The most under-hyped, but most important, technology since seat belts | Hacker News Reader

The most under-hyped, but most important, technology since seat belts

scobleizer.com

115 pointsMaysonL134 comments

The thing about ABS is that it operates on the principle of least surprise: It takes what the user expect brakes to do (slow the car down) and makes it work even when traditional brakes would fail to perform that task.

So it took a technology that the user expected to work in a certain way, and made it work better in that way.

It made brake behavior less surprising to the user.

This technology seems to do the opposite (at least, the way Scoble described it). It massively changes the behavior of the car based on some sensors, and then expects the user of the car to deal with the different behavior of the systems, all in a high-stress situation with alarms going off.

Driving cars is very dangerous. The earlier we radically change how we use cars the earlier we can avoid deaths on the streets.

I want an autopilot for cars as soon as possible.

Absolutely. The moment we develop an autopilot that is safer than average human drivers by a statistically significant margin, it should be mandatory in all cars. I am hopeful that we will see this during our lifetimes.

The introduction of such systems (probably expensive at first) would also be a great opportunity to switch from car ownership to widespread car rental (ala zipcar).

The technology might get there soon but not the legal framework. I can't even imagine the legal implication of cars on autopilot killing pedestrians and drivers.

That is also probably one of the reasons we don't have flying cars yet.

http://www.wired.com/autopia/2008/01/gm-says-driverl/

Sure, but that's not the same thing as saying they will be proven statistically safer, which will be the bar you have to pass before you can mandate them for everyone. I figure that's another few years off. But I agree that it won't be long now.

Scoble's text is different from what the Ford guy says. Ford doesn't automatically hit the brakes, only preps them and when you hit the brakes, it automatically applies them 100%. Prius automatically applies the brakes and Scoble says he no longer instinctively hits them himself. I wonder if that loss of instinct is worth the benefit.

In practice what it means is that if you're about to crash into something, and you hit the brakes, that it hits them 100% right off the bat, instead of the 50% or 80% that most people would go for.

I've not used the system in the Prius, but I've used a competing system and it's not at all surprising.

In many cases you're better off trying to steer around the collision than (or in addition to) stopping short of it. I've seen a few crashes where people could have just steered onto the shoulder or median to avoid the crash but it seems that they just applied the brakes and hoped for the best.

I don't think driver's education in the USA teaches emergency maneuvering much. How is it elsewhere?

I believe these systems don't have the effect until the point at which steering is no longer the option. The idea isn't so much to avoid the accident, but rather to try to bleed as much of the force of the accident out through braking as possible.

If you can get the brakes fully engaged you can reduce the severity of an accident pretty substantially, even if you can't fully avoid it.

My drivers education course (in the US) emphasized steering to avoid a collision first, and then braking afterwards. We were taught that colliding with a pole at 35 mph is better than doing so with oncoming traffic for a total effect of 70 mph. It also reduces the chances of injury.

After reading that pilots are the safest drivers, I've made sure to always look around my car as I drive.

Remember that F = ma.

Hitting another car of the same size coming from the opposite side when each car is doing 35 mph is pretty similar to hitting a concrete wall at 35 mph. There is no 70 mph effect. The reason the pole would be safer is that it has some give. Lower acceleration = lower force.

If someone could explain why I was voted down, that would be great. Thanks.

I didn't vote you down, but you cannot assume the cars are of identical size and mass, or indeed that it's a car at all. Hit a truck at 35, while he is also doing 35, and you'll feel a 70mph effect all right. For a few milliseconds.

Unfortunately, poles, especially old ones, do not have as much give as one would like, and large trees have basically none. Plus, their narrow aspect is a very bad thing to hit; cars are designed to collide with large, flat, wide objects and to crumple accordingly to wash off energy. Hitting an immoveable narrow object is outside this design consideration and cars have a nasty tendency to wrap around it.

I was just making the point that justifying swerving out due to some undefined 70 mph effect is incorrect if you don't know what you're dealing with. And that whoever came up this 70 vs. 35 mph effect justification should realise that it mostly applies to e.g. the truck example you gave and somewhat to cars of significantly different mass / size. Therefore it's confusing advice and obscures other possible good reasons for swerving out.

Ditto. The thing that seemed to trip up most people in the simulators was remembering to correct for the swerve. Most would turn hard right/left to avoid the obstacle and then forget to make a fast counter-turn to avoid running off the road.

Nothing taught in Australia, at least conventionally. My driving instructor was a retired police driver trainer, I benefited greatly from his experience and understanding of vehicles. Never had an accident, but avoided several via quick thinking + knowledge of how the car will react and what options I have when driving.

For what it's worth, that's the main focus of the Motorcycle Safety Foundation course. If I remember correctly, 35 mph is the speed at which is easier/faster/safer to swerve than to try to stop a two-wheeled vehicle.

