Mercedes-Benz's Self-Driving Concept Car Is Here

cartechboy writes: Mercedes-Benz has finally taken the wraps off its autonomous concept car, dubbed the F015 Luxury in Motion Concept. Shown at the 2015 Consumer Electronics Show in Vegas last night, the concept is a self-driving, fully-connected, hydrogen-electric plug-in hybrid that touts a vision of driving in the future. Mercedes says this concept is not only a means for getting someone from one point to another, but also a usable space for entertainment or work as well as a platform for communication and interaction. The hydrogen-electric plug-in hybrid system is unique in that it produces zero emissions at all times. It consists of a hydrogen fuel cell stack, a lithium-ion battery, and two electric motors. The F015 has a driving range of 124 miles with a fully charged battery, and up to 684 miles with a full tank of hydrogen. While not intended for production, Mercedes shows us that it has the technology today to produce a zero-emission vehicle that can drive itself. In related news, Audi has just shown off an A7 that drove itself 550 miles from San Francisco to Las Vegas for CES.

Comments

"while not intended for production"

[turkeydance]

that's the kind of responsibility i want.

Re:"while not intended for production"

[QuantumReality]

None of today's self-driving car is ready for production and will NOT be until we make real AI. Too many variables for that kind of system, pattern recognition and prediction systems that is unreachable for current algorithms and electronics. Maybe in 30 years from now.

Re:"while not intended for production"

[ubergeek09]

I would like to see mandatory personal vehicle ownership go die in a fire. I hate having to own a car just to go to work. If I can get a reasonable alternative, say a company has a fleet of these things to pick people up and drop them off I would be so happy.

Re:"while not intended for production"

[0123456]

If I can get a reasonable alternative, say a company has a fleet of these things to pick people up and drop them off I would be so happy.

I expect you'd like a pony, too.

Best thing is, they're self-driving, so long as you don't really care where you end up.

Re:"while not intended for production"

[bill_mcgonigle]

It's like bug spray for bureaucrats.

Re:"while not intended for production"

[Scottingham]

Man will you be surprised in 3-5 years.

Just look at voice recognition's progress within the last two years alone. It had reached a plateau of about 80% accuracy, even when speaking slowly, clearly, and without regional accent. Now? I can mutter drunkenly and it'll get it more often than not. I can say very local business words like Phydeaux (upscale pet store) and get the spelling right!

It's called machine learning. The more experience/scenarios etc these self driving cars get, the better they'll be.

I'd say I'm as confident as you (but in the opposite direction) that within 10 years computers will out-drive humans in all scenarios. It won't even be close. 360 degree sub-millisecond informational input versus our meatbag eyes and reflexes??

Re:"while not intended for production"

[0123456]

Meanwhile, back in the real world, my car can't even figure out when I say 'redial', and gets about four of of ten numbers wrong when I try to dial a number directly.

Re:"while not intended for production"

[RingDev]

Time is moving faster than you expect. The model year 2017 vehicles in the US are all required to have a V2V communications systems in place when they roll of the line. Most of the V2V interfaces also support V2I communications.

We will be seeing some radical advances in automation, crash avoidance, and information services to motor vehicles in the next 10 years.

And the autonomous vehicle with self-drive functionality doesn't need to understand 100% of all possible traffic/infrastructure interactions. As soon as it encounters a pattern it is not designed to handle, it can bring the vehicle to a stop.

To be fair though, have you ever done the San Diego - Las Vegas drive? It's nothing but desert highways. It's such a boring drive you're at more of a risk due to falling asleep than you are to any road hazards.

-Rick

Re:"while not intended for production"

[jklovanc]

They do, they are called taxis, public transit, van pools, etc. There are many alternatives to personal vehicle ownership.

Re:"while not intended for production"

[ubergeek09]

Cars are expensive and spend a lot of their time parked. There are already companies like zipcar, which have memberships where people can use the vehicles and for many it's more reasonable than owning their own. If a company would take over both the civil and criminal liability associated with an autonomous car driving me around and let the passengers sleep/whatever I would line right up for a membership. I'd even pay to 300-400 a month for it too.

