Re: What about pedestrians?

[AvitarX]

Most neighborhoods aren't littered with stoplights now, so I'm not convinced theyd need to be littered with brushes.

Re:What about pedestrians?

[Barlo_Mung_42]

It would work just like today. Pedestrians and bicycles press a button to let the system know they're there. Everyone else stops for a bit to let them cross.

Re: What about pedestrians?

[dAzED1]

that doesn't make sense. People can just hop in a non-driverless-car not and go a couple blocks away. It also ignores big cities like manhattan, where most people don't own a car - driverless or no.

Re:What about pedestrians?

[dAzED1]

Ok - pedestrians, bicycles, animals, and...the backwards compatibility issue. Sure, that bit could eventually work itself out, but even if you mandated all new cars to be driverless by 2017, you'll still have a majority of the cars of the road for years to come be older than that and not have such tech. So you're just going to say people can no longer ride motorcycles or drive old cars? Really think that will fly? Not any time soon. So unless they're talking about some distant future where there are alternate crossing spaces for non-cars, and old cars have been phased out in such a way that everything is already driverless and we're just finally killing off stoplights some decades after the fact, then sure...but won't we have flying cars by then anyway? :)

Re: What about pedestrians?

[breakermelvin]

People will just walk..... I have seen the future and it's hilarious. . Car riders will be hacking their software to get priority....

Re: What about pedestrians?

[phizi0n]

If it doesn't make sense to you then it might be that you're not considering that uber/lyft are already preparing to have fleets of driverless cars so the majority of cars would be either be run by businesses or possibly even government owned. It would be incredibly cheap and quick to pay with NFC to travel short distances and the streets would be full of driverless taxis to take you anywhere. With the current driver operated cars nobody really considers them to go a few blocks even if they're tired or walking or w/e because the wait to get a ride far outweighs the small convenience of it but if nearly every car is a driverless taxi then people would use them to go near or far.

Re:What about pedestrians?

[StormReaver]

Build pedestrian crossings over traffic, like we do now in many locations. Then no one has to stop for anyone.

Re:What about pedestrians?

[beelsebob]

On the plus side, this will allow the US to reverse it's dumb decision to stick pedestrian crossings right at the same places as complex traffic intersections, and stop the ever-present issue of people making right turns, only to discover (sometimes too late) someone crossing the road right in front of them.

In more sane countries, the pedestrian crossings are put a decent distance away from the actual intersection, and barriers erected to stop people crossing in the intersection.

Re:What about pedestrians?

[dgatwood]

You don't need a full stop light for that. A button-operated lighted crosswalk system (lights going across the roadway that flash) is sufficient.

Re:What about pedestrians?

[Drethon]

So the preference is to double the number of places cars need to stop? I've never had trouble with people crossing when I turn, except for bikes on the sidewalk that don't slow down when going through the intersection.

Re: What about pedestrians?

[dgatwood]

Grand Rapids, Michigan is built that way. In fact, lots of other northern cities are also built that way for temperature/snow reasons; that's just the only one that I have personal experience with.

Basically, in Grand Rapids, buildings are connected together at the second floor level. There are periodic pedestrian tunnels connecting across streets to adjacent buildings as well, also on the second(-ish) floor. As a result, you can walk from your hotel for a dozen blocks and never even see the street. Of course, sometimes, you have to go up or down a ramp that compensates for height differences, and if there's enough of a height difference, you might leave the second floor of one building and end up in the third floor of another, IIRC, but either way, you stay indoors.

It would be nice if that practice would spread to the rest of the U.S., because it is much safer than having pedestrians and cars share the same space.

Re:What about pedestrians?

[Anonymous Coward]

I saw a low tech solution to this when I lived in Mexico. Down there, cars DO NOT stop for pedestrians. At certain uncontrolled intersections during high traffic periods (aka rush hour) the police would monitor certain uncontrolled (no lights) crossings. Cars would be free to drive through most of the time. Once a decent number of pedestrians built up and were waiting to cross, the police officer would walk out and block the road. This would let any build up of pedestrians cross. It might sound crude, but it actually worked very well. There was a good balance of pedestrian vs. vehicle traffic. I never waited so long as a pedestrian that I was frustrated, and the same was true for people driving cars.

Re: What about pedestrians?

[PPH]

you walk down into one and someone robs/hurts you.

We've got that covered with the Second Amendment.

Re:What about pedestrians?

[martin-boundary]

That's not efficient. Right now pedestrians don't have to press a button, so for example they might arrive at the crossing halfway while the crossing light is green, and they can just cross without slowing down. With your system, *every* pedestrian has to press the button and wait, which just causes delays for pedestrians. I argue that, especially in cities, pedestrians are more important than cars, and should get priority. Your system takes away convenience for pedestrians in favour of drivers. FAIL.

Re:What about pedestrians?

[martin-boundary]

You're not qualified to talk about traffic if you don't realize that putting pedestrian crossings at intersections is the most efficient way to let pedestrians navigate the cities. People don't know how to walk through walls you know.

Re: What about pedestrians?

[jenningsthecat]

...eventually pedestrian crossing could be completely eliminated since anyone could just hop in a driverless car and go a couple blocks away instead of walking and if you wanted to go for a walk/run then you would take a driverless car to a nearby park or gym.

WTF? Having to go to some special pedestrian-designated place just to walk or run in the city? Better yet, having to try to get a dog into a car which probably has a no-pets policy just so you can take Fido for a walk? You're trolling, right?

"First you use machines, then you wear machines, and then ...? Then you serve machines." I was going to say "No, thanks", but I'll amend that to say "Hell no. No. Fucking. Way."

Re:What about pedestrians?

[Barlo_Mung_42]

You're right that a button is antiquated. They would just need cameras at the corners to recognize them and priorities things. Matching the speed of the cars to all pass through without slowing down is only the best case scenario. As someone walks up to the intersection there's no way to know if they're going to cross or turn the other way so they'll mess up the timing regardless. Large cities a long line of pedestrians could cause major traffic jams if there's no way to share the road. You're wrong, people in cars are also important so there needs to be a system of sharing or two kids could just walk back and forth and tie a city in knots.

Re: What about pedestrians?

[mikael]

Cities in Canada had undergound shopping malls constructed in downtown along with access corridors between the different street blocks. That way, someone can walk from their apartment or from the metro station to the shopping area and never have to deal with the extremely hot or cold weather.

