Very interesting piece, Junko. I wonder what automotive seat belt systems engineers are working on: the car won't move until all the passengers where seat beats? Or the seatbelts automatically latch themselves?
The numbers suggest that only 46% of the 32,000 deaths could be saved, but almost all crashes are avoidable, and each fatal crash avoided by a self-driving car not making the mistake a person would have means fewer deaths whether the occupants were wearing seat belts or not.
The way these numbers were presented is a bit confusing... But yes, basically, self-driving cars, if functioning correctly as assumed by carmakers, could have saved some lives among those who got killed even though they had their seat belts on...
I've been in cars where the top of sholuder belt was attached to the door and in other cars where the top of the shoulder belt was on a track and autonatically moved forward when the door was opened and moved bak when the car was started.
This technology is at least 10 years old but I haven't seen that it's increasing in prevalence. Of course, that may be because I'm driving a 14 year old car...
A system that tries to make you wear a seatbelt even if you hate it cannot work: you'll always be able to cheat the system (lock the belt behind you, for example).
In order to make me wear my seatbelt when I was 10, my father stopped the car as abbruptly as the breaks and tires would allow from a speed of 20km/h (~10mph). That made me fall on the floor... and made me WANT to put it on.
I'm not sure how well a self driving car would be received if its software allowed it to do that, though...
It is amazing the # of Fatalities has dropped from over 55,000 to about 35,000 in the last 35 years. Both my parents had injuries from driving or riding in cars with no seat belts(the standard at the time)
The idea of the car not starting till all belts are latched is quite nice. If it fails, no harm is done. The idea of it not letting you unlatch till you turn off the engine is scary though. If that fails, there's a possibility of someone being stuck in a car when they would need to get out.
The less required of the driver, the more distracted they will be, and the less capable they will become should a problem arise. If seat belts are the only answer to the safety situation, then vehicle speeds will have to be limited to 7.5 MPH, as the human body is not designed to handle impacts at speeds greater than 15 MPH.
I know that planes and busses are safer than automobiles, but I still prefer my car for travel because I feel that I have at least some degree of control of the situation should a problem arise.
I feel differently I guess. There are many times I would love to be able to hand off the driving to the car and just enjoy the ride. Unfortunately, it would ultimately result in me just sleeping for the entirety of most of my trips.
Seat-belt discussions aside, I think the point the Google Car safety director is making here is that no safety measures could become a perfect solution for zero fatality. And that, by the way, includes self-driving cars.
There are some odd ambiguities in the way people evaluate risk. The same people that would not trust self-driving cars may cut people off in traffic or change lanes without checking the blind spots. No question that there are low-hanging fruit like seat belts that should be an easy choice, and also no question that self-driving cars will not have a perfect safety record, but both of these can reduce risk by significant amounts. Logic says that people should wear seat belts, but many do not. I also believe that self-driving cars can do a better job than people do. Add the two together and the safety margin increases substantially.
I'm glad to finally hear someone say that no safety system is 100%, and it is concerning that anyone would play the "zero fatality" card. As mentioned, the first fatality in an autonomous vehicle will probably set the technology back 10 years -- even if the overall safety record of autonomous vehicles is far superior to human-controlled vehicles.
If anyone doubts the negative impact that the first autonomous vehicle fatality will have, just consider the viral response we witnessed to a video of an EV getting burned to the ground, despite the fact that this was a very rare if not unique event, and despite the fact that a great many gasoline combustion engine vehicles catch fire every single day.
@Junko and Frank: Agreed. There is a pr nightmare waiting here of the distopian future of automated crashes very like the burning Tesla. But the reality as we know it is automated systems and EVs willbe safer than distracted human drivers and gas cars.
How do we communicate that to the mainstream media and avergae driver?
"How do we communicate that to the mainstream media and avergae driver?"
A tough call, but not because it should be. Even on this august site, there are many who seem to believe that driving a car is the ultimate impossible task for automation. And yet, if it's so difficult, how come so many barely competent (or for that matter mostly irresponsible) people seem capable of it?
But then, Rick, when I look into some software bugs and electronics glitches that might have caused Toyota's electronic throttle control to play a role in unintended acceleration ( see my other story http://www.eetimes.com/document.asp?doc_id=1319897&), i can't help but wonder how safe those software-packed autonomous cars will really become.
"Have you ever wondered: Who, really, needs a self-driving car?"
