BMW shows off 'drifting' self-drive cars | Hacker News Reader

BMW shows off 'drifting' self-drive cars

bbc.co.uk

67 pointsalsutton010203139 comments

Last time I checked this was Hacker News, not Reddit, and I am getting sick of articles like this making it to the front page (as I am aware other commenters have noted).

I didn't even need to click on the article (but did in a vain hope) to know that there was absolutely no information as to what they actually did from a technical perspective that deserves notice on HN. The most technological it gets is "Using 360-degree radar, ultrasonic sensors and cameras, the cars sense and adapt to their surroundings.", but then what does it do that is different from what the Google Car and other autonomous cars do? Is it any different from the Audi that drove itself up Pike's Peak as quickly as it could?

The rest is offtopic, but as a long time read-only user, I must come out of the woodwork and make my first comments (other than a few days ago) to be to beg of submitters and upvoters to please only submit/upvote stuff that is actually relevant to what should be the audience of Hacker News, not the general reddit population...

I got sick of reddit and 'general' news years ago and don't want to go back to it.

I agree this is a very superficial article, although it is somewhat relevant to folks interested in self-driving vehicles. Despite regular procrastination on Jalopnik, I hadn't previously heard about self-driving tech applied to anything in the motorsports arena.

Nevertheless you raise a good point re. the fluff content here and encroaching redditification. Lately a lot of "witty one-liners" have appeared up in the comment section, and downvote-enabled HN users should feel a duty to discourage them.

Audi has been doing autonomous sports cars for years. In 2010 they had a TTS climb Pikes Peak, albeit almost three times slower than the human-driven record for that course.

I sympathize with your feeling but this is far from the worse I've seen on HN's frontpage lately. At least it's not about politics.

In the time you spent venting your frustration about this, you might have been able to dig up something interesting for the conversation in here. Also, you can beat stuff like this for raising interest in technology. Case in point: http://www.youtube.com/watch?v=WsnKzK6dX8Q

You have some control over this, by the ability to upvote stuff on the 'new' page, helping the sorts of articles you like to appear on the front page.

I try to run through it once a day. Then, whenever a story I upvoted makes it to the front page, I get a small sense of satisfaction.l

Here's a promo video of it: http://www.youtube.com/watch?v=IL_enMPWT7s

I think I spotted an overturned cone at the 55 second mark ...

Wired magazine had a small piece about the presentation, including that video: http://www.wired.com/autopia/2014/01/bmw-builds-self-driftin...

Two in fact.

a lot of time in racing with cones they will lay one down next to another to indicate which direction to go.

source: I did a lot of autocross racing.

edit: looking at it again I'd be extremely surprised if it was anything but a direction marker.

Stanford have done some interesting work on this: http://www.youtube.com/watch?v=gzI54rm9m1Q

IIRC Stanford's motion planning relies on blending from a plan based on a physics model to a choice of learned operations programmed into the car by a stunt driver. Makes sense for demos, but I don't know how well it would work if surfaces or driving conditions were different from when the operation was taught.

This is bad news (despite being a BMW owner)

In future driving will no longer be fun, if you want to drive "non-automated" car or disable "auto-pilot" your insurance company will automatically jack up your premium, since they would consider self driving cars "safer"

More terrible than killing someone:

Spending more money to do it.

In future driving will no longer be fun

It's funny seeing BMW try to be a forerunner in this space, after all of their "BMW: The Ultimate Driving Machine" ad campaigns. They were one of the few car makers that actually spoke of driving fun! (Or, well, driving pleasure)

Note that "The Ultimate Driving Machine" doesn't imply that a human is the one driving. :-)

Is the cost to society (in lives lost) worth the cost to individual enjoyment?

The cost of all your enjoyment is your own life; you too will die.

It depends on how much pleasure you get out of driving. I certainly think it's worth some fraction of a life. And of course, a certain amount of the risk comes from inattentive drivers who don't enjoy driving, and would rather read a book or eat cereal (or both) on the road.

Car accidents have unfortunate side effects, in that generally more than one driver's life is put at risk.

Indeed. That's why I ride motorcycles instead.

There is, I believe, a certain amount of risk simply be being on the road at all. No self-driving car is going to prevent that.

What I'd like to know, though, is that as self-driving cars become developed, the labs are calculating the risk factors of various intersections, streets, exchanges, etc. That is, different roads can be analyzed by the robot cars and independently assigned a risk factor for a given section of street.

The resulting impact to traffic engineering and city resource prioritization could be a very interesting side-effect to the introduction of self-driving cars.

Life becomes less valuable when you can do less with it. In the extreme case, we could all spend our lives in styrofoam boxes hooked up to feeding tubes, but what sort of life would that be? I would argue, a valueless one.

Is a collective devaluing of human life worth saving a few individual lives?