Re: What about pedestrians?

[turbidostato]

"eventually pedestrian crossing could be completely eliminated since anyone could just hop in a driverless car and go a couple blocks away"

I don't want to go two blocks away, I just want to cross the street, dammit.

Re:What about pedestrians?

[turbidostato]

"That's not efficient. Right now pedestrians don't have to press a button"

No, they have to wait for the green light, disregarding if there's traffic or not. Much more efficient. And of course, the fact that there already *are* traffic lights of this kind on places with low pedestrian traffic (with retention times studied for the standard flow) shouldn't concern you. And the fact that whenever something like this is applied, it'll be because *everything* will be much more connected, so pedestrians arriving will also be detected by the system, shouldn't be a concern for you either.

Your lack of imagination is daunting.

Re:What about pedestrians?

[turbidostato]

"You're not qualified to talk about traffic if you don't realize that putting pedestrian crossings at intersections is the most efficient way to let pedestrians navigate the cities."

Putting pedestrians crossing on intersections *with* traffic lights is quite convenient. Putting pedestrians crossings on intersections *without* traffic lights is quite efficient... if you want to get rid of pedestrians.

You are right about qualifications. Basically everything is already invented, so what I've seen is like some 80% of people commenting here seemingly haven't driven beyond their suburbs since the various designs for roundabouts, where to put -or not, traffic lights and pedestrian crosses depending on the kind of crossing, number of branches intersecting, with how many lanes, pedestrian flow and relative priority... all seem to be novelties to them.

Re:What about pedestrians?

[stu72]

Why force the humans to beg to cross safely? Why not all green for pedestrians and the machines have to ask permission to cross?

Re:What about pedestrians?

[stu72]

Thank you, came here to say this 1000x

Re:What about pedestrians?

[stu72]

When you say heavier, I'm going to go ahead and assume you mean, "with higher kinetic energy". A heav car at 1mph can still kill but only if it's victim is unusually slow. A light car far at 100 mph will kill anything in its path.

Do you also support applying this of logic to other objects possessing fatal amounts of kinetic energy? How about bullets? We could just dispense with the whole notion of law and justice and police and just say, "if you don't want to get shot, don't walk in front of bullets"

People choosing to drive are choosing to escalate the energy available for property damage and harm to life. They should bear the responsibility of this, even if that means, *gasp* sowing down.

Re: What about pedestrians?

[AvitarX]

Pretty sure the car that can handle slipping through cross traffic can handle stopping at a stop sign and checking for pedestrians.

If anything, you won't be able to rolling stop.

Re: What about pedestrians?

[Etcetera]

Basically, in Grand Rapids, buildings are connected together at the second floor level. There are periodic pedestrian tunnels connecting across streets to adjacent buildings as well, also on the second(-ish) floor. As a result, you can walk from your hotel for a dozen blocks and never even see the street. Of course, sometimes, you have to go up or down a ramp that compensates for height differences, and if there's enough of a height difference, you might leave the second floor of one building and end up in the third floor of another, IIRC, but either way, you stay indoors.

It would be nice if that practice would spread to the rest of the U.S., because it is much safer than having pedestrians and cars share the same space.

Every time I visit anywhere that has those (downtown Dayton, OH has one) I always feel like I'm in some weird dystopian '70s sci-fi movie.

You should move to San Diego. We don't even build car or light-rail tunnels unless there's literally no other possibly option because we always want to be able to see the outside. We also have outdoor malls, which tourists seem to find astonishing.

Re: What about pedestrians?

[dAzED1]

so...your future utopia involves the roads being cram-packed with driverless taxis just waiting to pick up someone for a 2 block walk? You don't think Manhattan isn't already cram-packed with taxis now? Do you really, *really* think the main reason - or even a minor reason, for more than just a handful of people - that people walk 2 blocks instead of calling a taxi to drive them 2 blocks, is that they'd have to wait for the taxi? Err...ok. No sense arguing with you, if that's the case. Is a taxi going to drive you to your desk in the office, too?

Re: What about pedestrians?

[dAzED1]

oh, and on the "robs/hurts you" bit - your argument was for a travel of a couple blocks. If you're too out of shape to walk 2 blocks, the mugger is going to catch up with you anyway. With your swarms of taxis on the road that are so prolific there is always one right next to you ready to go, traffic won't be moving well thus the mugger can just walk along side you the two blocks and get ya then. Also, that means you live in an area where muggings occur, so 2-block-taxis or no, you're gonna get beat up at some point.

"We can make you safe..."

[Nova Express]

"...you just need to surrender your individual autonomy and hand over total control to us. It's necessary for increased efficiency!"

You know, this argument seems strangely familiar...

Re:"We can make you safe..."

[spire3661]

The screams of the 30,000 people that die each year in cars should be enough to silence this argument as we move toward autonomous cars. Its time to remove human ego from transportation. We may live to see a day where self-drive is illegal and i cant wait.

Re:"We can make you safe..."

[WaffleMonster]

The screams of the 30,000 people that die each year in cars should be enough to silence this argument as we move toward autonomous cars. Its time to remove human ego from transportation. We may live to see a day where self-drive is illegal and i cant wait.

Why assume self driving would always carry the same risk going forward? In a future where electronics can drive reliably what prevents the same electronics from rendering self driving safe?

Re:"We can make you safe..."

[pslytely psycho]

People die every day. I prefer the government keep the hell out of trying to make my life 100% safe. I like some risk in my life. And I love to drive.

Cruising down the freeway in a Corvette at 80mph with the top off and eight cylinders rumbling is the height of non-sexual euphoria.

Self driving sports/performance cars would be the most boring thing ever created.

At least self driving motorcycles would be further down the road, so I still would get to drive my Harley.

Re:"We can make you safe..."

[dgatwood]

The only way to do that would be to allow the computer to override your decisions. If the electronics are able to take control over the car without your permission, you aren't really in control; you just have the illusion of control. At that point, you might as well just let the computer drive so you can kick back and check your email, send text messages to your friends, and make out in the back seat... or whatever.

Re:"We can make you safe..."

[spire3661]

Too bad we arent talking about your wants and desires. In the context of public policy, roads are for the utility of transportation and nothing else. We will end self-drive within the next 40 years (or at least make the insurance so expensive it will cost too much ). Count on it. oh, by the way, Motorcycles will be the FIRST to be outlawed, not the last... They are an unsafe conveyance at highway speed. If they were invented today, they would have never made it past legal..