It's not taught in South Africa, although we have a number of Advanced Driving Courses which teaches you anything from skid control to collision avoidance.

It is so reliable I no longer impulsively reach for my brakes. Let’s say the car in front of me speeds up after slamming on its brakes.

Hmm. That sounds pretty dangerous :(

Yeah, I could see this becoming a problem in the future. Imagine getting used to this, then for some reason, needing to drive an older car. "Oops, I didn't realize this car required manual braking".

There could be two classes of license, like in Japan where you can have a limited qualification to drive only automatics, or a full license including manuals too. In this case, however, the "automatic" would be the fully automated vehicle, and the "manual" would be the traditional human-controlled one.

why would you need a license for a fully automated vehicle?

The real reason for drivers licenses is social control. My father's first driver's license was never supposed to expire, my State ID was never supposed to expire, in both they changed the rules, and forced stores and banks to no longer accept "expired" IDs as proof of identity. It's not like your identity or age or ability to drive or anything else changes when your ID expires, they just want to keep you on their chain.

Don't ignore the departmental revenue from license renewals.

Taken out of context your quote does sound dangerous. If he were used to this and tried to drive another vehicle he may have slower reactions when braking is needed.

The acceleration in your quote is because the road ahead is now clear and he can accelerate up to his cruise speed unless he gets too close again. THe full paras:

>I pull onto a road, say freeway 280, and I set the cruise control. I set the top speed the car should ever go. Say 80 mph. But it doesn’t go 80 unless there’s no cars in front of me. Usually in Silicon Valley there’s traffic. So, the car in that case follows the car in front of me.

>But they just slammed on their brakes to avoid something. What does my car do? It slams on its brakes too. It is so reliable I no longer impulsively reach for my brakes. Let’s say the car in front of me speeds up after slamming on its brakes. My car speeds up too. It’s like there is a rope between my car and theirs. It is like nothing you’ve ever experienced.

Yeh sorry the accel section got copied over by accident.

I do think the first section highlights one of the big concerns though. Regardless of scobles excitment this tech is still in the infancy. There could be a dangerous transfer period.

I have a car equipped with a similar system, and have a similar reaction (my car will slow down to a stop if I let it).

I've had no issues thinking that rental cars were the same, because the driving experience with "standard" cruise control and radar/laser cruise control is very obviously different.

According to the Insurance Institute for Highway Safety, electronic stability control is the most important technology since seat belts.

Yet, as recently as a couple of years ago, the Volkswagen Rabbit (Golf) was the only car under $20,000 I could find with the option, other than Suzuki. As an example, the Honda Civic did not have it, and you had to buy an Accord or an SUV to get Honda's version of it.

An electronic stability system "makes the vehicle go where you're pointing it by applying the (computer-assisted) brakes. I think this is definitely a milestone." This includes driving on ice, or making tight turns (http://www.youtube.com/watch?v=j-hHWSQhKuc)

There are some really interesting risks associated with electronic stability control, though. One system used by GM on the Corvette works by comparing the input from an inertial yaw sensor to the current steering-wheel angle. Any disagreement results in selective application of braking at one or more wheels, as needed to make the two sensors agree.

Now: what happens when one of those sensors fails? http://forums.corvetteforum.com/c6-tech-performance/2336065-...

I wish I could find the video I saw earlier, where a Z06 doing about 90 MPH did a hard right turn into the wall at a drag strip for no apparent reason. I don't know with any certainty that the active-handling system was at fault in that incident, but it was enough to turn me off of the technology in a 500+ HP car.

Ford is really doing a phenomenal job recently, doing its homework and engineering to try and get the company back on track. 5 years ago, I would have said they were 10-15 years behind an equivalent Japanese car. Now I think they they are only 3-5 years.

Their recent offerings are really very impressive. I'd love to see Ford be a benchmark for American car companies as well as imports. They aren't quite there yet, but their progress is amazing.

I'm sure they are getting a lot from Mazda. The Ford Fusion, for example, is basically a rebadged Mazda 6.

I wonder how it works - Ford own Volvo, who are pretty good with safety features (e.g. the 3 point seatbelt) and who have been building blind-spot awareness systems, brake-priming systems, etc. and their concept for the next S60 had distance-to-car-in-front cruise control and pedestrian detection and so on. ( http://www.honestjohn.co.uk/carbycar/index.htm?md=1468 )

Maybe Ford are getting it from Volvo?

Maybe. Technology sharing is a big reason for a lot of these deals. For example the Chrysler 300/Dodge Charger uses a lot of Mercedes-Benz suspension components (though Daimler has since sold off their interest in Chrysler).

Incidentally, Ford has recently entered an agreement to sell Volvo to a Chinese manufacturer.

http://www.guardian.co.uk/business/2009/dec/23/volvo-sale-ch...