Consider if we cut down all trees within a dozen or so meters of roads, and installed guardrails everywhere. We could probably save many lives, but at the cost of devaluing the beauty of the countryside. Everyone's lives would become just that much more depressingly monotonous and sterile, for the sake of a preventing a few roadside deaths.

Safety should not be our society's primary concern. Quality of life is a much better thing to strive for.

> Consider if we cut down all trees within a dozen or so meters of roads, and installed guardrails everywhere.

You realize that we did that and it resulted in more deaths, right?

Whatever, bad example then. My point stands unless you lack an imagination.

We could string safety nets between all skyscrapers and under all bridges. We could ban convertibles. We could put fences along all rivers and ban backyard swimming pools. We could ban fireplaces.

There are any number of things that we could hypothetically do to prevent accidental or untimely deaths that we do not do because they would degrade the quality of our lives.

Well, I wasn't trying to debate your point, but if you insist on being huffy,

Skyscrapers generally don't have windows that can be opened or easily broken, and very rarely have balconies. When's the last time you heard of someone going through a skyscraper window? It's a safety net.

We probably couldn't string safety nets under all bridges. However, we have already been fencing them off to prevent jumpers.

I'm not clear on why you are targeting convertibles specifically; I'm unaware of any special danger they pose. However, I would argue that banning cars entirely would raise our quality of life, if combined with other measures.

I'll concede the river one since there are too many possibilities there to argue exhaustively.

I could prevaricate on the swimming pool thing, since my parents own one and our quality of life probably would be better off without it, but not worth it.

I'm not clear on enough building codes to know how much we discourage fireplaces, but most new buildings don't seem to be built with them. We appear to have invented central heating over the last several decades, which is both safer and more effective.

The point being that the reason we don't implement a lot of "safety" measures is not because they "degrade the quality of our lives" but because they don't work. In fact, most of the so-called safety measures you luridly imagine actually should be done because they'd increase the quality of our lives... so really, the question is why we haven't done them, if QoL is such an important factor.

Disagree. Driving will now have increased insurance prices associated with increased risk. The roads will be safer.

Either that, or you'll take on some sort of hazard liability, and they'll find a way to see that you waive all coverage during the periods in which you're personally operating the vehicle.

Insurance companies have to compete on rate pricing to some extent, so there isn't too much room to move radically north there. But I could definitely see their getting into real-time pricing models, and/or denying coverage as a means of minimizing their responsibility in your agreement.

If you want to have fun, there will no doubt be private tracks where you can go and drive in a fun way. As someone who just gave up commuting about 1 traffic-laden hour each way, driving stopped being fun long, long ago.

Too many of my commutes were made an unnecessary 50% or more longer because of drivers who were probably having a lot of fun until they caused an accident and backed up the road for everyone else.

As someone with a similar commute, I lament I can only upvote this once. Every time I-95 severely backs up because some asshole managed to crash on a day with perfect weather and visibility, I think to myself: "I hope it was worth it."

As bad as being stuck in traffic is, usually you're having a better day than whoever was involved in the wreck, most of whom probably didn't do it on purpose.

I would LOVE to have the car drive me autonomously to the track and back. Highway driving is tiring/dangerous/boring.

I autocross and take my car to a private track for time trials, I'm starting to face the fact that this is not the best plan because there is now a karting track (0-120 in 6 seconds) that is closer to me.

My car, that I justified it's sportiness because of this hobby, is more expensive at the track, burns more, fuel, has more maintenance issues, and could just never perform the 3gs that a kart can.

Embrace the Kart. My dad used to do that as a hobby and they're impressive machines. Just remember that you've got no crumple zones and your legs are right at the front.

At the same time, the entire continent is finally pushing back on that 1-hour commute. People are returning to cities, transit and infill and dense construction are becoming political issues, etc. Part of the impetus for this re-urbanization is the misery of commute.

Beyond commuting woes, density is also more efficient for providing services and for environmental concerns.

So I'm slightly worried that a move towards self-driving cars will allow us to backslide on this progress. Obviously I want the technology - it will be safer, more efficient, and healthier - but I'm worried about how culture could shift around it.

When did people leave cities then? Which cities were shrinking but now aren't?

This article gives a good overview in the USA: http://en.wikipedia.org/wiki/Largest_cities_in_the_United_St...

I can't see any cities there that significantly shrank and then grew again. Care to name some if you know of any?

New York's population was steady from 1950-1970 before dropping about 10% between 1970 and 1980. Considering that the US population increased by about 50% between 1950 and 1980, I'd consider that a rather significant decrease.

Backslide, or return to making forward progress against crowding? ;)

I think the lives of humans are more important than some thrill seekers who want to drive around a few tons of metal without paying extra.

What's the point of a life not enjoyed?

The issue is only people putting other lives at risk. Driving is statistically unsafe for other people, not just yourself.