I see the answer on roads every day, Junko. Those people who are much busier on their cell phones, either talking or texting, and obviously not very engaged in their driving chore. Women who have the rearview mirror turned to put on makeup, again with this pesky driving nuisance always getting in the way. Drunks. Drivers who feel the need to move at 25 mph in a 35 mph road. Drivers who seem to go catatonic at stop lights, wasting a good chunk of the green cycle to snap out of their stupor.
Of course, the bit about safety never being 100 percent verges on banality. That's true no matter what you do, or even if you're at home!
My prediction is that many years from now, people will marvel at how ANYONE could ever have been allowed to drive themselves. How reckless!
No doubt: a lot of people should not be driving, so having an option for them to be driven around by a better driver (a self-driving car or a chauffeur) is the best solution for the fully functional drivers with whom they share the road. (Think about a self-driving taxi -- it could put a whole group of people out of work, too.) That is, assuming the self-driving car will be a better driver.
Can the consumer trust the auto industry to create a good self-driving car? The level of distrust out there considering the auto ndustry's past reputation in the U.S., and current Toyota ligiation where software flaws may have caused deaths, I don't think a well-functioning, experienced, safe, and alert driver would be readily give up the control of the car to a complex system embedded system until they know the system is close to flawless.
"Can the consumer trust the auto industry to create a good self-driving car?"
The consumer trusts automakers to make wheels that don't fly apart, no? Or engines that don't seize up. Or automatic transmissions that work essentially flawlessly (certainly flawlessly compared with the typically incompetent way drivers drive stick shifts!). Consumers trust ABS, stability control, automatically operating lights, spark advance, choke, fuel-air mixture, catalytic converters, air bags, and a host of other automated controls in cars, that most consumers aren't even aware of anymore.
What makes driving so different? Virtually all of the automatic systems I listed were not always automatic. But no one questions anymore that the automatic version far surpasses what humans were capable of doing manually. A catalytic converter, for instance, would never survive the manual controls of a Model T Ford.
By the way, when air bags were being introduced, their main selling point was that occupants would be protected even if they didn't bother to buckle up. They were introduced essentially as the self-buckling seat belt.
In the speech made by Google's Medford, he made it clear that the purpose of the self-driving cars, in his opinion, is "freedom." People are now being freed from the complex task of driving.
Where the disconnect occurrs is when the auto industry pitches autonomous cars to the authorities who must approve that self-driving cars can run on the same roads with other cars, carmakers always use "safety" as their pitch.
"Zero fatlities" is a mantra that sounds good, and everyone uses it. And it almost legimizes self-driving cars. But the interesting part of Medford's speech was this. To paraphrase it: don't expect zero fatality just becuasae it is a self-driving car. And if it gets into an accident, don't blame the self-driving technology for it, because even what is supposed to be recognized as the most effective automotive safety measure -- seat belts -- is not 100 percent effective.
I think that truly autonomous driving should probably not occur on the same roads as manual driving. Driver assistance, sure. Or if autonomous driving does occur on regular roads, then the margins for error, stuff like distance between vehicles, should be way increased compared to efficient autonomous driving.
"Safety" is always the who-can-possibly-argue-against-it excuse for whatever people want to push. The painfully obvious example being, "safety" is the excuse why so many people drive obese, wasteful, SUVs. When people "play the safety card," I can feel my eyes rolling.
Bert, I am in total agreement with you when it comes to "the safety card" being played by anyone in the auto industry. I cringe. I just hope that marketing folks don't push this safety angle too hard, every time when they talk about autonomous cars. Because in the end, it will come back to bite them back.
What happened to driving talent and driver training? Am I to believe the ultimate goal is to issue a drivers license right along with the birth certificate? Sure would save a lot of time and money. Especially standing in line at the motor vehicle department. I have driven cars that no teen could drive today. Must be for "freedom" from learning.
It is quite easy to realize. Trucking industry. Especially long distance trucking. And in order to simplify navigation through narrow complex cities, it maybe that just truck can go on autopilot from one specially designed meeting lot on the outskirts of the city to another one like that in other city. From that lot to final destination inside the city, human driver can drive the truck. Big economic and safety benefit (software never sleeps behind steering wheel).
What are the engineering and design challenges in creating successful IoT devices? These devices are usually small, resource-constrained electronics designed to sense, collect, send, and/or interpret data. Some of the devices need to be smart enough to act upon data in real time, 24/7. Specifically the guests will discuss sensors, security, and lessons from IoT deployments.