> Just like it was criminal for car makers to drag their feet pushing out seat belts, air bags, and anti-lock brakes in previous safety fights

Apparently Honda couldn't get consumers to pay for ABS and had to install it by default at no extra charge in order for people to have it.

I agree that is part of the issue, many customers don't care. I think there is another relavent point.

The video of the Ford guy illustrates how companies are competing on improving this tech. If it were mandated, might that be like forcing commoditization? Companies with inferior tech would push it out too soon. There would be less competition because customers would accept the minimal and it would be harder for companies to differentiate. As opposed to now: forward-looking customers with the money can look for the best system which pushes the tech forward.

And let's not forget that there are still laws on the books in most (all?) states to compel passengers to wear seat-belts.

People are very peculiar about which safety features they wil like and use.

I suppose I would have to experience first hand to really weigh in on whether I would use it or not, but from an initial reaction standpoint, I would think I'd prefer an all or nothing approach. Meaning, either fully automated driving or don't help me at all. I just can't help but imagine that I'd never be able to get used to this type of selective assistance. Someone cutting you off doesn't necessarily mean you're headed for a crash and no cars in front of you doesn't necessarily mean you can drive 80 mph without caution. If you grow to rely on a system such as this, it only takes one incident where the vehicle doesn't believe it's about to be in a crash when you really are for it to ultimately fail. Or worse yet, for the car to sense you may be in a crash and apply counter measures which causes panic by the driver and ultimately leads to a crash anyway due to overcorrection or something similar.

There's a lot of interesting automobile safety research going on in Silicon Valley. For example, Mercedes-Benz has a building in the Stanford Research Park in Palo Alto where they design next-generation safety systems. My former roommate works there, and, though I probably shouldn't give details, I can say that there are some neat safety systems under development.

One tidbit: a number of stoplights in Palo Alto and Redwood City broadcast information about when they are going to change. A properly equipped car can receive that information and decide whether to proceed as usual or begin braking, even before there is a visible change. As a corollary, if the red light is about to turn green, the car can decide not to slow down.

The slowing down for a red is nice but the not slowing down because it is going to turn green is scary. My friend died in a car crash because he timed the light and saw that it was going to be green. A mild red light runner t-boned his car and that was that. I say mild red light runner because had he not timed the green perfectly he probably would still be alive.

Don't get me wrong this is promising technology but I would be hesitant to rely on it to do too much.

One more argument is that there is usually a time gap between one light turning red and the other turning green. This gap should allow most of mild red light runners to pass through safely. (At least in my city in India this is true, the time gap is as big as 5 to 8 seconds)

The systems here have a time gap too but it was not sufficient in this case. I think it's at most 1-2 seconds here.

I understand that argument but I am of the school of thought that the machine should not make decisions that can impact my safety unless it is an emergency. shaving off a few seconds on my commute is not what I would call a much of a benefit when there is a substantial possible downside.

These types of technology are unlikely to make it out of research for quite a long time. Assist and emergency operation tech seems to make it to deployment a lot faster because of exactly these issues.

The transition between research and deployment is an important period where things that recently became possible can be fully scrutinized. The goal of research is simply to make things possible. The rest is a long hard slog that not all tech makes it through.

In short, I really wouldn't worry about anything they're working on in a research facility, that's just so far away from shipping product figuring out which parts are more or less useful probably isn't worth doing at this point.

Here's the reason I expect to hate this "feature": I already hate the way my seat belts work and they have nearly gotten me into multiple accidents. Specifically, if I see a hazard coming from the sides/rear and slow down and try to avoid them, I need to be able to turn my head quickly and sharply to see them. And the GODDAMN SEATBELT IS PREVENTING THIS because the slowdown and my quick sitting-up-straight has caused it to lock me tight against the seat. I am now PARTIALLY BLIND with respect to whatever threat is incoming and I have to hope for the best. This is terrifying and it is not a safety improvement. Based on the descriptions in this article, the situation in these cars of the future will be even worse, because the seatbelts will lock down more often and the brakes will become hypersensitive and work differently than your muscle memory has been trained to expect.

No matter how much training and experience you have with your car, as soon as you get in a situation that requires precision handling you are suddenly transported into a car that behaves and handles completely differently, that you've never driven in before. Put me down in the "no" column.

Sure, buddy. And not wearing a helmet would doubtless make you a better motorcyclist, too. Fortunately, in the case of motorcyclists, people who actually believe this tend to kill themselves earlier, sparing the rest of us from having to listen to their bullshit.

Edit: I suggest not slouching to start with.

Note that I'm still wearing the seatbelt, and I'm not pushing for the repeal of seatbelt laws. If you'll read a little more carefully, you'll see that my objection is to the mechanical change in handling, visibility, etc, just at the moment when you need it the most, not to the presence of safety measures in general.