1) you can still drive traditionally on private courses. 2) your right to enjoy hurling yourself around in a 2 ton missile shouldn't be more important than my right to life

What's the point of life if you die? I don't want to die, therefore I strongly suggest we make it illegal for humans to drive vehicles on public roads. The trade-off is in my opinion really small compared to the millions of lives saved.

If you want to have fun, have it at places where it doesn't kill people. It's like smoking, outside is fine - you are only killing yourself - but don't do it in bars/the office/hospitals/other places where you might kill others.

What's the point of a life wasted when someone loses control of their car and injures someone who was driving responsibly?

I’m pretty sure people in one century will still be easily able to drive around for fun. Just like I can go horse riding for fun right now, despite horse riding being obsolete as actual transportation.

I’m also pretty sure that there were people waxing poetically about the joys of using horses for transportation a century ago and predicting doom and gloom because of the joyless and horseless future they foresaw. Those people probably didn’t shut up about it for the rest of their life, but in the end it all didn’t matter because they died and that’s that.

The thing is that most people on the street do not drive for fun, they drive to get from point A to point B and that’s that. Self-driving cars improve the quality and safety of going from point A to point B for them.

Why the scare quotes on "safer"?

If self-driving cars prove objectively to have fewer accidents and resulting injuries and deaths, wouldn't that just make them a superior technology?

In that case, why shouldn't insurance charge more for traditional driving if self-driving were available on your car? That's just how insurance works, nothing new there.

A self-driving car system might be statistically safer than a human drivers in overall conditions, and still be prone to failure in an exceptional situation a particular driver wouldn't have been. If you have the potential to be that driver in that situation, I could understand feeling the scare quotes are warranted. Which percentile of the population does an automated system need to outperform to be safer rather than "safer"?

I am (in my own estimation) a very good driver.

I drive a manual transmission car, and enjoy using the flexibility that provides in my daily driving.

I drive emergency vehicles of various sizes, in a lot of less than ideal conditions (I live in upstate New York, so weather is a big factor in driving conditions).

I have no problem believing that a computer will be able to do an objectively better job of driving than I can. It will be able to react more predictably in routine driving, and respond _much_ faster to exceptional situations. In a collision, a modern car is _done_ processing the event (detecting the impact, deciding which safety restraints to deploy, and deploying them) before your body even recognizes that it happened.

A computer controlled car will make fewer mistakes and react better in emergencies than a human being will. It has nothing to do with the skill of the driver. It has everything to do with the amount of data that the car can gather (do you have 360 degree radar vision or direct access to sensors built into the car), and the speed with which it can process that data. This technology still has a lot of maturing to do, but it has made amazing strides in the past decade, and I'm confident that it will continue to do that.

I, for one, welcome our new robot chauffeurs.

Further, once all the cars on the road are automated, they can alert one another to their own changes in routing, mechanical failures, etc, allowing everyone else the chance to reroute and avoid as needed. We could eliminate traffic lights and stop signs and we'll not only get to our destinations safer, but also sooner.

Not only that, but emergency vehicle will be able preempt roads directly, rather than relying on the 18 year old kid with headphones on to a) recognize the lights and sirens, b) care, and c) do something not stupid (like pull into the way of the fire truck...)

> they can alert one another

Until someone misuses the protocol to gain an advantage or to cause a disruption (assuming it is peer to peer, decentralized).

If it is centralized, then um, there are other worries (privacy, control, etc).

So you feel that 100% is easily within reach? As a coder, I can't help but think it'll be a while before we iron out all the bugs in the driving AI, and as a result, we'll see self-driving cars on the roads long before they are able to outperform 100% of human drivers in all situations, or even your average sober, attentive driver in all situations.

Even if their reaction processing times are orders of magnitude faster and the sensors give you significantly better awareness, if there are conditions where the sequence of AI responses and priorities are conflicting or wrong, then you'll just be faster and better at getting the wrong solution, sometimes tragically so.

I don't doubt that we'll ultimately be much better off with self-driving cars being the rule, but it may be a bumpy road to get there.

I never said easily. In fact I said the technology still had a lot of maturing to do. I also never said it would be 100% (though I'd dare you to find me any human who would meet that 100% criteria either).

I absolutely agree there is still a ways to go before this is mainstream technology. I also agree that we will get there eventually, and the world will be a better place for it.

The more interesting question might be the percentile where the auto-driver makes driving safer for everybody (by weeding out the worst drivers).

I would guess it is more like 15 or 20 than it is like 50 (but of course I'm guessing).

We (automotive active safety engineers) don't look at it as a percentage-of-inept-drivers, but more like the Trolley Problem[1], that is, overall harm vs benefit. To put it another way: seatbelts, airbags, and heart surgery hurt/kill some people, but the benefit outweights that harm.

[1] www.wikipedia.org/wiki/Trolley_problem

I can't see any reason why we can't have both; you control the car until it detects a dangerous situation and then takes over. Sounds ideal really, you could go faster than you can now, safe in the knowledge that if things go bad the computer will kick in and take over.