Re:"We can make you safe..."

[FunkSoulBrother]

Cruising down the freeway in a Corvette at 80mph with the top off and eight cylinders rumbling is the height of non-sexual euphoria.

You need to try more drugs

Re:"We can make you safe..."

[WaffleMonster]

The only way to do that would be to allow the computer to override your decisions.

Does AEB override your decisions? What about ABS? Traction control? Power steering? Speed/Rev limiters? CTA? Cruise control?

If the electronics are able to take control over the car without your permission, you aren't really in control; you just have the illusion of control.

In the typical scenario where I get in my POS, drive to work and no electronic dumbass driver alert goes off my driving experience is none different.

The only difference is in the exceptional case where the dumbass alert goes off people don't end up in hospitals or morgues.

You are free to argue adding constraints is the same as "illusion of control" but I'm not sure how such statements are any more useful than arguing "free will" is an illusion.

At that point, you might as well just let the computer drive so you can kick back and check your email, send text messages to your friends, and make out in the back seat... or whatever.

I think *IF* this automation can be made to work reliably it would be great. It would give drivers more options than they have now. If they don't like driving or don't feel like driving they don't have to.

Where I disagree is the assertion manual driving will be necessarily denied because it is deemed to be unsafe. If the technology exists for vehicles to drive themselves at or better than human capability certainly the technology would also exist to assist drivers who decide they want to drive themselves.

Re:"We can make you safe..."

[dgatwood]

Does AEB override your decisions?

Yes. Try going off-roading with that feature, and you'll see the problem. :-D

What about ABS?

Yes, but only to a very limited degree. There are situations where ABS makes accidents more likely. They're just far less common than situations where ABS reduces the risk of accidents.

Traction control?

Again, yes, but only to a very limited degree, by reducing the amount of power to the wheels. And again, there are rare cases where this is the wrong decision.

Power steering?

Of course not. Power steering is a purely passive system that converts mechanical motion into stronger mechanical motion.

Speed/Rev limiters?

Yes. Massively, yes.

CTA?

California Teachers Association? I have no idea what you mean here.

Cruise control?

Nope. It does precisely what you tell it to do.

The only difference is in the exceptional case where the dumbass alert goes off people don't end up in hospitals or morgues.

The original question was what prevents computers from rendering manual driving safe, which I read to mean "as safe as computer driving". Although there are some limited automated safety systems that are less complex than full computer driving (e.g. automated emergency braking), those limited features won't get you to that point. The only thing that will get you all the way there is a computer that monitors your driving, doing basically the same things that it would do if it were driving, but only taking control if necessary to prevent a collision. If the computer is that advanced, then it is also capable of driving for you. So why would you want the hassle of driving when you can just let the computer do it and be free to spend that time relaxing and/or doing something more productive?

More to the point, why would any car manufacturer spend the extra expense of adding a steering wheel, gas pedals, brake pedals, a gear shift, a handbrake, turn signals, manual wiper and headlight controls, etc. when 99% of users will just put it on autopilot anyway? Maybe a few car makers will initially, but I'd expect them to quickly taper off, as those things greatly increase the complexity and cost of the vehicle, while providing little benefit.

Also, it is probably physically impossible to make manual driving as safe as computer driving will eventually be. Why? Steering wheels kill. In an accident, if everything else is equal, you're less likely to die in the passenger seat than in the driver's seat, because of that unmoving object right in front of your body. Steering wheels are a lot safer than they used to be, mind you, but the added risk is still decidedly nonzero and probably always will be (where "always" is defined as "until steering wheels go away" in this case).

Where I disagree is the assertion manual driving will be necessarily denied because it is deemed to be unsafe.

It is also massively inefficient. Do you know how much money we spend on traffic lights every year, both in terms of maintenance and power bills? Just putting one in will cost you $80-100k, plus about $1,400 per year in electricity, plus having to rewire it every time you do any minor road paving work, plus other maintenance. There are over 300,000 traffic lights in the U.S., which means we spend about half a billion dollars every year just keeping them lit. The only reason they exist is to limit the flow of non-automated vehicles. Get rid of the manual drivers, and you can see some significant cost savings.

Knowing government, that alone is sufficient reason to assume that manual driving will eventually be highly restricted or eliminated. :-)

But pedestrians will need.

[niftymitch]

Pedestrians will need implants!
It astounds me daily how many volunteer to be a Darwin award winner
as they step off the curb without looking both ways expecting vehicles to stop for them.

Near hotels in the UK and Australia are words on the curb reminding Mericans to
look "the other way" too. Those little reminders are fully lost to the goof with
his or her nose in his phone.

I recently found a crazy turn signal in this area where vehicles were given
a right turn green arrow at the same time pedestrians were given a walk indication.
Traffic folk are so worried about automobiles that they ignore bicycles and
pedestrians.

Re: But pedestrians will need.

[BarbaraHudson]

Ignoring pedestrians and cyclists is a feature, not a bug. How else are they going to get pedestrians and cyclists to usecself driving cars except by making walking and cycling too dangerous, hence illegal? And then it will be your fault when you get hit, and what pedestrian wants to pay the high cost of insurance just to walk?

Re:But pedestrians will need.

[silas_moeckel]

That is the correct response, get the people off the road. Elevated platform tunnels etc etc etc.

It realy does not matter as automated cars have to take pedestrians into account outside of cities we have plenty of rarely used crosswalks that do not have stop signs and traffic laws to always yield to crossing pedestrians. It actualy works here in the burbs/rural setting.

Um, Pedestrians?

[Striek]

This is all great - I can see it might actually work (but then again, it might not).

But I am typically a pedestrian. How will I cross the street? Do these smart cars just do their "ballet dance" around me? Do I press a button and tell all cars to stop, while I proceed through the intersection? All the article can say is

seamlessly knitting together flows of cars, pedestrians and bikers.

How, exactly? I can see a system of autonomous cars being able to do this, but TFA mentions squat about how it will handle bicycles and pedestrians; only that somehow, magically, it will. There are no pedestrians in the promotional video, so it's pretty hard to tell. How will a computer system predict when I will cross the road? With no "light" to instruct me, how am I supposed to know when to cross?