How many years of driving experience do you have? I only ask because this sounds so much like something a new driver would say.

Your car is not a finely tuned Formula 1 Ferrari. It shouldn't be driven like one.

Humm... turning your head? Something wrong with your rear-view mirror?

Please avoid CAPS in HN. This is a civilised place here. Thanks

Adaptive cruise control has been on high end cars for over a decade - http://en.wikipedia.org/wiki/Autonomous_cruise_control_syste.... It would be good to see it standardized as it can significantly reduce traffic congestion that's caused by over braking. I think BMW claimed that this would be happen if more than 30% of cars had the feature.

"Ford’s version of radar prepares the brakes so that all you need to do is touch them to get full braking pressure if the car thinks it’s headed for a collision."

That sounds like a really good way to get hit from behind. Does the system try to stop you as fast as possible, or as slow as possible while still avoiding the car in front? The second option is kinder to cars behind without such sophisticated safety systems.

It'd be especially cool if it took both front and back radar into account. If nobody is behind you break more liberally but otherwise exercise more caution.

(It does seem like there would be room for such an "aggressiveness" parameter, since you can't predict the future breaking behavior of the car in front of you and it might have stronger or weaker breaks.)

Either way I hope all of this stuff works out to be a net win. We should know more in a year or two!

Wouldn't the ideal step be to harmonize the radar/lidar technology and have a system enabled that forces a safe driving zone. It's all too common on highways that vehicles can't stop within a safe distance if the vehicle in front of it stops, which leads to 5 vehicle pile ups.

You'd only need a moderate percentage of vehicles with the radar/lidar technology to keep a highway reasonably spaced. EG if you're tailgating, your vehicle will brake, if someone is tailgating you it will speed you up (within the speed limit) to get you to a safe distance, or (when capable) shift you safely into a different lane when you're in cruise.

Eventually you could simply have vehicles communicate, so when the guy in front of me brakes, my vehicle gets instructed that it's approaching a decelerating object. It'll be small 'safety' steps that will eventually remove human error from vehicle accidents.

I don't think increasing your speed will avoid tailgaters, rather it (in my experience) makes them go all the faster, to continue with thier tailgating ways.

I just saw a Top Gear episode where they tested the last Mercedes S which does just this. (talk about serendipity and having new ideas!)

The goal is to minimize the force of a collision not to stop 10 feet behind the car in front of you. As a collision avoidance system, if you can stop with a reasonable distance from the car it's not going to start and if the other car is not stopping as fast as expected then it can relax some.

"Think about that. If we could cut down on car crashes by even 5% we could save more than 2,000 lives!"

I don't know whether or not this is true, but the argument is fallacious. Without further evidence, it is a bit premature to assume that the fatal:nonfatal ratio in crashes prevented will be the same as that of crashes in general. It is possible, though unlikely, that a chance might only prevent nonfatal crashes.

Seems that I've read about studies that show that people have a certain, nearly unconscious risk tolerance. For example, ABS and all-wheel-drive made driving in the snow and ice easier, so people drive a bit faster or more aggressively until they reach their previous risk tolerance level. Thus the potential safety benefit is not realized in actuality. In fact, if people overestimate the added safety of the new technology, they may be more likely to have an accident than without it. Witness the common sight of four-wheel-drive vehicles spun off the road in a snowstorm.

Looking at the data, the steady decline in deaths per passageway mile suggests improving safety actually works. Now, it may be useful to make people feel less safe, but IMO most accidents are caused by people not really paying attention and automated systems which tell them something bad happens should help as long as they alert the driver that something bad almost happened.

PS: The fact you can take little damage while rear ending someone at 65MPH is not something the average person wants to test out.

Agreed, my dad always described 4 Wheel Drive as a way for people to get stranded farther from help than they otherwise would have.

I agree with this. I am far more comfortable driving faster with 4 wheel drive on, so I disable it as soon as I don't need it.

They used could save not will save for that vary same reason. Most likely the net effect would be much like seat belts which transform many fatal accidents into non fatal accidents and many non fatal accidents into non accidents.

PS: Unlike airbags seat belts do help avoid accidents. But, it's rare to see people talk about that outside of drivers education.

Just to put in context, a link that was here last week or so about the Automation Paradox:

http://spectrum.ieee.org/computing/software/automated-to-dea...

Will there be a compromise between traditional crash deaths and safety systems induced deaths ?

Not to dismiss the new systems, I'm an optimistic by nature.

How is it not still a net win? The automated systems fail a lot less often than humans. Even if they can't get the failure rate to 0%, it is still making things safer to replace humans by robots.

GPS-enabled Governors should be put in all cars too, limiting your speed to that of the speed limit, adjusted for weather conditions. If you're in a 35 MPH zone, you car will stop accelerating (and ideally auto-slow when going downhill or onto slower roads) once you hit the limit for that part of the road.