I think part of the safety boost of self-driving cars is that they don't get impatient, and neither will you since you're reading a book/email/facebook/whatever.

Self-driving cars ought to also be capable of driving like grandmothers. That's part of why I expect them to be safer. The idea that you can drive a car to the brink of a high-speed collision and then magically expect it to out-react a seasoned driver into avoiding an essentially unavoidable accident is juvenile fantasy.

There's no reason why you can't drive it to the brink of a crash and then have it recover for you. The 'brink' in this case is the latest the computer can reliably recover - not a point 100ms before an unavoidable collision.

But why would you want that? Why wouldn't you want to use a system that can get you to your destination much faster, cheaper and safer[0], only because you want to experience the "fun" of being in a traffic jam?

[0] - self-driving cars constantly talking with each other could drive much closer to each other than current "safe distance", much faster, while utilizing aerodynamic effects to conserve fuel, and that system would still in principle be safer than human drivers.

I'd think that very often by the time things have gone wrong, it's too late to prevent the crash even with perfect intelligence. Unless perhaps you have the system take back over so often that it's more frustrating to manually drive than to just leave it on auto-pilot anyway.

This is already the case today, albeit limited.

My Audi

- makes the steerwheel vibrate if I veer off my lane, actively works against that (limited 'self-steering', but really, really crappy, useless and stupid in its current incarnation)

- has lights in the side mirrors that light up when someone's left or right behind of me. If I set the turning signal to a 'blocked' direction, the lights flash. This system is immensely satisfying for me, works awesome.

- stops the car on its own if it believes that a crash cannot be avoided and while I still go 30km/h or less. It breaks _immediately_

- issues a high pitched warning noise if it assumes that a collision might be imminent but outside of the safety zone (> 30km/h)

- supports automated cruising control, that breaks/accelerates on its own, depending on the traffic ahead and to the side (it won't overtake a car while you're right of it).

That said: I love driving. I love leisure rides, love to drive on the Autobahn. That car's rather 'slow' (compared to previous company cars) and tops out at 220km/h. When it is safe I really like to drive at that speed.

I think the resistance to self-driving cars is based on mistrust (you give up control, after all) and ego mostly - but it doesn't help the industry or supporters of that idea to paint everyone that likes to drive as a potential 2 ton bullet. A little perspective on both sides would help a lot here.

The 2014 Mercedes E- and S-klasses that I've tested also do active intervention if you try to change lanes towards a car in your blindspot. Via braking on the opposite side.

Interesting. Would love to test that/feel how that works.

Here that's not combined, if there's something in my blind spot I just have those lights (flashing if I set a turning signal). There's an assistant that is supposed to keep me in my lane. In theory it ignored lane changes only if I set the turning signal (-> intended by me) and otherwise steers in the opposite direction (I can still beat it, it just tries to convince me otherwise). Plus, it vibrates the steering wheel.

It doesn't work, though. Audi was obviously so afraid that people start leaning back and drinking a nice cup of coffee that they added a very annoying 'are you still alive' test. That assistant is going to power down on its own as soon as it things you're not steering anymore. Which often happens without reason (just driving straight).

Basically that thing is only working if you hold the steering wheel like a 50's/60's Hollywood star, turning left and right on a dead straight street. Maybe the next incarnation will improve this design..

We should expect that, assuming that self-driving cars are in fact safer, insurance rates will go down for people with self-driving cars, and remain the same for manual drivers.

The rates for manual drivers will be a function of how safe that pool of drivers are. If a large portion of the manual drivers do it for fun and are more reckless than the current average driver, their rates may in fact go up.

They will go down for manual drivers, too. If you hit your brakes now, you get hit from behind. If the car behind you is a robot, it will come to a stop at 5cm from your car.

Unless, of course, the manual drivers counteract by driving more dangerously. Why stop for a red traffic light if all other cars are driven by computer? Can you try and get an autonomous car to drive into a ditch by moving your car very close to it?

I think there will have to be laws governing such behavior ('harassment of an autonomous vehicle'), with video cameras in the autonomous vehicles providing the evidence.

That also is a reason that I think driverless cars will not work everywhere for quite a number of years. In the typical old city in Western Europe, an automated car will have to negotiate with pedestrians and cyclists for room. Once the pedestrians know that a car will brake if 'challenged', they will change their behavior. I know I would (I guess driving a self-driving car through the center of London or, better yet, Amsterdam in rush hour must be one of the hardest challenges)

It might work out better than you think in your last paragraph. Auto routed vehicles will reliably choose the 'main' route through an area (instead of whichever they feel is fastest), so segregated routing should work better (and with little additional physical infrastructure you can have one way rush hour streets and such).

In many old city centers, there isn't much room for segregated routing.