Re:"while not intended for production"

[QuantumReality]

What are you comparing ? You seen how hard was it to crack captcha only ? for VISUAL recognition ? Because VISUAL not sound is important here. Sound is EASY like 1 million times easier. You don't know what you are talking about, trust me, we are not even near the real solution. You have so many different shapes in real world, colors, different motions that human can understand and predict next but algorithms just can't, don't be naive. Without real AI forget about that cars on streets.

Re:"while not intended for production"

[ubergeek09]

Not in the suburbs.

Re:"while not intended for production"

[Anonymous Coward]

It doesn't need to be perfect to be useful. It just needs to be better than a human.

Autonomous cars are getting better at driving. Humans aren't. They don't have that far to go.

Re:"while not intended for production"

[SuricouRaven]

We don't need real AI. We need good-enough-to-be-safe AI, with human intervention available when the AI calls for it.

Re:"while not intended for production"

[Anonymous Coward]

Voice recognition algorithms haven't improved much. What has improved are the massive databases behind it. Did you notice you're using an online, service based recognition engine and not a local program? Every sound that the app hears gets stored then data mined along with a lot of metadata about the sound. The data mining is what's improving, not the recognition.

Re:"while not intended for production"

[QuantumReality]

Ye, stopping from 60 to 0 every time it will not recognize something... and it will do it so often.... really secure solution.

Re:"while not intended for production"

[QuantumReality]

Humans can recognize things in real world which machine can't without learning first. Another thing good luck with recognizing lanes by that car on really snowy day, or how they turn. This are things that machine can't understand on it's own, like it can't understand that it exists.

Re:"while not intended for production"

[nospam007]

"Meanwhile, back in the real world, my car can't even figure out when I say 'redial', and gets about four of of ten numbers wrong when I try to dial a number directly."

That's because it's not a Benz.

Re:"while not intended for production"

[nospam007]

"Humans can recognize things in real world which machine can't without learning first. "

Alas, no. People see what they expect to see, their brain is hardwired that way.
That's why they never see the biker, they expect to see cars only in the mirror.
Machines do not make that mistake.

Re:"while not intended for production"

[0123456]

Machines do not make that mistake.

You're right. Machines see a bunch of random blobs and think it's a bike.

Re:"while not intended for production"

[ubergeek09]

During peak there would be the possibly of carpooling effectively creating a privatized transportation system. Those that don't want to carpool could pay a much higher cost. Also not everyone is on the road at the same time during rush "hour". Where I'm at it rush at around 6:30 am and runs till 9:30 am and starts again at 3:30 pm and runs till 6:30 pm. If you assume a half hour commute and two riders per vehicle you could potentially cut the number of vehicles used in the morning by a factory of 12. Probably a bit less since the vehicles will have to pick people up at two different locations and drop off as well as driving between rides, so I'll round down and say you can decrease the vehicles used by a factor of 10. So now you have a vehicle that before would only be used by one person is now being shared by 10. Even if you assume a 200% markup on costs you will still see a decrease in costs for the consumer by a little over a factor of 5. I say a little over because now you don't have to worry about parking.

Re:"while not intended for production"

[fustakrakich]

As soon as it encounters a pattern it is not designed to handle, it can bring the vehicle to a stop.

Well that would suck if it happened while crossing the railroad tracks.

Re:"while not intended for production"

[Toshito]

Voice recognition is a joke. 9 times out of 10 my gf's android phone is way off and can't recognize half the words.

And yesterday I tried Youtube's automatic close captionning, on a video with a guy speaking correct american english, in a very quiet room, and it was really funny, as ALL sentences where complete and utter garbage.

Re:"while not intended for production"

[fustakrakich]

Let them have the damn pony. Then they can clean up the shit and pay for the vet.