Seems to me they left out a rather important thing about intersections - the humans using them are inherently unpredictable.

Re:Um, Pedestrians?

[dgatwood]

Seems to me they left out a rather important thing about intersections - the humans using them are inherently unpredictable.

Not at all. When you're driving and you see a pedestrian walking up to the intersection, you have several seconds of visibility before he or she reaches the edge of the curb. If you'll arrive first, you keep driving. If the pedestrian will arrive first, you stop. If you aren't sure, you stop.

And with computers driving, the car five cars ahead of you spots the pedestrian, and your car knows that it will have to stop, so it starts slowing down ahead of time to minimize or avoid the stop, as do the cars behind yours. By the time your car reaches the intersection, if the pedestrian has already crossed in the gap, your car speeds back up. If not, it stops and waits for the intersection to be clear, then speeds up.

The only really unpredictable things are A. whether the pedestrian will recognize that your car is slowing down or will wait for you to completely stop, B. whether the pedestrian will speed up or slow down between first detection and arrival at the curb, and C. whether the pedestrian will ever finish crossing the road or die of a heart attack in the middle of the driving lane. And all of these things are detectable with cameras, and can be compensated for, at least to some extent.

Re:Um, Pedestrians?

[R3d M3rcury]

How will I cross the street?

You'll call up an autonomous car, obviously!

"Hello, car. Please take me across the street."

Re:heh...

[Intron]

Except wouldn't slowing down the cars or making them stay apart by "safe distances" decrease the traffic flow anyways? So you're just back where you started from

Not true. They just added a traffic light at a "T" intersection near me. Now it only handles 1/3 of the traffic since each leg gets green only 1/3 of the time. Traffic is now backed up on all 3 legs where before it was only on the left turn leg.

Roundabouts

[Greg Merchan]

We can dispense with them already by using roundabouts.

Re:NOT SAFE!

[Anonymous Coward]

Automated cars would be communicating with each other and road systems.... so there would be no need for visual-based things like traffic lights. This is a pretty simple concept.

If you remove humans? yes.

[Lumpy]

Problem is humans are at the core selfish assholes, which is why we have traffic lights.

It turns out...

[Etherwalk]

"...you just need to surrender your individual autonomy and hand over total control to us. It's necessary for increased efficiency!"

You know, this argument seems strangely familiar...

It turns out that's how it works. There is often a tradeoff between safety, efficiency, and civil order on the one hand, and personal autonomy on the other. There is a cost to personal autonomy. If we give everyone nukes, sooner or later someone blows up a city. If we give everyone AKs, sooner or later someone shoots up a school or a mall. If we give everyone cars, sooner or later people kill each other with them, deliberately or by accident. If you give people the ability to control dangerous things with their free will and imperfect responses, other people die and maybe the actor dies.

So you have to make a judgment call about the level of autonomy you desire and how many lives it's worth. If you can reduce traffic accidents a thousand fold, isn't that worth at least making automated driving the default?

Wow MIT figured this out, wow.

[EmperorOfCanada]

I believe that I have seen animations for driverless cars careening through intersections for years, as in going back to the 70s.

What I find much more fascinating is the economic impact of this sort of thing. How much economic activity is generated by traffic lights. Obviously there are the companies making them, maintaining them, their electrical usage, the cost in having people idle at them, and the ticket taxes generated by having police ticket people for not obeying the tax laws.

I would not be surprised that the savings to the taxpayer and the public by removing a single unimportant traffic light could be well in excess of $100,000 per year. For instance there was one major downtown street near my old house where they had the lights perfectly timed so that you pretty much missed all of them. Thus the average speed on that street in low traffic was maybe 15mph tops. With about 8 lights and the street being 1.5 miles the savings in time alone to get that up to 25mph would be astounding, let alone in gas.

Also many busy intersections are pretty much car accident factories. So to remove those would be just another layer of costs removed.

But what is interesting about all the above costs is that they are all very parasitical. Most of the costs in having a traffic light don't really "benefit" society. Obviously a typical traffic light today massively reduces accidents and other problems but when we have 100% SDCs their removal will only be a net benefit to all.

Where this is also going to get interesting is that some traffic lights are political. For instance there is a neighbourhood in my old city where a 3 way stop was replaced with a traffic light. This then encouraged people to take a short cut through a rich influential neighbourhood so within about 10 days the light was removed and went back to a 3 way stop. I can see attempts to prevent self driving cars from "navigating" through rich neighbourhoods but that is going to impinge upon fundamental freedoms and those laws are going to be hard to sustain. But with enough political influence there will be a way to keep the plebs away from the rich. Which will simply be part and parcel of the many many stupid laws that I see coming when politicians don't realize that every stupid traffic law they implement will be diligently followed by computerized cars. I can see every squeaky wheel along rural highways calling for the speed limit to be dropped in front of their house because of "the children" thus the speed limit will be very much an indication of how influential any given household is in rural communities.

via Internet?

[DogDude]

I don't know about Europe, but in the US, Internet connections are crap. Utterly unreliable and slow. There's no way in hell vehicles will be using the Internet for mission-critical stuff. No way. We have no infrastructure to support it, and as long as the government is bribed to not regulate it, we never will.

Re:via Internet?

[gweihir]

Europe? Wired Internet is usually pretty reliable, wireless networks are very reliable. And the last electrical outage where I live was more than 10 years ago. But this is about vehicle-to-vehicle communication over short ranges, an entirely different thing.

Re:via Internet?

[gweihir]

Never, obviously. But there are occasional storms strong enough to blow down roof-tiles (and blow over people that are not careful). They usually do not do much damage due to much better building quality. Incidentally, I have been hit with two earthquakes strong enough to wake me up from the shaking. Damage in total: a single-digit number of smaller houses developed cracks. Now, if you go to Turkey (not really part of Europe), you find a lot of people killed by earthquakes due to shoddy building quality.

Your suggestion that the US has higher environmental risks is flawed.

There are Simpler Ways

[nukenerd]

Before doing that it would be very advantageous and far simpler to review the algorithms controlling existing traffic lights, which seem to have been devised by idiots.

There are junctions near me where the lights hold everything stopped for anything up to 30 seconds (and there no pedestrians crossing, in case you were wondering). Another junction has a pedestrian crossing 25 yards before it, and the two sets of lights are not synchronised; usually they are out of phase so when the junction lights go green only four cars are released with the rest of the queue held back at the pedestrian lights, and apart from those four cars nothing moves for about a minute.

