Comment by jodrellblank | Hacker News Reader

Comment by jodrellblank | Hacker News Reader

Parent Full thread

Not even slightly excited about robot cars. Driving is fun, for a start!

human errors make up 93% of the causes of car accidents

What kind of human errors? Running out into the road in front of cars? Driving while too tired? Arguments with passengers? Not looking when changing lanes? Too busy looking for road signs to see a stopped car ahead? Driving horribly to try and go back for a missed turn?

Technology could help with a lot of these problems without replacing cars altogether. Current Volvos can have a blind-spot warning system, they can ready the brakes for you as you get close to a car in front so they will react quicker, they can silence incoming phone calls when the car detects too much activity (e.g. sharp cornering) so it wont distract you.

I look forward to the day when all road signs are gone and replaced with informational broadcasts that my car will interpret, and on the dashboard I get a neat summary of the current speed limit, the GPS navigation route to where I'm going, and that's about it.

You also aren't clear which figures you mention are worldwide and which are USA only. Worldwide, there are countries which don't require regular car maintenance, and people drive around with tree branches instead of steering wheels, for instance. Is that kind of thing 'human error'?

Don't worry, I don't want to pry the steering wheel from your cold, dead hands. Brad Templeton (the writer of the essays I linked to) predicts that robocars should be able to co-exist with human drivers (though there would be some advantages if all cars had autonomous capability, such as increasing road capacity by reducing follow distance, and optimizing the design of parking garages).

My interest is primarily on reducing deaths. I'm fortunate enough that I've never lost a close friend in a car accident, but two friends of mine have been injured (one moderately, one seriously), a close friend of mine lost his best friend, and the 21-year-old receptionist at my previous employment was killed in a head-on collision. I expect most people have stories like these or worse.

The 93% figure comes from Wikipedia, and the citation is a report that doesn't define human error. Templeton thinks that robocars could eliminate almost all of the million car accident deaths per year. But even if it's only 50%, that would prevent so much avoidable loss.

Templeton: "The creators will need to make an urban test track and fill it with swarms of the robots, and demonstrate that they can walk out into the swarm with no danger. Indeed, like a school of fish, it should be close to impossible to touch one even if you try. Likewise, skeptics should be able to get onto bicycles, motorcycles, cars and hummers and drive right through the schools of robots, unable to hit one if they try. After doing that for half an hour and getting tired, doubters will be ready to accept them on the roads."

If that can be achieved, robocars will drastically reduce deaths.

People who enjoy driving will have to do is in such way as to not endanger regular commuters - e.g. on the racetracks or private roads. 45,000 deaths per year is a lot of dead people. Unlike smoking victims who brought it upon themselves a good half of driving accident victims did not and could not see it coming.

People who enjoy driving will have to do is in such way as to not endanger regular commuters

As if the two groups don't overlap! As if the regular commuters are completely innocent and the people who enjoy driving both have and cause all the accidents! As if robocars could not and would not have any accidents! As if people who commute to work by car are forced into it and have no choice about being road users!

I disagree with all four implied statements.

a good half of driving accident victims did not and could not see it coming.

Then why do so many other countries have fewer road deaths per hundred thousand people and also fewer road deaths per miles travelled?

http://www.driveandstayalive.com/info%20section/statistics/s...

See how the USA is way down in 40th position - more dense countries, countries with more dangerous driving weather, other big sparse countries, other rich countries, much poorer countries are all above the USA. Switch the table to deaths per vehicle miles travelled and the USA is still barely in the top 10.

Just luck? Or better driver training so their drivers can see potential accidents coming and avoid them?

I know, arguing not hacker news. But someone is wrong on the internet. It might even be me!

I think the implication was that having any human drivers on the roads negates a lot of the potential benefits and safety provided by robotic cars. The drivers don't have to be bad to be worse than a robot. Driving a car yourself might go the way of riding a horse for transportation. You can still do it, and have lots of fun, just not down most city streets! Robocars would still have accidents, but hopefully more on the frequency of planes :)

I expect "no human drivers permitted on public roads" to be the law because human drivers are fallible. Robots are also fallible, but they tend to accumulate their lessons across entire robot population (at least within one manufacturing company) whereas each new human driver needs to be taught anew. Even well-taught, humans are still prone to errors and it's only a matter of time before standard robot is 10 times safer than 95th percentile of drivers.

The only problem with robots is systemic risk - some sort of common bug (e.g. Y2K) that would take out many machines at once or another thing that fails in a non-graceful manner. Car companies are not strangers to designing proper failure modes so I think it will pan out ok in the end, but software engineers aren't as good at it yet so we do have that risk for a while.

Driving is fun, to be sure, but there are enough of us that would love to be done with it. I currently commute an hour and a half each way in DC traffic. I LOVE driving, but due to other people, all the fun is taken out of getting to work. If I could hit a button and just 'get there' via auto-pilot while I finished work, that would be ideal (for me).

I could turn off auto-pilot, but then there wouldn't be any point. I think that the magic in getting auto-pilot to work in cars is in enforcing that enough other cars on the road are using it too. I truly believe that so long as there are any cars on the same road as you not on auto-pilot, that it will negate most of the benefits for everybody.

Regarding human error, I think that most of what you listed would count. The car is a human operated machine, which means that most of the time, wrecks are our fault. Unless the wheel falls off unexpectedly, or the brakes fail, you could pretty much always attribute a failure to the operator -- or the operator of the other car where applicable.

I would also not like to be rid of driving as a pastime, but until you do, I'm also not sure that there's any point in automating the driving experience. So while I am sympathetic to your point, the pain of my daily commute is sympathetic to the other.

Reply on news.ycombinator.com