The best system will be like some places do with bikes now. You swipe your library card and use the 'transport pod' to go where you're going.

Re:"while not intended for production"

[blue9steel]

Certainly, and they all currently suck. In very specific cases they can, with a lot of work, be made to work semi-acceptably but for the most part the experience is dirty, crowded, slow, dangerous, inconvenient or all of the above.

Re:"while not intended for production"

[0123456]

During peak there would be the possibly of carpooling effectively creating a privatized transportation system.

People can do that today. Most don't, because they don't want to have to spend ten minutes waiting outside someone else's house and end up late for work because they couldn't get out of bed on time.

None of these utopia schemes work so long as everyone has to travel at roughly the same time.

Re:"while not intended for production"

[ubergeek09]

I was more or less imagining a carpooling system like some cabs do today where they pick up multiple riders en route, but okay you don't want that. So you can decrease costs by a factor of 2.5. Still a significant improvement.

Re:"while not intended for production"

[jklovanc]

I agree with your assessment of the alternatives and I choose not to expose myself to those issues and therefore choose to own a personal vehicle. This is a choice and not mandatory as the GP stated.

Re:"while not intended for production"

[RingDev]

I disagree with your assertion, and it just so happens that I have a software development background that includes voice recognition, visual interpretation, and vehicle communication systems.

Voice recognition is a really cool subject. There are a finite (and surprisingly small) number of sounds that make up the English language. But for each of those sounds there are nearly an infinite number of dialects. This is combated by looking at the context of each individual sound to find known sound patterns. This can still result in thousands of different words, so we look at the context of the sentence. And through out this effort you have to deal with people talking quickly/slowly, while emotional or out of breath, with a southern drawl or a New York accent, or ESL folks that may say put words in an order that makes the context obvious. Heck, we did it for early onset dementia patients. The amount of work it took to dial in the VRE even when we knew the specific user and their inflection/dialect was massive.

When it comes to visual recognition, the AI for a car doesn't need to understand what a dog is, it just has to recognize that there is an IR source on the current trajectory. The AI doesn't need to comprehend what a boulder is, it just needs to recognize that there is an obstruction in the road.

The AI doesn't need to be able to identify that there is a lady wearing a yellow spotted sun dress doing interpretive dance while high on mescaline in the middle of the road, it just needs to identify that there is something in the road and respond accordingly (slow down, swerve, stop, etc...)

The exact same thing is true for humans. For example: the other day I was driving my wife's car. I was pulling out from the gas station and something caught my eye after I thought traffic was clear. So I slammed on the brakes thinking I missed something. Turns out it was a reflection of a dash light on the side window.

Yeah, there will be false positives. A vehicle may decide that a pothole is actually an obstruction, or that the railroad track is the end of a road. But in the vast majority of cases these are targets that the vehicle can identify from a significant distance away. It's not like you're going to be driving on the open road then come screeching to a halt for no reason.

And each of these false positives is something that will be handled through refining the AI system. Through IR, RADAR, real-time 3-d surface mapping, V2I communications, V2V communications, etc...

It's coming, and quickly. And many municipalities and states are looking to leverage the V2I systems. If you have a V2I enabled vehicle in Las Vegas, you can actually get the system to tell you how fast you should drive to avoid hitting red lights. Most major metros already have systems in place picking up tire pressure and blue tooth signals to determine traffic volume and speed, which is how those fancy "12 minutes to exit 123" signs get populated.

As the 2017 cars start rolling out and more data becomes available, we'll see technology leaping ahead. For example, in Wisconsin, our 511 site has a public facing developer API, so even if the state can't invest in some cool new apps and vehicle information systems, individual developers and manufacturers can: http://www.511wi.gov/Web/extra...

-Rick

Re:"while not intended for production"

[MadMartigan2001]

Actually, both sound and light are identical in complexity, both are waves. Interpreting sound is no more "easy" than any other wave of information. Imagine trying to teach a machine to listen to you on a sidewalk with semi trucks and people talking and yelling for cabs. The fact that the human ear can distinguish between a cacophony of sounds and decipher what one person is saying is remarkable. With that said, we do not have a machine that can listen accurately on a street corner so even though sound is no more difficult that light, it still may not arrive for a while.