There are some very easy wins to be had without waiting for pipe dream technology to appear.

Re:There are Simpler Ways

[dgatwood]

You're operating under the mistaken notion that the primary purpose of traffic lights is to increase safety and maximize throughput. In reality, the primary purpose of traffic lights is to make city streets so annoyingly slow that drivers will be forced to get on the freeway as soon as possible, then sit there at 3 MPH. It's called "traffic calming", mostly ironically, as the primary effect appears to be an uptick in road rage.... :-D

Re:There are Simpler Ways

[nukenerd]

You're operating under the mistaken notion that the primary purpose of traffic lights is to increase safety and maximize throughput. ...It's called "traffic calming",

I'm not under that notion at all, well aware that it is for "calming", as you can see from other posts of mine here, now and in the past :-). As you say, it's infuriating, not calming, and I have even complained to my local council about it.

Bad headline

[jgotts]

Neither MIT nor anybody else has shown me how their autonomous cars will work with the potholes on Michigan roads, the black ice common in this area, ice in general, the lack of municipal funds to plow many roads, driving on the highway with 6-12 inches of snow on the ground with more coming down, or driving with whiteout conditions where the road is not visible. In Michigan in many places no indication of lanes is visible. The local residents simply memorize where the paint used to be. It's very exciting to us to see a freshly painted road surface.

Nobody can afford to maintain their vehicles so that they're like new. How are autonomous vehicles going to work when the alignment is off, or a tire is leaking, or the suspension needs work? Due to potholes, everybody in Michigan drives vehicles long past when these problems should have been fixed. Not because we want to, but because the average car in Michigan needs $500 in suspension work every year.

Once they have reasonable solutions for those problems, then we can talk about removing stop lights.

Re:Bad headline

[ThosLives]

As a fellow Michigan resident, I'd rather fix whatever policy or lack of oversight that allows our roads to be so bad (when they are paved) and the madness that is refusing to pave roads in major metro areas in the first place. I almost feel ashamed to live in this area - why are people so against making first-class infrastructure? (Cue theories about the mob, wanting to encourage more new car sales, and similar here.)

And I second the comment on the terrible lane paint they use here; any time the roads get the slightest bit damp the markings vanish. It's like they never heard of reflective paint or something.

And save gas

[pubwvj]

My son and I were just talking about this idea a couple of weeks ago. What we were focused on is how much fuel and break pads would be saved by eliminating traffic lights. Here in Vermont they've been putting in a lot of round abouts which have a similar effect but it would be even better we thought if auto autos communicated with each other and solved this.

Re:If you remove humans? yes.

[rwyoder]

Problem is humans are at the core selfish assholes, which is why we have traffic lights.

Agreed. In the excellent book "Traffic", (by Tom Vanderbilt), he cites studies showing that in comparison to traffic lights, roundabouts:
- Move a lot more traffic.
- Have a fraction of the accidents.
- Have much less severe accidents, resulting in a fraction of the fatalities.
 

Meh...

[sycodon]

...Happening already

Re:Meh...

[Tony Isaac]

This is funny, but it's probably a pretty true-to-life illustration of what automatic navigation without traffic lights would look like. Whoever designed the MIT simulation, obviously doesn't live in Houston (or any other big US city)!

Now THAT....

[Ferretman]

....is an interesting possibility!

Ferret

They won't go away.

[Khyber]

All it will take is one idiot on their cellphone trying to cross the light-less intersection to show MIT just how short-sighted this idea truly is.

Re:NOT SAFE!

[Calydor]

Most likely the autonomous cars won't be slamming on the brakes and revving up from a stand-still like you and your parents seem to. Barring emergencies it will probably be doing the usual slowing down, which you will anticipate by seeing an intersection ahead, followed by comfortable acceleration, which you will anticipate by NOT seeing an intersection ahead.

I don't think anyone's talking about windowless boxes just yet.

I am funnily reminded of a Donald Duck story I read several decades ago, in which Scrooge got an idea for a fully automated road network from watching the kids play a (VERY early) racing game on a (REALLY VERY EARLY) computer. The mayor of the city was in a panic after trying it out because he had no control over the city, but reached the conclusion that the younger generation would probably adapt better.

Unfortunately the wireless system to make the cars communicate caused massive interference on the TV signal in the city, so the whole plan ended up scrapped. I guess it's a good thing all TV signals have gone digital since then! ;-)

What about crosswalks for pedestrians, then?

[Rinikusu]

Some of us walk.

Re:NOT SAFE!

[BronsCon]

The post you are replying to just explained why that is incorrect.

Like traffic circles

[mauriceh]

Already been done with traffic circles.
In some places these are problematic due to some people being too stupid to drive

Assuming all cars cooperate

[Hotawa Hawk-eye]

It seems like the assumption is that all cars are willing to participate in the communication necessary to coordinate the traffic flowing in each direction to avoid collisions. When that happens, I think it would be like Japanese synchronized walking. But if you throw an older, human-driven car into the mix, does the coordinated effort break down? If you threw someone who hadn't practiced with the group into that YouTube demonstration, it would quickly devolve into chaos.

Not everyone is going to want to turn over driving to the machine, and even if they are required by law to do so there will almost certainly have to be some scenarios grandfathered in. How will those situations be handled?

But, but ...

[PPH]

... with no traffic lights, how will I know when to accelerate like I do now on an amber?

Re:NOT SAFE!

[William Baric]

Like pretty much everybody, I never look at what's ahead when in a bus. The bus stop and go, and I just don't care. It doesn't "damage" myself. Also, my parents never said "whoa" and "go" to signal me anything when I was a kid.

Sorry, but your comment is idiotic.

Re:NOT SAFE!

[Dunbal]

Your parents had to say "whoa" and "go" to signal you to adjust yourself.

I grew up in an age when my parents didn't do that. In fact they'd smack me if I made too much noise in the back seat. Kids nowadays have no idea how spoiled they are.

But what about

[rossdee]

Pedestrians? Cyclists?

Not everything is going to be controlled by computer

Re:But what about

[aberglas]

What makes you think that not everything will be controlled by a computer?

http://www.bostondynamics.com/...

Re: NOT SAFE!