Also, why would pedestrians or even cyclists care about one way streets? Even if they do, consider the case where streams of cyclists enter a one way street just in front of an automatic car. That could easily happen in Amsterdam (for an impression, see http://bicycledutch.wordpress.com/2013/03/28/amsterdam-bicyc... or http://www.nytimes.com/2013/06/21/world/europe/a-sea-of-bike...)

If there are 3 streets, the pedestrians and cyclists can take the one with no cars on it. In Amsterdam, cars could reliably be routed out and around (I think lots of people probably already use this strategy, but the point is that the auto router would never cheat).

Did you understand segregated routing to mean separate streets?

I don't think this will be the case.

Automated driving is only useful if you know where you're going. If you're on a leisurely ride through the country or are looking for a place with no address, you can't use automated navigation. So disabling it would be a necessity for a variety of driving purposes. As an aside, motorsport enthusiasts would be up in arms if they had to pay a premium just to be able to drive their expensive sports sedan the way they want on a closed track.

What will probably happen is a black box will record if the car was in automated mode when a crash occurs. If it was driving itself, the manufacturer will probably be at fault and forced to pay the insurance company. If you were driving it, the standard rules apply.

I have to say though, automated drifting is really depressing. I used to drive a Toyota MR2, which is an ass-heavy car that was very susceptible to snap oversteer. Unlike most other cars, if you felt like the car's rear end was about to slide out of control, you had to PUNCH the gas to get traction back. You basically have to fight all your instincts and do something crazy in a split second or end up spinning backwards towards god knows what. It's the most exciting thing i've ever experienced in a car (even more than sitting in on hot laps with an indy driver in a Lexus IS-F), and I would never get to experience it with automated driving.

As for non-drifting, automated driving is definitely an amazing and useful feature. Writing code on my way to work sounds like fun. And I know it sounds controversial, but if there was some way to couple a breathalyzer with automated driving, we might be able to use our cars as taxis that pick us up and drop us off without us having the ability to drive. But perhaps that's hoping for too much...

Please stop "leisure driving". Traffic is a tremendous drain on my life as it is with superfluous cars on the road.

I suspect people looking for a "leisure drive" abhor traffic as much as you. They're probably on quieter roads with fewer people, cars, and cops.

Please stop breathing. Carbon dioxide is a tremendous drain on the life of the planet as it is without your hot air added to it.

Leisure driving isn't usually done on roads that are full of other cars all in gridlock trying to get home. It'll be on those same roads at much later times in the day, when the traffic is thinner, or it'll be off on a country road or mountain pass somewhere.

I love to drive. I do not enjoy dealing with major backups. When I'm leisure driving, it's out to a small town in the middle of nowhere later in the afternoon or when everyone is at work; or, it's out in the boonies because wow that view when I take a certain turn is just absolutely spectacular.

> If you're on a leisurely ride through the country or are looking for a place with no address, you can't use automated navigation.

I don't think those are impossible problems to solve. For the leisurely ride, define a radius or area and let the car do a 'random' walk (or ride on only small roads), or visit beautiful spots where g+ users have shared pictures, etc. For a place with no address, give a lng,lat or drop a pin on a map.

You could even use way-point coordinates...like with an Xbox controller. Tell it where to turn and the computer will safely do the maneuvers that take you that route.

I like controlling the car around those leisurely bends. Feels good to experience the wheels rotating in a certain way around a banked turn WITH an amazing view.

Insurance won't necessarily go up for manual driving. At least at first, it will be lower for automated cars (once insurance companies have enough data). Assuming sufficient competition, insurance companies charge just enough to cover their costs plus some profit. Statistically safer cars means fewer claims, and lower cost. A good opportunity to undercut any competitors not on the same page.

Sure you would pay more than owner of automated car, but probably not more than you are already paying. People on bicycles don't pay any insurance at all, I don't see anybody complaining about them (for that reason anyway).

The scenarios in your first paragraph can be handled by treating the controls of the car as an input for the self-driving system. The accelerator asks the self-driving system to move forwards, whilst the steering wheel asks it to turn into the street to your right/left. Or, recognise that the existing controls no longer fit the purpose and come up with a better way of giving such input. For example, you might just point to where the car should go, or say "stop here".

As a BMW owner myself, I say: "good riddance to driving being fun." I used to get a kick out of it when I was younger, but now that I'm married with a kid it's just an exercise in paranoia and frustration. People are fucking nuts in the way they drive. My friend just got T-boned hard enough to flip his car 360 degrees a couple of months ago, by some asshole speeding in the shoulder as traffic was backed up. I see an accident at least once a week commuting into Philadelphia. Teenagers (16-20) just plain shouldn't be on the road the way most of them drive. It's no wonder that traffic accidents are the leading cause of death for people 5-35. The government should ban non-self-driving cars as soon as feasible. If you think about it: it's utterly ridiculous that a publicly-funded infrastructure with a crucial utilitarian purpose should be used for "fun."

"by some asshole speeding in the shoulder as traffic was backed up."

fun driving and reckless endangerment of the people on the road around you, including aggressive disregard for the rules of the road, especially driving on the shoulder are two very, very, very different things.