Re:"while not intended for production"

[Ralph Wiggam]

That's unfortunate that you bought a crappy car. My voice recognition works great.

Re:"while not intended for production"

[Archangel Michael]

I'd also say, that as more of these cars are on the road, they'll create more accurate represenations of those roads to navigate by, and changes in conditions would require attention will be diminished. The reality is that scaling does matter. More cars = More data = Better decisions = better experience.

I can easily imagine, GoogleCar/Lyft changing how we commute. Imagine autonomous busses.

Re:"while not intended for production"

[mbone]

It's a concept car. Concept cars never are almost never turned into a product, they are intended to generate press and publicity, which this one is succeeding in doing.

Re:"while not intended for production"

[Firethorn]

You're right. Machines see a bunch of random blobs and think it's a bike.

Doesn't really matter. On average you don't want the car to be hitting said blobs, whether it's a bike, sign, people, another vehicle, animal, or random road debris.

Of course, the 'race for the self-driving car' seems to be shaping up to be a bit like the race for the first heavier than air flying vehicle. You ended up with multiple companies working on multiple solutions, and many of them more or less succeeded in a relatively short period of time. In some cases they had to carefully consider time zones to figure out who 'invented' some bit first.

Re:"while not intended for production"

[mjwx]

Man will you be surprised in 3-5 years.

By surprised you mean, wont see any autonomous cars then yes, yes I'll be surprised.

Just look at voice recognition's progress within the last two years alone.

Actually it hasn't really made any. it's just received a lot of marketing in the last two years. Realistically voice recognition isn't any better than it was back when Google Voice Search was released 4 years ago. It still struggles with accents, it completely fails with slang or mispronounced words, accuracy suffers when you're not in a perfectly quiet environment. Beyond this, voice recognition on your phone works by sending a recording your voice to a remote server that has a lot of computing resources, it isn't done locally.

Voice recognition is the ideal example of why there wont be autonomous cars in 5 years.

I'd say I'm as confident as you (but in the opposite direction)

And you'll be very surprised in 10 years when that doesn't happen.

Re:"while not intended for production"

[Glarimore]

Phones are in the real world too, so I guess you bought a shitty car...? I'm afraid I don't get your point.

Re:"while not intended for production"

[C0R1D4N]

From what I've seen/heard the autonomous cars will never choose "swerve out of their lane" unless something another vehicle is heading right for them. For deer/pedestrians they brake and they do it damn well.

Re:"while not intended for production"

[Headw1nd]

"And the award for most contrived example goes to...." I kid, I kid.

But seriously, how many human drivers would pass your test? I'd imagine they're do worse that the automated cars by a long shot. Here's why: a human driver in your example is going to to be overwhelmed with information as soon as the snowboarder appears, in this confusion they are likely to forget the dangers of swerving right or left.

In contrast, the machine does not become overwhelmed. The world passes very, very slowly to it. What does it see? Everything in your example, but in a simplified context. It sees the guardrail. The guardrail is a known item to it, it signifies danger. It cannot cross the guardrail. It sees the truck in the other lane, it is a large object and moving quickly, it cannot move into the path of this object. There is a new object in the road, an object that maybe be a person or animal. It may have to collide with this object. It begins to apply the brakes in a controlled fashion, and lightly impacts the snowboarder. All because it wasn't driving 60mph on an icy mountain road like a goddamn idiot.

Re:"while not intended for production"

[gl4ss]

a car doesn't really need a real AI.

a real AI is so much more different concept than something that logically drives a car perfectly in all possible situations. it just needs to drive a car, it doesn't need to be able to come up with the concept of what a car is on it's own.

that kind of says how far we are from the hipster shit singularity in reality.