[dilvish_the_damned]

People did that when automobiles were first becoming mainstream because the driver was used to talking to their horses. Are you sure your not just old?

Except for...

[Berkyjay]

...those pesky humans who want to cross the street. If only we could find a way to autonomously walk, we could program humans not to be so annoying.

I already knew that

[Tablizer]

Ain't the future, I've been ignoring them for 30 years.

Re:If you remove humans? yes.

[Lumpy]

not impractical at all on the smaller level. They are adding them all over the place in the residential areas around here. The huge advantage is that they slow down the idiots that blast through the neighborhood at 50. most people now drive far close to the speed limit.

Re:heh...

[turbidostato]

"Except wouldn't slowing down the cars or making them stay apart by "safe distances" decrease the traffic flow anyways?"

In practice, no. Human reflexes are a bit slow when compared to traffic speeds so under changing environmental conditions (like crosses and intersections, or even on almost saturated straight highways) amplified waves tend to appear and going slower almost always guarantee better throughput)

Deja vu

[Ol Olsoc]

This is all beginning to sound like those ancient Popular Mechanics visions of the future.

Flying cars for everyone!

Weee here I go!

[AndyKron]

I'll pretend to have this technology as I go screaming through the intersections.

Don't Hold Your Breath

[JimSadler]

In order for this to work all vehicles with human drivers would have to be taken off the roads. Also, classes of vehicles such as motorcycles might have to be eliminated. And I suspect that passengers would have to be blindfolded and the driver unable to interrupt the computer in control. People entering an intersection at 40 miles per hour would be terrified to see another vehicle, at a 90-degree angle that would only miss hitting them by a foot or so. Considering peoples' passions for driving I don't see this happening for many decades.

Rushing in

[fluffernutter]

We seem to be really rushing in to this whole automated car thing and people are going end up dead. Seriously.. .Google gets into an accident with a BUS two weeks ago and this week they are trying to justify cars without steering wheels? Have they gone back to the drawing board and tested all possible scenerios related to the accident two weeks ago? Have they tested in all kinds of weather? I invite them to where I live, where lanes on a road don't matter and a four inch deep rut can send a car sideways if you don't line the tires up with it. Now put that rut in traffic.

Seriously, they are not ready and I wouldn't expect them to be ready for at least another five years at least.

Self-driving car

[BlackBindy]

So, the self-driving cars should be spreaded all around the world first. for now, I cannot imagine the road without any traffic lights, but I cannot say that that is impossible either because I've seen so many impossible things became possible.

Re:NOT SAFE!

[BronsCon]

It's not about the device, it's about the people inside the device. They still need the cues to avoid injury and/or motion sickness, just as examples. It was explained in the very first sentence.

Re:NOT SAFE!

[turbidostato]

"It's not about the device, it's about the people inside the device."

Yeah, I know: that's why I said it's irrelevant.

"They still need the cues to avoid injury"

On one hand, no, they don't. It is not cues what avoids injuries, but low deltas on speed/direction vector changes, which only depends on the driver's -either human or artificial, ability to anticipate and react on its surroundings. On the other, *even* if cues are in fact needed (say, once it's admitted there will be situations where awareness/reaction speed is not guaranteed to be below a required threshold) why do you think the cues need to come from an specific external device, be it traffic lights or any other, when the vehicle itself has already gathered all the required info and this is in fact, the very reason why it can go without external traffic lights? For all that matters, the car could very well be completely opaque and create all kind of cues on a 360Â screen.

"and/or motion sickness"

The way a car is driven and/or gathers information about its surroundings in order to take some action is, again, orthogonal to the visibility it offers to their passengers.

But note ---

[purplie]

If you think improvements like this are going to speed up your commute, read about

• Marchetti's Constant --- technology doesn't improve commute times
• Jevons paradox --- increased efficiency leads to increased resource usage
• Lewis-Mogridge Position --- traffic expands to fit the available road space

Re:NOT SAFE!

[BronsCon]

If the discussion is about whether or not we'll need traffic lights in the future, the fact that people need them is extremely relevant. Since I can't seen to illustrate that well enough for you with people inside, let's consider the ones outside, who will still be riding bikes, running, jogging, and walking. All of this will occur adjacent to, on, or across roadways and any of it could result in abrupt speed deltas (think cyclist hits small rock, or swerves suddenly around it, ending up in the path of a vehicle, or pedestrian trips and falls into street). And pedestrians needing to cross? Lights.

As for motion sickness, that actually has damn near nothing to do with delta and basically everything to do with cues not matching up. No matter how fast the vehicle can throw up its own cues, there will always be some delay (as the signals travel through a mile or more of wire, trace, and computational pathway, and as the data moves from sensor to RAM to CPU to signaling device, and as the vehicle's on-board computer analyzes and processes to determine if a specific cue is important) and any delay is enough to cause issues for a large number of people. A window solves that and a light helps. So yes, it is important to the matter at hand, making the people in the vehicle relevant to the discussion.

Sure, the computer can process and react more quickly than a human, even a well-trained human, but there is still (and, the laws of physics dictate, always will be) considerable delay.

About half of ny examples relate to windows on the vehicle, moreso than lights on the roadway, and there is a good reason for that. The next logical step from "the cars don't need lights" is "people in the cars don't need to see where they're going". These are both false statements, however, as cars will still need lights to control how they affect the flow of traffic they cannot communicate with (pedestrians, for whom a light will still be necessary, even if the car can be signalled by radio, though I'll also illustrate why that is a horrible idea) and people still need windows to avoid motion sickness and be able to brace when a sensor fails (or a pedestrian with a smartphone ans $20 worth of electronics sends a BRAKE HARD signal) and the vehicle fucks up.

This could get really "interesting"

[herbierobinson]

In order for this to be implemented in the real world, there will have to be a standard.

The standard will be subject to creative interpretation.

Vendors will compete on how good their creative interpretation is at getting you through the intersection faster.

In other words, if this happens, it will turn into a high tech version of the game kids call "chicken".

Re:NOT SAFE!

[turbidostato]

"If the discussion is about whether or not we'll need traffic lights in the future, the fact that people need them is extremely relevant."

Not that I ever said otherwise since that was not what this thread was about: "It's not about the device, it's about the people _inside_ the device.". That's what you said.

"let's consider the ones outside, who will still be riding bikes, running, jogging, and walking."