Speed limits would likely have to be re-adjusted in some cases as they may be too low in some area's, but it would certainly cut down on accidents.

An exception would be allowing people to speed in short bursts. While rare, sometimes you need to accelerate to avoid an accident. Not allowing for this would be unfair.

Definitely wouldn't be popular with a lot of people, or police departments generating revenue from speeding tickets. But driving is simply a means of getting from A to B in a safe and efficient manner. So long as people are getting hurt or killed in between A and B, something is very broken.

How about we put a GPS-enabled governor around your neck, in case you ever think of stepping on private property.

Different context which doesn't apply to my initial post, and worded in a harsh manner to attempt to demonize my idea without presenting an argument against its merits (i.e. "around your neck"), but I'll bite.

Stepping on an individual's property doesn't endanger others' lives needlessly. It's pretty easy to call the police and have a trespasser removed/arrested. Plus, in some states, you can shoot and kill trespassers so long as you've given them fair warning, and can legally carry a firearm. People also have attack dogs, surveillance systems, etc...

You can't shoot and kill someone who's speeding endangering your life. If you call the police about a speeder, they'll usually look for them if they're not too busy. But police departments would quickly be overwhelmed if they had to police "every" speeder on the road. There's simply not enough police. Also, you're defenseless while being in the presence of their illegal driving behavior.

Trespassing, on the other hand, is a rare occurrence. Penalties tend to be harsher than speeding (jail time rather than a fine). It must be much easier to speed than trespass since I don't know anyone I've met in the last year who hasn't sped, yet I can't pick a single trespasser out of any of them. I'd imagine this is the same for most people.

People have little incentive to trespass too. Why do it? To steal something? Murder? Most people don't need to steal things or kill others.

Speeding, on the other hand, has a number of incentives tied to it which provide immediate benefit without significant perceived risk (tickets are rare - habitual speeders may only receive a couple a year | accidents are even rarer). For one, it might save you time^. It also makes people feel good (release of aggression, showing off, feeling of freedom, etc...).

My initial idea isn't one of punishment, but one which increases safety. What's the goal of driving? Going from A to B in a safe, efficient manner. Why wouldn't you want it to be safer, within reason? So you can cut 4 minutes off your 60 minutes of travel time?

^This isn't always the case. Roadways are much like assembly lines. Their slowest points set the pace of everything on the line. Speeders mostly zip from one bottleneck to the next, never really decreasing their net travel time. Time yourself driving the speed limit to work and back one day, and time yourself speeding the next. How much time did you save? A minute or two? None at all?

NOTE: Most vehicles already have governors which limit their top speeds, though aren't used in the manner which I suggest: http://en.wikipedia.org/wiki/Governor_(device)

I demonized your idea not because it's impractical, but because it's evil. It's evil because the idea is designed merely to make a more optimal society. It's evil because it's designed to enforce the law by preemptively making lawbreaking impossible. It's evil because it forces carmakers to make their machines in a certain way, to achieve this petty goal.

This proposal wouldn't increase safety until all cars on the road had the GPS-governor feature, which would take 15-20 years for old cars to be phased out. In the meantime, the new cars would have an annoyance (no speeding) without providing any measurable safety improvement.

You're correct on some points, though it may slow other cars by way of simply being in the way. It wouldn't work on highways with passing lanes until a high adoption rate, but a single slow car on one-lane side roads will "bottleneck" any speeders. This alone would most likely increase safety.

Gradually as more people got the cars, things would get safer and safer. Once, say, 90% of cars on the road have governors, it'll be hard for a speeder to find the opportunity to speed.

Economic reasons could speed up adoption too. For example, insurance carriers would likely charge less for those with the device, and more for those without it. Governments could give those with the device a break on their income taxes.

But honestly, I don't think the idea has much traction. It's too emotionally charged which would make it difficult for politicians to implement it, and there are large switching costs from current cars. Self-driving cars will likely appear beforehand. Radar and governors and what-not are really all just bandaids for the real problem with cars; their human drivers.

It's not clear to me that slowing traffic will increase safety in any meaningful way. Slowing traffic will, for any constant amount of "demand", necessarily increase traffic density. Increased traffic density would seem likely to have a positive correlation with accident rates that would partially offset the decrease you're hoping to see.

During the transition period, I predict you'd see a marked increase in the accident rate due to the increased instances of overtaking and broader spread of speeds. (12 cars all travelling at 80 mph on an interstate is safer than 6 cars travelling at 60 mph and 6 cars travelling at 80 mph. When you figure that each car governed to 60 will have to drive 33% longer, you'd actually have 8 cars at 60 mph and 6 cars at 80 mph to get the same throughput to destination on the highway.)

Interesting point. I know broadly that lowering speed limits tends to reduce accident rates, but that's certainly different than this (and may just be spin I've heard through the media).

