It's not just driver errors. Apparently the crash detection system kicked in during the Berlin ISIS attack and stopped the truck before the attacker could kill more people

(sorry, link in German) http://www.spiegel.de/politik/deutschland/anschlag-von-berli...

That's cool.

This seems to be a fresh news story but I hope it will be more widely disseminated. Among so much aimless hand-wringing over terrorism, security theater "designed to stop it", and dragnet surveillance and intelligence gathering, here is a clear example of a new technology reducing the impact of a terrorist act and saving lives.

It kinda goes against the click-generating narrative of autonomous driving technology being half-baked gimmicks rushed to market, fatal accidents waiting to happen, and Tesla and its celebrity corporation ilk flying too close to the sun on homemade wings.

Saving any lives in great, but trucks are a very small part of the problem.

Out of 33883 US traffic deaths in 2009, 3380 involved trucks, and truck driver error was listed as a cause 26% of the time.

Self driving long-haul trucks probably won't save "thousands" of lives, but certainly hundreds.

http://www.trucking.org/ATA%20Docs/News%20and%20Information/...

The 26% statistic is misleading in this case-- just because the death didn't involve truck driver "error" doesn't mean a different movement pattern for the truck wouldn't have made a difference.

For example, in the original video, if the Tesla autopilot hadn't braked, and the car had been involved in the crash, it's not clear to me that error on the part of the Tesla driver would have been listed as a cause.

It should be, though. Car #1 stopped shorted and car #2 hit it. Car #2 was at fault for hitting Car #1.

Car #3 (Tesla), if it didn't stop, would be at fault for not maintaining stopping distance, hitting Car #2.

You're implying that minimum following distance <= stopping distance. A brief Google search yields a stopping distance of 315 feet for travel speed of 70 mph.[0] We don't train drivers to leave a football field of space in between cars traveling at highway speed. At least where I live, the legal standard for minimum following distance is that a driver "shall not follow another vehicle more closely than is reasonable and prudent".

If Driver #3 can't see Car #1 past Car #2, and Car #2 is traveling at highway speed, I think it's a stretch to say that Driver #3 could be faulted.

[0] http://www.driveandstayalive.com/stopping-distances/

Yeah, for sure. I was implying that the other 74% will be avoided by cars with self-driving features, since the parent comment was suggesting that getting driverless trucks was more important.

It's not more important. Automating trucks is easier, and the economics is all on the side of automation.

Automating cars will take much longer than trucks.

Exactly. There's a lot more economic impetus to automating trucks. It'll happen faster there than in cars, by a long shot.

Saving a single life is worth the tech and disruption it will cause. That is one family that won't be missing a father, a mother, or a child.

Any saved is a win in my book, though point taken.

Autopilot would take over trucking even if it was no safer than humans under the current safety regime applied to human drivers, simply because it could drive nearly 24/7 without degradation, improving asset utilization.

I think you underestimate the ability of the trucking industry to poorly maintain automatic pilot systems.

The trucking "industry" is probably pretty okay with losing drivers - its the only class of employees affected and they've had a shortage of qualified drivers for years [1]. Heck there will be a whole new class of higher paid mechanics and technicians maintaining the auto-pilot systems. There will also undoubtedly be a higher bar government mandated maintenance for self driving trucks.

They're almost incentivized for this change.

[1] http://cerasis.com/2016/05/03/driver-shortage/

Poorly maintained autopilots would still be safer than poorly maintained drivers.

But at least if it's a poorly maintained driver the company can blame the driver and not change their practices.

They can still use the owner-operator model if they want to, and have the vehicles owned by a different entity than the freight company.

But in all seriousness, although I understand that people like to out-cynic each other, over the road trucks seem to be pretty well-maintained. Whether that's the result of safety regulations and enforcement or simple economics (a broken-down truck is not a moneymaking truck) is left as an exercise to the reader, but either way you do not see a lot of barely-making-it or broken-down commercial OTR trucks on the road. That level of poor maintenance is generally the province of private vehicles.

I'd hope a poorly maintained automated truck which crashes would result in extremely heavy fines and lawsuits, to the point where it could seriously threaten to put a trucking company out of business. Make it uneconomical not to maintain the trucks.

Drivers maintain themselves, so this is pretty much purely maintained drivers vs. little-to-not maintained autopilots.

That said, I agree there's plenty of incentives for the trucking companies to go the autopilot route.

>Drivers maintain themselves

Fatigue and sleep deprivation are chronic problems in the trucking industry. Clearly, many drivers are unable or unwilling to maintain themselves.

https://www.nytimes.com/2014/06/17/business/truckers-resist-...

http://www.huffingtonpost.com/entry/congress-made-trucking-d...