Raise the speed limits and make fines and enforcement more frequent for not using turn signals, aggressive lane-changing where other people are on the road, tail-gating, driving in the shoulder (REALLY???) and so on.

Oh, and make that driving test harder.

I'd say lower the speed limits. And implement ANPR systems on every straight part of every road, make them enforce those limits in a ruthlessly automated way. That should finally let traffic engineers do some real work optimizing the road network.

And then of course ban non self-driving cars from public roads the moment it becomes feasible.

EDIT: disregard the "lower the speed limits" part. I'm not sure why I wrote it (it was a long and tiring day...). Speed limits are mostly fine, but adherence to them is not. I stand by the rest of the comment.

Roads are already engineered in the general case for faster than the speed limit. Why do you want them to go even slower?

You're right; I'm not sure why I wrote about lowering them.

> fun driving and reckless endangerment of the people on the road around you, including aggressive disregard for the rules of the road, especially driving on the shoulder are two very, very, very different things.

The two are not disjoint things. A lot of people, particularly teenagers, engage in reckless driving behaviors because it's fun.

And that particular style of 'fun driving' should be disallowed. It's not even allowed on race tracks; but there are other kinds of fun driving that I don't think should be stopped.

You're associating the two more closely than I am. Then again, I have a lot of fun driving many of the roads that I drive on and the fun that I have doesn't include the things I mentioned in my negatives list.

You'd pay a premium over those that exclusively use auto-pilot. That premium should still be lower than what you pay now. Reason being, you'll still be driving yourself, but now you're surrounded by auto-pilot vehicles. Therefore, you should be involved in less accidents as auto-pilot cars react faster to your mistakes, avoid following you too close, avoid merging into your lane, backing into you in a parking lot, etc. So, driving on auto-pilot might be 90% safer, and driving without auto-pilot, in a world where the majority use auto-pilot, you might be 60% safer than today. Either way, less accidents, safer roads, lower insurance.

Considering insurance companies already insure airplanes, and you can turn the autopilot off without losing insurance, I somehow doubt this future. Airplanes are both more expensive and cause more damage when you crash them, and yet, you can rent them and insure them all the same.

IMO, that's because the AI is not yet proven to be actually safer than human. And it's pretty interesting what will be happen if the AI eventually comes up. Anyway, it doesn't seem to happen anytime soon.

As I understand it, most car accidents are caused by driver error, and the belief is that self-driving cars will have a better miles-driven-to-accidents ratio.

However, I'm not aware of the same being true for airplanes, plane crashes, and autopilot systems.

Indeed, the GA autopilot systems are pretty simple: keep wings level, altitude hold, vertical speed hold, etc. On the other hand; you can take a lot more risks in your plane that will cost the insurance company a lot of money -- tailstrikes, flying into IMC, flying while sleepy, etc. If you total your Piper Cub, that's going to cost the insurance company a lot of money. If you total your Prius, that's cheaper.

On the other hand, it's a lot hard to get a pilot's license compared to a driver's license. (What are the requirements for a driver's license these days? Having a pulse? No, I don't think they check for that...)

I think that's absolutely reasonable that they remove your premium if you want to remove other people's safety just for your own fun.

Driving cars for most people is only fun in ads, where people always enjoy empty streets, wide roads across beautiful wilderness, and the joy of utter freedom. Truth is that almost all of real life driving is a terrible chore, and "having fun" on the road generally means taking irresponsible risks for you and others.

This is great news to me.

Automated cars open the doors to:

- A potentially larger market for private tracks where you'll be enjoying racing, or even pleasure driving. It would be totally possible to have roads that are only here for pleasure driving. I can picture the Izu Skyline or other really beautiful roads that you pay to get a marvelous driving experience.

- A much safer world. If every car on public roads would be driving itself, you would not have to worry about a driver having an abnormal behavior and hurting/killing you in the process[1].

- Cheaper insurance? I don't really believe this, but it could happen, as there would be a lot less accidents.

I enjoy driving, I really do, but I'll use a self driving car any day, and I'm totally willing to pay an extra to drive in a safe and controlled environment like a private track. I think it's a win for everyone. If you enjoy driving, you'll enjoy staying safe, and self driving cars are bringing this safety.

[1]: I recently had an accident while coming back home, an old driver panicked because he didn't know where to go, and suddenly pushed hard on his brakes.

> - Cheaper insurance? I don't really believe this, but it could happen, as there would be a lot less accidents.

I haven't run the numbers myself, but it'd be worth checking to see what insurance numbers look like for private yachts or horses.

Good news is this future won't be in our lifetime :-)

Wonder when Google will have their self-driving fleet taking people from the Las Vegas airport to their hotels during CES? It would be a great publicity stunt.