Re:"while not intended for production"

[RingDev]

And here's some counters:

The autonomous vehicle, detecting slippery road conditions, and not being an over confident idiot driving way over a reasonably safe speed would slow down. Making emergency maneuvers less risky.

The autonomous vehicle, using V2V information, knows exactly where that oncoming truck is, and is able to communicate it's knowledge of surrounding risks and travel plan (ie: I'm going to stay in this lane, you stay in yours!) reducing the risk of a head-on collision.

The autonomous vehicle, using V2I information is aware of a precipitous drop off the right shoulder, indicating that there is no easement to pull over onto.

The autonomous vehicle, using wide angle IR/RADAR sensors, detects the movement of a possible obstruction before it gets to the road (seriously, we're already doing this with deer detectors). It can identify the trajectory of the object which gets added to the travel plan.

So we have a vehicle that is already traveling at a slower (safer) speed, has more knowledge about the road, the ability to immediately communicate with surrounding vehicles, and a strongly enhanced ability to identify and calculate threats to the travel plan.

To suggest that a human being would be a better driver in that situation is ludicrous. And I say that as a guy that loves driving, fast, even in snowy conditions, who has never been in an accident.

If I were driving in that situation, with all likelihood either I'm dying, or the snow boarder is dying. If the autonomous vehicle is driving, odds are my trip will take longer, but no one winds up dead.

-Rick

Re:"while not intended for production"

[neoritter]

The first option is always to break. Contrived example I think goes to you, lol.

Re:"while not intended for production"

[RyoShin]

A vehicle may decide that a pothole is actually an obstruction, or that the railroad track is the end of a road.

Very nice post. This particular line made a question pop into my head: Do we have any human-driven cars that have a companion AI "driving" a virtual car, where the AI is doing risk aversion and noting where the human differs from the AI for later review and/or machine learning?

Using your example, the vehicle "sees" a railroad track but, because of the sudden shift in terrain, thinks it's an end of the road. The AI, in its virtual car, starts applying the brakes, but the human maintains speed and keeps on going. The AI notices that the terrain has resumed expected road conditions, that they haven't crashed or fallen off a cliff, and marks that point as an event to review.

Re:"while not intended for production"

[RingDev]

I'm not aware of anything like that currently, but I'm on the other side of the equation at the moment. I work for a state DOT, so I'm seeing things primarily from the V2I perspective.

I know that years ago Cadillac had an infrared detection and HUD alert system, but I don't believe it did any analysis of it.

I would be surprised if the major autonomous vehicle players aren't looking at machine learning though. Someone had previously mentioned a vehicle with collision avoidance that would force a vehicle to stop a "safe distance" from a parking gate, making it hard for them to reach the parking ticket box. One would assume with some rather simple machine learning algorithms, combining GPS location, speed, peddle conditions, velocity changes over the last few meters, etc... that you could pretty quickly develop a set of scenarios to disable the collision deterrence system.

I wouldn't be surprised to see more short range infrastructure tools pop up as well. A gate that communicates with your car to establish license plate/ID to eliminate the need to get a ticket at all, and the ramp that can tell the car exactly where the open spots are, and where the spot closest to your desired exit is. etc...

This stuff is coming. It'll take time, but the technology exists, it just needs refinement and price drops to gain wider acceptance. And the 2017 model year will spur a massive growth of the technology.

-Rick

Re:"while not intended for production"

[RyoShin]

Interesting, thanks for the reply!

When they test these autonomous cars...

[Anonymous Coward]

Do these tests cover use cases such as encountering a flash flooded road where what used to be a road is now a river? Driving on ice? Etc.

Re:When they test these autonomous cars...

[Qzukk]

Did your driving test cover that?

Re:When they test these autonomous cars...

[ubergeek09]

That's why there's still controls to drive it non autonomously.

Re:When they test these autonomous cars...

[Russ1642]

Don't forget the all important jet engine test. (you won't get that if you're not a Far Side fan)

Re:When they test these autonomous cars...