So you change the issue, then. OK, I'll tell you are not starting right: there're two different sets there: the ones within the road and the ones not in the road. Different sets will probably be managed in different ways.

"let's consider the ones outside, who will still be riding bikes, running, jogging, and walking. All of this will occur adjacent to, on, or across roadways and any of it could result in abrupt speed deltas"

So what? I already concluded that speed/direction deltas and traffic lights have no relationship.

"think cyclist hits small rock"

I do. How in hell any traffic light can have any relationship with that?

"And pedestrians needing to cross? Lights."

Yes, traffic lights *can* be a solution, it doesn't mean it's the only one. And given the perspective of autonomous vehicles in the road, it's probably a very bad one in fact, since it's utterly opened to sabotage and disruption. Think a bit yourself about the issue and you'll understand why and you also see around you solutions that are already in place, albeit not in street cross-sections (hint: automatic doors).

"A window solves that and a light helps"

A window is something that just has appeared on your argumentation *now* and even then, all you are using "your" window for is for "everything *but* the traffic lights", so now I don't really now what your point is: are traffic lights needed or are windows? Or both? And by the way, signals on wire will go basically at the speed of light (moreso if we are talking about fiber optics or wireless), just the same speed than the one going through the windows and then, you have the computer's reaction time versus the human's reaction time, and the computer assisted mechanical interface versus the human-actioned interface. Can you bet which is faster?

"The next logical step from "the cars don't need lights" is "people in the cars don't need to see where they're going. These are both false statements."

They are *so* false that passengers in a bus *already* don't see where they are going and what happens?... Uh, nothing at all.

"or a pedestrian with a smartphone ans $20 worth of electronics sends a BRAKE HARD signal) and the vehicle fucks up."

Or a pedestrian just makes a sudden movement into the road (either maliciously or not) and the current very human driver also fucks up. So?

Re:NOT SAFE!

[BronsCon]

You explain that electrons through a circuit move at the same speed as light through a window, which, while technically correct, is a fallacy as, as I explained (and you ignored), the circuit is a much longer path and the delay is measurable. You pick the points I am making, rather than the facts that back them up, to attack; you do this because you have no facts of your own to argue with and the audience here is generally smart enough to see that.

Also, you can compromise an automatic door with $20 worth of electronics, no smartphone needed. And how do you propose the system you describe would signal the vehicle? Hackable, recordable, replayable, jammable radio waves? Lights are a fair bitbetter suited for this type of signaling, as they not only provide signal data, but also easily verifiable sourxe data. Where is that radio wave coming from? You can approximate the direction and guess, but you can actually see where the light is. The same applies to cars.

Speed

[Etherwalk]

What would be the comparative advantage with respect to a roundabout?

I know they are not very popular in the US, but they can be very efficient, and prevent the frustration of waiting at busy intersection (especially if going in the non-popular direction).

Speed and wear-and-tear. A computer system can keep traffic moving through an intersection at speed in both directions, interleaving cars. But if there is a roundabout, both of the cars have to make (usually) tight turns, resulting in centripetal force and a corresponding need to slow down.

For human drivers, roundabouts cause fewer deaths (because people slow down) but more collisions (because more judgment is involved in determining stay-vs-go).

Re:NOT SAFE!

[turbidostato]

"as I explained (and you ignored), the circuit is a much longer path and the delay is measurable."

It _is_ mensurable, yes; it is also non-significant: human average reaction time for an unexpected circumstance on-road: around 1,25 s (yes, *that* bad). Even reaction time for an *expected* event on road is like 0,250s.

Time for light to linearly cover 100m: 0,0003s. Time to cover 1000m: 0,003s. So the travel part, even with cable being 100 times longer than line sight, is like 0,25% of total time. In other words: irrelevant. Heck, even over standard Internet the round trip from my position to a well known site about 6000 miles afar, with its hops and routers in the way, is like 0,160s!

"You pick the points I am making, rather than the facts that back them up"

I pick the points because your overall position (even if you made it clear, which you do not, as far as I can get it) is ludicrous. At least your intents on fine points are clearly questionable: trying to debunk "waving hands in your general direction" is much more difficult, I confess, than trying to set why self-aware cars, driven in a full-aware environment don't need traffic lights nor time-managed traffic slides. Just to pick one, the argument that sorroundings-aware cars still need traffic lights in order for their occupants not to crush their heads against the bodywork is probably the most ludicrous of all.

"Also, you can compromise an automatic door with $20 worth of electronics"

Yes. And how many times have you seen an automatic door compromised? How many times, in fact, do you see compromised a door with a simple "do not trespass" panel on it? Having a door, or a swinging barrier is not in order to make them invulnerable but to provide an obvious semantic context (just like a traffic light does) -in this case minimizing the inconvenience of a dumb child jumping over just because he knows the cars in front will forcibly and reliably brake.

"Where is that radio wave coming from? You can approximate the direction and guess, but you can actually see where the light is. The same applies to cars."

Ok: so the problem here is that you just lack the required modicum of engineering knowledge or how -even current, traffic signaling works. It is not a problem: not everybody should know about everything. It's just that it makes this conversation tad non-productive.

Re:NOT SAFE!

[BronsCon]

human average reaction time for an unexpected circumstance on-road: around 1,25 s

I not only acknowledged that, I explained why that doesn't matter. Simply put, it's because we're not worrying over reaction time, we're worrying over perception of all inputs (to the human, not to the vehicle) agreeing with each other, as motion sickness is triggered when they do not. It does matter, because any perceptible delay (and the difference between data traveling through a mile of circuit, with various delays as it moves from acquisition, to memory, to processing, back to memory, then to the signaling device, be it auditory or visual, versus light traveling through 1/8" of glass, is perceptible; there's a reason people get motion-sick with virtual reality) will trigger it. Again, you're ignoring what I'm saying and arguing with what you want me to be saying.

Ok: so the problem here is that you just lack the required modicum of engineering knowledge or how -even current, traffic signaling works.

Electromechanical or purely computer-driven? That I know to ask should tell you something.

Re:NOT SAFE!

[turbidostato]

"as motion sickness is triggered when they do not"

There it comes again... How the heck do traffic lights relate to motion sickness!!!??? Specifically, how the heck should the relationship between traffic lights and motion sickness for an occupant of a self-driving car in a future scenario would be any more relevant than the traffic lights to motion sickness relationship for the holder of a seat in the fifteenth row of a human-driven bus on a present-day one?