The only way to tell would be to run something like this through a traffic simulator, and if that found things to be safer, next study it in a sample state or country (would be difficult to do). So you're possibly right, pending results of an imaginary study. :)

Besides radar I think there's a lot of potential for automated systems to help drivers cope with snow & ice. For example a sensor scanning the road able to warn the driver of black ice, or at least inform the driver that they may be seeing rain on their windshield but it's freezing on the ground. Maybe speed controls based on real time weather conditions. No one should be going over 40MPH on any road in the snow. How about using cars as a way to measure road safety? Monitor the road conditions and the cars traction and report this information to the DOT so a plow/sander can be dispatched to particularly bad areas.

Interesting that it doesn't always perform in severe weather conditions as this seems like it could be when it was most useful, eg avoiding a pile-up in a whiteout. He mentions that you shouldn't be driving in such conditions but unfortunately Michigan is famous for rapid and surprising weather changes. Other than that it seems like very cool technology that hopefully becomes standard soon. It's nice to see that car manufacturers are willing to take things slow to avoid rejection because they think it is important as well.

The assisted braking 'feature' where the system moves your calipers closer to your rotors sounds really stupid and dangerous. When I press the brakes, I am expecting a buffer period between initial contact with my pedal and the actual deceleration of the vehicle. Because of that buffer, I tend to push my pedal down further to initiate braking. Eliminating that buffer with drivers who have come to expect it will make for very sudden braking, which would probably increase the chance of being hit from behind.

Presumably the ABS still works, meaning you won't lose control of the car while slamming on the breaks.

As far as being hit from behind, I suppose this is (and always has been) an issue in accidents and other emergency situations. Still, the car behind you should allow appropriate spacing to prevent rear-ending you. I don't see how emergency breaking systems change really change the basic rules of the roads.

I don't buy it. If you need a radar to tell you you're about to crash, you're probably not paying attention properly.

We should tighten up driving tests. Driving shouldn't be a right, it should be something only competent attentive people can do.

It'd be interesting to know stats on Manual vs Automatic car accident rates. I'd guess that Automatics are involved in more accidents since the drivers are probably paying less attention. Similarly cruise control.

It's not that you need it to tell you that, it's that the human brain, even when focused, is relatively slow to respond. Not to mention many people have a panic impulse that causes them to yell first and then step on the brakes. In certain situations where a half second can mean the difference between death and a bruised knee, it's pretty impressive stuff.

That may be, but I still look with suspicion on someone who says he would have gotten in 2 accidents in 11,000 miles had his car not saved him. In my experience, it's very, very rare for someone to just cut you off with no warning. I can almost always predict that someone is going to switch lanes without signaling; usually 2 or 3 seconds before they do it.

Predicting un-signaled lane changes is almost a required survival skill on a motorcycle.

No doubt. He probably Twitters while driving.

This is a very good reason I prefer driving in the UK where lanes are only used for overtaking, and overtaking on the left is illegal.

Good for you, but how about the other 95% of the population who are less attentive than you and lack predictive powers?

You think 95% of the population gets in 2 accidents a year? Scoble is either overstating his case, or he's a wretchedly bad driver.

They shouldn't be allowed a license.

Sure but that's why you pay attention and drive properly, and look out for any dangers/issues.

The best way to cut down the number of accidents is to get rid of all the crazy bad drivers, rather than any sort of technology IMHO.

It's not even remotely economical to eliminate bad drivers, not to mention good drivers who driver badly at times. My sense is that during any test, people will appear much better drivers than they are on average, since they won't be drunk or shooting off an SMS message or whatever else.

It is economical to save a lot of lives with technology. This is a good tech. I don't know if I'd go with most important since seat belts, but it could save lives.

Agreed. Even assuming people do the right thing all the time and never get behind the wheel after drinking, when emotionally heated, without enough sleep, or while distracted by phones, etc., anyone can be distracted for a split-second and cause a major accident.

That kind of "Superman" thinking that things would be great if only you tried harder is BS in a lot of cases. Most people do not want to get into road accidents, deliver code late, or build a shoddy product. Fixing the incentives or process works better.

It's an incentive for people to pay less attention when driving. It's a bad incentive IMHO.

You could make the exact same silly argument about seat belts, crumple zones, and air bags, but the fact is all 3 save many lives.

Seat belts, crumple zones, air bags don't claim to prevent accidents. People aren't likely to say "Ah well, it doesn't matter I don't need to pay attention, since I have an air bag".

Radar etc though does claim to prevent accidents. Which will likely mean people pay less attention.

Modern cars make so much power that you really don't need to shift unless you're accelerating from a stop, in traffic, or on curvy/hilly backroads. I can comfortably drive in 6th gear between 55 and 100+ mph including passing, going up hills, and so on. Obviously it's not the ideal gear in all of those cases, but if you're feeling lazy it works.