[gstoddart]

That's why there's still controls to drive it non autonomously.

And if that is the response, that is why autonomous cars will NEVER work on public roads.

Either the car drives itself 100% of the time, or I drive it 100% of the time.

If someone thinks that you're going to be driving along in your car not paying attention to the road, and suddenly the computer is doing to say "fuck it, I don't know, you do it" they're complete morons.

Human reaction time will not allow a driver who is disengaged to suddenly be in control of the car.

There is no "sometimes the computer drives and sometimes you do", unless it's a complete transition. But when the computer is driving, I should be able to climb in the back and sleep, or read the paper.

Otherwise the entire system is doomed to fail, because it simply won't work in the real world.

I see zero value in an "autonomous" car which is periodically going to decide it isn't responsible. And I'm sure neither Benz nor Google plan on indemnifying you from legal responsibility.

In which case this will always simply be a toy.

Re:When they test these autonomous cars...

[TWX]

My driving test did not cover that, but my Cub Scout training as a young child sure did, at least as far as the flooded road/river part is concerned.

Re:When they test these autonomous cars...

[Russ1642]

The horseless carriage was a novelty at first. I mean, who wants to ride in a machine that has no mind of its own? Something that can't even avoid minor obstacles on its own?

Re:When they test these autonomous cars...

[0123456]

The horseless carriage was a novelty at first. I mean, who wants to ride in a machine that has no mind of its own? Something that can't even avoid minor obstacles on its own?

Your analogy is silly, unless it was a 'horseless' carriage that required you to pull a horse behind it just in case it refused to drive anywhere under its own steam.

Re:When they test these autonomous cars...

[vux984]

There is no "sometimes the computer drives and sometimes you do", unless it's a complete transition. But when the computer is driving, I should be able to climb in the back and sleep, or read the paper.

Agreed. This has to be the case. All the real value use cases for autonomous requires 100% autonomy and liability on the car.

From drop me off downtown and then go find a place a park on its own (empty) and then come pick me up when I call it; to dropping the kids off a school (no adult), and return home (empty).

It has to have contingency smarts to handle everything from construction and flag people sending the car down a detour to getting home and the garage door not working so it has to find some other nearby place to park to being able to deal with a mattress falling off the truck in front of it leaving an obstacle that blocks the road, to being hit by a cyclist.

Because if a driver is expected to sit in the driver seat, hands on the wheel, ready to take control, but not actually in control the entire time, then there is no reason to have an autonomous car.

Re:When they test these autonomous cars...

[0123456]

Because if a driver is expected to sit in the driver seat, hands on the wheel, ready to take control, but not actually in control the entire time, then there is no reason to have an autonomous car.

That's not entirely true. Hands-free cruise control, for example, could be useful even if you have to keep an eye on the road to check nothing disastrous is about to happen.

But, yeah, in a city, forget it. When my 'driverless car' starts sliding sideways on ice, I don't want it to suddenly tell me to take over and then blame 'human error' when I crash.

Re:When they test these autonomous cars...

[PRMan]

Stop being so extreme... I already have a car that will not automatically stop without hitting the car in front of me. Yes, in rare cases (extreme stop or someone cutting me off so close the car doesn't see it), I have to respond. But unless I see one of those two scenarios, I just cruise home while chatting hands-free with my wife. And it's wonderful.

Re:When they test these autonomous cars...

[bondsbw]

I disagree. It doesn't have to be a "sudden" transition... the transition can be in planned scenarios (like transitioning from freeway to non-freeway driving). The car could also pull to the side of the road when weather conditions become too adverse.

"Driver must be prepared to take control at any time" is a reasonable disclaimer for early self-driving cars, and may even become the law in many states.

Just because you won't buy such a car doesn't mean there are not plenty of people who would.

Re:When they test these autonomous cars...

[number17]

But unless I see one of those two scenarios, I just cruise home while chatting hands-free with my wife. And it's wonderful.