Re:NOT SAFE!

[BronsCon]

Go back and read, I addressed why half of my arguments have nothing to do with traffic lights.

Re:NOT SAFE!

[turbidostato]

"Go back and read, I addressed why half of my arguments have nothing to do with traffic lights."

Yes, you have been jumping from an irrelevant topic to another all thread. So just to recap:

No, what you think it will happen, it won't happen. And even if it happened, it wouldn't be relevant. And even if it were relevant, it would be easily coped with. And even if it weren't possible to deal with, it wouldn't be any different to something already happening today and accepted.

Be it traffic lights, motion sickness, people jumping into the road, bicycles trumping into rocks or whatever.

Sorry being me the one bringing you the bad news but no, today is not the day you find a critical flaw in the study it took a MIT team to come with.

Re:NOT SAFE!

[BronsCon]

You're an "engineer", aren't you? Not an actual, trained and licensed, engineer, but someone who fancies himself an "engineer". I can tell, because you're giving a typical "engineer" response: "Your argument is external to the problem and, therefore, irrelevant." Well, the "problem" under discussion is traffic lights on roads and pedestrians, while typically an externality when it comes to roads, do often have to cross them, which makes their reliance on such lights relevant. Also, motorcycles, those run on roads, and need to be considered. As for how I got to motion sickness and the need for windows in vehicles, well... anyone reading along and actually paying attention already figured that out.

Sorry being me the one bringing you the bad news but no, today is not the day you find a critical flaw in the study it took a MIT team to come with.

A real engineer, someone cut from the right cloth for it, considers such externalities as they actually affect the problem. You failed to do so; and the MIT team doing this study didn't need to as the study was intended to be purely theoretical. Of course, if pedestrians ceased to exist and all vehicles were autonomous, we could do away with traffic signals. That's hypothetical, at best, and that was the intent of the study; the discussion you interjected yourself into was about how the study fails to apply to reality which, in all honestly, is a perfectly fine thing for a theoretical study to fail to do.

Also, thank you for that broken English, letting me know this is not your native language. I feel as though that might explain (not excuse) some of your misunderstanding. Sorry to be the bearer of bad news, but no, I was not trying to find a flaw in the MIT study, so you have not prevented me from doing so. You're the one grasping at straws to apply a theoretical study to the real world, but the study didn't account for many variables which prevent this from working, because it did not have to. The study, in all honesty, amounts to "if there were no people, we wouldn't need traffic lights", which is about as obvious as it comes; no MIT education necessary.

Re:NOT SAFE!

[turbidostato]

You see? You are doing it again: "I addressed why half of my arguments have nothing to do with traffic lights" and then " the "problem" under discussion is traffic lights" (even qualified to "...is traffic lights on roads and pedestrians").

Well, another bad news for you: no, the problem under discussion is certainly NOT traffic lights on pedestrians side but traffic lights on road side, specifically how a fully surrounds-aware system (vehicle-to-vehicle (V2V) and vehicle-to-infrastructure (V2X)) could get rid of traffic lights (or any other slot-based traffic management system on intersections) *on the traffic side*. It even goes to specifically explain about the existence of a 'pedestrian' button since pedestrians, as long as they also share access to the intersection, would need a slot-based pass-through (not necessarily in the form of traffic lights on their side -and only on their side).

"A real engineer, someone cut from the right cloth for it, considers such externalities as they actually affect the problem"

You said it: as they actually affect the problem. They don't come with unrelated externalities just waving hands to make them into the problem realm.

"You're the one grasping at straws to apply a theoretical study to the real world"

What I see is you trying to debunk the theoretical study by putting into the scenario "real world circumstances" that somehow make it impossible to put it in practice, failing to both clearly state what those circumstances exactly are and how exactly they break any of the study conditions.

"The study, in all honesty, amounts to "if there were no people, we wouldn't need traffic lights""

No. Even at that very high level, the study amounts to "if there were no pedestrians (not "people"), we would increase load capacity getting rid of time-slotting traffic flow management systems (not just "traffic lights")".

Re:NOT SAFE!

[BronsCon]

... pedestrians aren't an unrelated externality. I'm not sure if you're a high-quality idiot or a low-quality troll.

Re:NOT SAFE!

[turbidostato]

"... pedestrians aren't an unrelated externality."

No, they aren't. That's exactly why the study takes them into account. It is still you the one failing to tell how the way the study takes care of pedestrians can't be put into practice.

"I'm not sure if you're a high-quality idiot or a low-quality troll."

What a coincidence!

No failsafe

[PPalmgren]

This method requires too many beneficial ifs and doesn't account for any of the adverse ifs. If there is a power outage and the intersection can't coordinate cars, what happens? If there's no line of sight to crossing traffic for the autonomous cars via weather or elevation, what happens? How does it handle pedestrians or unusual traffic like a windmill flatbed or a wide load? All of these things have to have a smooth transition mechanism.

News flash: America is a big, diverse country

[Etcetera]

Whenever someone says X won't be needed or Y is the future, I'm always amazed by the assumptions they've unconsciously made about how X or Y is used across the country. (To the extent that "meddlers" tend to be Silicon Valley or New York residents, the mistakes are often similar.)

The US is a big country, San Francisco, CA and San Diego, CA are different, let alone San Francisco, CA and San Diego, TX.

In some places, no one drives unless they have to, and many don't even know how. In others, if you don't have a car you're insane, a child, or literally are unable to acquire one. In some places, it's 72 degrees F for 95% of the year, in others you spend 5 months surrounded by snow.

It's a big country. Take a road trip or three before you offer your prescription for what people 1000's of miles away from you need. And when describing "how things are", always pre-pend with the phrase "In my State, ...."

Re:Sure...

[Shortguy881]

Yeah, you'll have to wait for all the people like me to die off. I like driving and the feel of the road too much to give it up. It wont be a decade or two before all I'll hear is this:

Wait... You drive a car... Manually?! Whats wrong with you! That's so dangerous!

I am not going anywhere...

[martinfb]

...until matter transporters are in place!

Re:NOT SAFE!

[toddestan]

Probably something having to do with today's cars, where the excessive rake and high beltlines mean that kids aren't going to be able to see out of car, especially when in the backseat. Wouldn't be needed as much in older cars where the visibility is so much better.