I would guess manuals are involved in more accidents and more fatal accidents, at least in the US, since the only people who buy them here in great numbers are performance enthusiasts, frugal bastards, and diesel truck drivers. Some enthusiasts are responsible and don't push their cars on public roads, but many aren't. If there is a difference, it's probably small. My parents have two of the same car, one in manual and one in automatic, and they never mentioned a large difference in insurance cost. Actuaries would charge manual and auto drivers different prices if their risk rates differed.

If somebody cuts you off, the radar is guaranteed to have a faster reaction time than you.

The point someone made up thread, is that unlike the radar, if you are paying attention, you will likely notice them approaching to cut you off which the radar certainly cannot.

"Another place the radar is invaluable? In fog. I drive over the Santa Cruz mountains every day to get home and there often is fog. One day there was a car in front of me that had no back taillights."

I bet the driver of that car was just using parking lights. I hate that! People who drive with parking lights on don't seem to realize the rear running lights are not on, parking lights are not meant to be used when driving that's what headlights are for.

Daytime running lights are mandatory in Canada. Not sure why they aren't in the USA. If the car is out of park, the headlights turn on automatically. It solves a handful of issues like this that are potentially dangerous.

It specifically does not solve the "no rear lamps" issue. In fact, it exacerbates it.

Not true. Every car in Canada engages headlights and rear tail lights when the car is out of park, regardless of the headlight settings. The only thing effectively controlled by the headlight knob outside of park is the dashboard lights.

Interesting; I did not know that. In the US, every DRL that I've seen lights only one pair of front lamps (usually the headlamps at partial power, but some models light other front lamps).

I've never seen a car where the "first click" of lights (what I presume you're calling parking lights) doesn't turn on the taillights. In most cars, they're even on the same fuse.

(Yes, there are cars with true one-sided parking lamps that will only illuminate one side front-and-rear to mark the edge of the car parked closest to the road.)

I suspect what you're actually seeing is drivers using only the daytime running lights. Those do not turn the rear lamps on.

What worries me about this technology are the corner cases. If it's 99.9% effective, eliminating the need for manual braking, the 0.1% of the time where it doesn't work, the user won't be ready to brake, which is dangerous.

I think the point is that rarely would a human succeed in such a corner case, let alone in 99% of all cases.

Do ya think it will have the unintended consequences of making people take more risks then they would have if they know of safety features?

Or maybe the increased safety outweigh the cost of false sense of safety?

BMW has a system in their top-end saloons that will safely pull your car over to the side of the road without crashing if the driver is suddenly incapacitated (e.g. a heart attack). These system are great and I hope we can overcome the social barriers eventually. Car accidents are still the #1 cause of preventable death worldwide.

EDIT: Here's another great system expected to start trials in 2011: http://nextbigfuture.com/2009/11/road-trains-and-modular-veh...

If you have a vehicle fitted with this system you'll be able to join 'road trains' where the front vehicle is driven by a professional driver (e.g. truck or bus driver) and the rest of the cars follow behind it - driven automatically.

Car accidents are still the #1 cause of preventable death worldwide.

That strikes me as unlikely, given that car accidents kill something like 50,000 people a year in the US. Just inside the US, an obvious candidate for prevention -- medical mistakes -- kills far more people. (Estimated at 200k per year preventable deaths in hospitals alone.)

Casting our glance worldwide, poorer nations have millions of deaths from causes that aren't even "hard problems", such as clean drinking water, food distribution (solved in Europe in about the 1700s), malaria (solved in America over half a century ago), etc, etc.

(Estimated at 200k per year preventable deaths in hospitals alone.)

For the United States ALONE? I don't buy it. But prove me wrong by citing an authoritative source if there is one.

Citing an authoritative source doesn't prove you wrong any more than it proves global warming is happening ;)

Be that as it may, see generally To Err Is Human by the Institute of Medicine (an arm of the National Science Foundation), which put the tally at 100k back in 1999, or Dr. Chunliu Zhan and Dr. Marlene R. Miller in a research study published in the Journal of the American Medical Association (JAMA) in October of 2003.

Thanks for the source. That was easy to Google with the clear information you provided. No, several of those causes of death would not fit the usual definition of "preventable deaths." Those patients may not have received the best care that they should have received, but it's overextrapolating to say that they died but for errors in treatment.

The Zhan & Miller study in 2003 supported the IOM's conclusion of ~98k deaths due to medical errors. The 200k number you quoted earlier was "found" by a study by some random website and was not published in any medical journal.

Not to downplay the severity of the issue, but there is a big difference between 100k and 200k lives.