"Currently California, Florida and Nevada have licensed autonomous vehicles to be tested on their public roads, and Google's fleet of 24 robot Lexus SUVs (sports utility vehicles) have clocked up about 500,000 miles of unassisted driving so far without any reported mishaps."

Can you imagine the emergency procedures they have in place to cover up any mishaps?

A few payoffs here, a few altered records there... no mishaps!

Do tell, how is one to search for news about a mishap? I think the Googlebot may have difficulty recalling such articles.

Insurance won't have higher premiums for manually driven cars, because in the future, manually driven cars will be banned.

It's absurd to insist that some people's enjoyment of driving overrules others' right to life. If you want to enjoy driving, you can do it on a private track where you only risk your own life and the lives of others who want to drive for fun.

I thought about this over and over. As it is usually the case with these articles (same with Google cars etc.) there's a group of supporters, usually with the 'It will be safer for the whole mankind' argument, sometimes with more mundane 'I would like to work in my car and let the thing drive itself' car-as-single-person-bus thrown in. And a group of people that firmly believe that this development takes away their freedom to (drive|steer|joyride).

I guess I should've put a disclaimer that states that I sympathize more with the latter kind of people, but try hard to find out why that is the case. For me, assisted driving is a no-brainer. My wife hasn't ever driven a car where the steering wheel is unsupported. Nor a car without ABS and related/improved technology. In my world that's a good thing. The car already, today, decides that it has to take control in a number of ways.

Further down in this thread I described my company car and its assistance systems. Most of those are nice to have and useful. I configured that car myself and included all those 'Let me help you here..' systems. This is not a matter of pride, no 'People that use assistance systems cannot really drive a car' attitude.

I guess for me the conflict starts when we talk about giving up _all_ control. For me the perfect solution would feel like a car today. It behaves as I expect it to, if I kick down the car responds. BUT I would be totally fine with more and more assistance systems that prevent me from screwing up. Not the 'Let the car drive' future, but rather a 'Let me drive (potentially optional), but step in if necessary'.

The result is more or less the same, but one feels like giving up something (a freedom, a privilege, a thing I enjoy), the other sounds like a reasonable and intriguing improvement.

I don't really want a self-driving car. I want an intelligent car that prevents me from causing trouble.

I'd rather have a car I can drive if I want to, or hit a button and have it just autopilot there instead.

I am all for this system as well. I don't think you and the grandparent are in disagreement, either. It looks like what they're saying is "give me varying levels of automation, including but not limited to full automation --- as long as I can have as much control as I want at the time." similar to a car with an option to disable traction control.

This sounds a bit like the car in I-Robot (which happens to be Audi built..).

The movie where the hero is blamed for causing an accident because he turned the autopilot off to avoid being killed by two large driver-less trucks sent to kill him.

At some point, your suggested improvements will result in a car that only gives you the impression that you are driving.

I agree. This is a slippery slope that I (think I can) accept. Perception matters. And that way the car (uhm - the technology?) gains trust slowly - and adapts to our preferences. It's one thing to talk about assistance systems taking more and more control (but being helpful in general, leaving me in control most of the time so far). It's a different thing to talk about self-driving cars that are tested today (even if they're far from ready yet). I don't want to hand over my current car tomorrow and enter a kind of cab that goes 130 max even if there's no traffic. Where I am nothing but a guest.

Do these lines converge? Possibly. Or certainly, even. But that needs to be a loooong process in my world and I prefer the route I described: By making the car I know and like smarter. Not by replacing it outright with a one-wagon-train or automated cab.

(and that's from someone that considers anyone not driving automatic already being much further down the road)

Drifting is suprisingly hard. I took a class and after two days was only able to do a little: http://www.youtube.com/watch?v=k-lN_jhvpqQ

I am thinking about how to organize an autonomous race in 2015 so this is pretty exciting for me.

Does anyone know if it was pre-programmed with the course and made minor on the fly corrections or was able to complete the course independently in real time?

"One 2013 study by the Eno Center for Transportation suggested that if 10% of cars on US roads were autonomous this could reduce fatalities by about 1,000."

This doesn't quite seem right. How does a 10% shift towards autonomous vehicles only equate to a 2% decrease in fatalities? Why wouldn't this scale linearly?

Who says it doesn't scale linearly? That certainly can't be concluded from the sentence you cited. Are you assuming that self driving cars result in 0 fatalities?

Roughly, yes. I'm assuming that at 100% autonomous cars there will be ~0 fatalities. From the article, at ~0 autonomous cars, there are 50,000 fatalities annually. I guess I'm just curious as to the curve that dictates the relationship between autonomous car adoption rate and annual vehicular fatalities.

I found the study referenced, but I haven't quite found their methods.

http://www.enotrans.org/wp-content/uploads/wpsc/downloadable...

If a car company has their own self-driving technology, will they ever pay Google for theirs?

Of course, why would they?