I agree. My point being, trucking companies don't have to maintain drivers, they push this onto drivers themselves. So if they simply s/driver/autopilot/, they'll have to start maintaining.

No way. I use semis as a highway safety feature. They can see farther down the road and are much less likely to get in an accident statistically. When visibility is poor, I follow one (with a large follow distance).

> (with a large follow distance)

Good. Some things to keep in mind: Semis can clear obstacles you probably can't - and at least around here are trained to not break because of some road hazards. Semis are not manuverable, and if they lose a tire the last thing they want to be doing is applying a now-uneven breaking force.

My car, with myself and a friend inside, almost got ran over by a truck when the driver didn't even bother to slow down for the yellow light and crossed at high speed the red light. Luckily he didn't hit anyone.

I totally agree with you.

I don't know your situation, but my experience driving big heavy stuff is that sometimes a yellow light does not provide enough warning to stop 55mph * crazy weight, and you sometimes find yourself with a dilemma of entering an intersection just after a light has changed, or a few seconds after the light has changed.

It sounds to me like the driver shouldn't be going 55mph in the first place, if they can't stop before the yellow light turns red.

It's similar to icy roads. If you go the normal speed limit and you slide into another car, it's your fault. The fact that the speed limit was higher than your speed doesn't release the driver from responsibility.

That part's not really up to the driver, though.

It's the responsibility of the traffic engineers to make sure the yellow light is long enough to accommodate reasonable braking time and distance for all of the traffic that can be on the road, travelling at the posted speed limit. They can lower the speed limit, increase the yellow light duration, or put up a sign limiting the max vehicle weight.

It's not reasonable to expect every driver to know detailed information like yellow light times for every intersection and section of road.

In this particular road and crossing there is a sign with flashing lights indicating "PREPARE TO STOP WHEN FLASHING". It was flashing while the crossing light was still green to indicate that it would turn yellow pretty soon, but the driver still didn't slow down.

Unfortunately there is also a profit motive in yellow light times [1]. This can cloud the decision-making significantly. [1]http://www.chicagotribune.com/news/watchdog/redlight/ct-yell...

In my state, the minimum yellow light length is 4 seconds, but it's obviously not uniformly applied. You can get a device to time the yellow lights from the state if you get a ticket, and then use that info to dispute it. Not much help if there is an accident.

It's a big enough problem that there's a group trying to do something about it.

http://shortyellowlights.com/standards/

EDIT: according to the link, the yellow light for a 55 mph road should last 6 seconds.

Icy conditions are maybe predictable (ok, it's under 0°C outside), but yellow light times (edit: the time between the light turning yellow and the light turning red) are not.

Yellow light turns are predictable in that you know that they'll happen, just not when. So if you have a heavy load or other adverse conditions you should slow down to a speed where you could stop at the correct distance. Software should do an even better job of predicting the weight to speed to stopping distance ratios.

You don't know how long the yellow light will last before turning red. That time is variable from light to light and essentially unpredictable to drivers.

What is a safe speed to drive when you take into account that the yellow light may last from less than a second to 20 seconds?

but the problem is you don't know how long an yellow light will be. They are all different. State to state. Town to town. Hour to hour. No standardization.

Chicago was caught making their yellow lights purposely short on some intersections to make more money.

Right. I actually wonder how self-driving Semis will handle such a situation.

Given that it's all math at that point, I wouldn't doubt if regulations surrounding minimum yellow-light timing are updated to more accurately represent the stop time of heavy loads. Otherwise it should be easy to prove in court that stopping in time would be impossible given the speed limit & weight restrictions (or lack thereof) in place at a given intersection.

You shouldn't be driving a vehicle through a non-freeway (i.e. a road with stop lights) at a speed where you can't stop in time. In many jurisdictions that is considered reckless driving. If your vehicle is too heavy then you (or the software) needs to slow it down.

I sincerely hope there will be such a technocratic solution to this problem. Recall that there was a trend in the late '00s where cities shortened the yellow light time so they could reap greater fines from red light cameras. Having a definitive length of time that yellow lights 'should' be will be helpful.

This road is primarily used by heavy trucks. There are signs with flashing lights indicating "PREPARE TO STOP" before the crossing exactly for that reason. That driver sure knew what he was doing.

Some were arguing yesterday that there'd be a strong negative emotional response when a driver-less car causes someone to die. I hope there will be an equal and opposite reaction when people are saved by this technology.

Not the least because truck autopilot don't play videogames while they drive!

A too common practice in Europe is to drive by ear. The truck driver will not watch the road unless the wheels are running over the side lines which makes a noise.