World stats (causes of death) according to WHO:

1) Coronary heart disease - 12.2%

2) Stroke and other cerebrovascular diseases - 9.7%

3) Lower respiratory infections - 7.1%

4) Chronic obstructive pulmonary disease - 5.1%

5) Diarrhoeal diseases - 3.7%

6) HIV/AIDS - 3.5%

7) Tuberculosis - 2.5%

8) Trachea, bronchus, lung cancers - 2.3%

9) Road traffic accidents - 2.2%

Although maybe it's more correct to say the leading cause of accidental death is car crashes. I guess 'preventable' is too broad a word.

Yes. In the public health course I took last semester, road traffic accidents were the most common form of death by injury, if I remember correctly. Turns out, tobacco use is the leading cause of preventable death worldwide, a label applied to diseases that can be avoided through the use of vaccination or healthy behavior.

"Preventable" has a fairly specific meaning in the public health community, as opposed to its common-use dictionary definition. Obviously, the latter isn't very useful since most ailments are "preventable" if you think broadly enough.

It appears that the leading cause of death for folks under the age of 35 is auto accidents (the bulk of 'unintentional injuries' if I understand the data correctly).

http://www.chdl.org/Projects/10%20Leading%20Causes%20of%20De...

N.B. saloon == sedan. (As with many of the Britishisms in my brain, I learned 'saloon' from The Economist.)

What, you don't watch Top Gear?

Though more entertaining than educational, I know the difference between my boot from my bonnet thanks to that show, as well as what a saloon, spanner, pillock and many other british-english mannerisms mean.

And here I was searching Google Maps for a classy place to drink and brawl.

how does that work if you're not right next to the shoulder?

my guess is lasers, they already check to alert you when you have cars in your blind spot, so it most likely checks those before moving you over.

GPS combined with various road / traffic sensors. VW has a system that can drive itself all the way to a hospital (or anywhere really) in city traffic. They're going to trial some of these systems in Germany this year. Because of more stringent liability laws in the US you won't see it in America any time soon though.

uh...then why are the darpa challenge guys having so much trouble? surely darpa would just give the money to one of the big car manufacturers if this was a solved problem?

The DARPA challenge was five years ago! This is the result of advances made in those challenges.

The Stanford guys have advanced so much that they'll take it on a rally track this year: http://www.youtube.com/watch?v=-CuVVZq9GfY

erm...the urban challenge was 2007.

The first DARPA challenge was in 2004. In 2005 and 2007 most of the teams easily completed the challenge so the problem is definitely 'solved' now. There's no more prize money anymore. Its over! The tech is here now.

I've had the radar-assist cruise control for two years (complete with braking) and everything, and I view it as a first step towards automated driving.

People aren't likely to accept automated driving (even if it's statistically safer than the average driver) but they will pretty readily accept driving aids that watch the car in front of you, and watch if you're in your lane, and what not. I expect those "aids" to continue to advance and expand with them being "aids" for a long time, until liability problems are solved.

And FWIW, I've had no issue when driving rental cars that don't have the radar/laser based cruise control. The experience of regular cruise control and radar/laser cruise control is sufficiently different that there's really no way to mix them up.

On long stretches of Highway in the American West, there really isn't a whole lot of driving to do except keep the car in the lane. It's pretty clear that we're getting painfully close to fully automated highway driving.

I wouldn't describe fully automated highway driving as painful in the least. I'd be quite happy about it.

After all, highway driving isn't remotely interesting from a "driving" viewpoint, so I'd hardly miss it.

"Painfully" close as in very close. I don't think he meant to say that automated highway driving was painful.

Right, that's what I meant, apologies for lack of clarity.

As usual, Scoble misses the boat by several magnitudes in his aim to over-sensationalize something he read on techmeme. I would personally put safety glass, airbags, anti-lock brakes, traction control, and crumple zones ahead of radar. Those are technologies that improve safety for everybody, whereas radar is for people who don't pay attention and should not be driving in the first place.

whereas radar is for people who don't pay attention and should not be driving in the first place.

So, "for people", then. I'm confident that there is no driver who is fully paying attention to driving at every moment while driving.

Radar is not for the people who don't pay attention. It's for the people who are driving in front of them who are going to get rear-ended otherwise. The less control idiots have over cars, the better. I would feel far more comfortable dealing with a bunch of automated vehicles than a similar number of typical drivers, to be frank.

Radar will just be another crutch for a bad driver. Good drivers make sure that they always have an out, check their mirrors and blind spots as often as possible, stay safely away from other drivers, and do their best to control the situation. Adept use of brakes and speed can control how far people stay from you, as well as your position within a lane. These are skills every motorcyclist learns very, very quickly.

I use my seatbelt as a crutch to lean on.

Those careless people do drive and radar might stop them from causing an accident that kills your friends or family one day. Radar also has a huge potential to improve the safety of pedestrians who otherwise might briefly admire the traction control system of a careless driver's car before they are crushed to death.

Reply on news.ycombinator.com