The interesting thing is the motivation. BMW want to make a better car, to sell more cars. Google want self-driving to free up driver's time to consume media, which will be covered in advertising - exactly the same as releasing Android to drive mobile Internet usage. But just as Google acquired Motorola, which automaker is in their sights? GM?

That's what I've been wondering. I know VW, Daimler AG, Toyota, Volvo, BMW, and others have making pretty good strides in self driving tech, but it seems most American car companies are pretty far behind already. I wonder if Google intends to license the tech out to the companies who can't or won't do it themselves, or if they decide to pull a Motorola and enter the manufacturing market at some point in the future.

Google could be hoping to diversify away from "everything free but ad supported" as they are with things like App Engine.

They could plan to license the technology to car manufacturers for a per-car fee or subscription; along with the detailed road condition data gathered by streetview cars and android phones.

I feel sorry for Google's shareholders if this is what they think.

BMW has a joint research lab with Google in Munich. My guess is this is the result of its work.

Google has various patents on self-driving car technology.

It would be nice if the auto manufacturers spent the effort on making their existing products better before they distract themselves with these projects.

Case in point? My car uses 1 quart of oil every 1000 miles (123K on the odometer) with a perfect maintenance record, and the manufacturer say this is Normal. Raise your hand if you like to stop in the middle of a family outing to pour a quart of oil into your car.

Or the power window motors that failed every 60K miles. Or the vinyl drivers seat that cracked to pieces. Or the traction control that engages on wet manhole covers. Or...

> It would be nice if the auto manufacturers spent the effort on making their existing products better before they distract themselves with these projects.

Actually, any effort expended on bringing closer the time human-controlled cars are banned from public roads is a Good Thing. As in, lots of fuel, time and human lives saved. Can't wait.

So, as a cyclist that commutes to work everyday on my bike--are any of these manufacturers making sure that their vehicles don't squish pedestrians, or even worse, people on bikes?

Yes, of course they are. That's a major part of the challenge.

They probably do since a collision might damage the car. And with all the sensors, there must not be a single blind spot left. That means it becomes much safer to ride a bicycle or motorcycle.

They will, in fact, do a much better job of it than a human driver.

An automated car will detect the obstacle faster, make a better decision on evasive maneuvers (based on a much better understanding of the current conditions), and apply that decision _much_ faster than a human driver could.

I mentioned this elsewhere in the thread, but a modern car processes a collision _very_ quickly. It detects it, decides what to do about it (which airbags to fire, which seatbelt pretensioners to deploy, etc), and implements those safety measures _before the driver is aware the collision has occurred_. Apply that same processing latency to evasive maneuvers and you'll see just how much safer automated cars can be.

EuroNCAP[1] requires increasing levels of automated pedestrian protection -- i.e. autonomous braking or avoidance -- to get a 5-star safety rating starting in about 2015. Because World Health Organization predicts 3% of fatalities in 2030 will be automotive collisions, and over half of those 3% will be pedestrian deaths[2].

[1] EuroNCAP.com

[2] Could't quickly find the 2030 stat, but pedestrian fraction: http://www.who.int/mediacentre/news/releases/2009/road_safet...

So what will be the point of laying down a tonne of cash for a "BMW - Sheer Driving Pleasure" if you can't drive it?

Pleasure for BMW, not for the drivers.

There is all this talk about safety but isn't drifting your car inherently dangerous? It's not like you're drifting your car around turns on your way to work. And if you are, you're a jackass and putting everyone else at risk.

The benefit comes from automated driving systems being able to handle emergency out-of-control situations safely and predictably, not in regularly being driven around like this.

Let's say you're in a self-driving car and get pegged and flung across a highway by another car going 20 mph faster than you. In this case, a driver with high performance training would be better able to land the crash than an average driver that panics and makes unsafe corrective inputs.

You want the self-driving car to be able to respond like the high performance driver, which means knowing how to safely drift.

Anybody know if self-driving cars can distinguish between safe and unsafe object; say a big stuffed animal, and a rock in the middle of the road. Would they make big maneuvers to avoid both?

Human beings aren't especially good at that either...

Tell me about it. I am constantly surprised by how close I need to get to the target in low light conditions to realize it is really not what I earlier recognized it to be.

It usually makes me wish I could enable some sort of Splinter Cell heat sensor.

Or a rock inside a big stuffed animal?

Yes! That's how robots should behave. Robot cars drifitng and bipeds doing backflips and pushups when handstanding on one arm.

"Fast and Furious 1X" will be starring self-driving cars i guess.

Things like this make me rethink my decision to stay in academia.

Great. They taught computers to drive like a dick as well now...

0_o Why don't we work on self-driving cars that drive correctly, first? Then the "HERP-DERP, I'MMA DRIVIN' SIDEWAYS!" enthusiasts can work on their own cars (hopefully on their own closed tracks and roads).

Being able to drive sideways may be tech that applies to recovering from an emergency situation like black ice.

Reply on news.ycombinator.com