Its not extreme. If you have ever driven on the highway in a snowy climate I am sure you have driven into a section of the highway that has open fields on both sides. This causes the snow to drift onto the highway covering the lane markers. Humans would slow down and continue to drive where they believe the lane exists. If you aren't the first person then you would follow the tire tracks of the previous cars. Stopping isn't an option and would be ludicrous.

The driverless car has to deal with this scenario as it is an everyday event during winter months. The car can't give up and stop in a live lane. The car can't just hand over the wheel at a moments notice when going 100km/h. You can't cop out and say that the car can't drive in snow.

Re:When they test these autonomous cars...

[0123456]

Yeah, when we bought a new car last year, we were laughing at all the models that now come with 'lane assist' based on tracking the lane markings, because you can't see the lane markings on our roads for about nine months a year...

But, hey, I'm sure a 'driverless car' that's been tested on a highway in California will handle it just fine.

Re:When they test these autonomous cars...

[number17]

The car could also pull to the side of the road when weather conditions become too adverse.

The car would still need to be able to get to the side of the road. If it can't handle driving in a snow storm then how would it possibly pull over if you drove into a snow squall on the highway? By definition a snow squall is a sudden moderately heavy snow fall with blowing snow. Its not something that happens with any warning.

Re:When they test these autonomous cars...

[TBoon]

And if that is the response, that is why autonomous cars will NEVER work on public roads.

Either the car drives itself 100% of the time, or I drive it 100% of the time.

I would love a car capable on cruising along pre-approved rural interstates, but handing over control to me for city driving. Set your destination to a city 7 hours away, and watch a movie, read a book, or maybe even take a nap, while the car keep a more vigilante "eye" on driving than a human possibly can for any extended period of time.

Of course it should be 100% capable of driving itself when in charge, and not suddenly hand over to a human. When it needs refueling, or if the weather deteriorates it needs to notify the human well in advance, and be capable of safely pulling over and stopping if the driver fails to take over in time.

Re:When they test these autonomous cars...

[twitnutttt]

Do these tests cover use cases such as encountering a flash flooded road where what used to be a road is now a river? Driving on ice? Etc.

Drivers in California can't drive with water either, even a little bit of rain. That's why when it rains in LA the traffic incident map is like a Christmas tree. Ice? Fugghedaboutit.

Re:When they test these autonomous cars...

[Bonobo_Unknown]

>If someone thinks that you're going to be driving along in your car not paying attention to the road, and suddenly the computer is doing to say "fuck it, I don't know, you do it" they're complete morons. It sounds silly but that's exactly how autopilot and fly by wire systems work in Airbus and Boeing aircraft and they have hundreds of passengers at a time.

Re:When they test these autonomous cars...

[Bonobo_Unknown]

That's why there's still controls to drive it non autonomously.

And if that is the response, that is why autonomous cars will NEVER work on public roads.

Either the car drives itself 100% of the time, or I drive it 100% of the time.

If someone thinks that you're going to be driving along in your car not paying attention to the road, and suddenly the computer is doing to say "fuck it, I don't know, you do it" they're complete morons.

It sounds silly but that's exactly how autopilot and fly by wire systems work in Airbus and Boeing aircraft and they have hundreds of passengers at a time.

Re:When they test these autonomous cars...

[Anonymous Coward]

By definition a snow squall is a sudden moderately heavy snow fall with blowing snow. Its not something that happens with any warning.

It's not something that happens with any warning that you can notice. For a computer with a suite of superhuman senses, real-time access to weather satellite data, and no capacity for distraction, there's plenty of warning.

And that's assuming the car isn't programmed to simply refuse to drive itself in dangerous weather in the first place. If you can look up your route on the Weather Channel and say "Nope, not doing this now", so can your car.

Re:When they test these autonomous cars...

[0123456]

You're apparently an idiot who's getting all hyped up about cars that can 'drive themselves' in extremely constrained conditions that bear no resemblance to real life for most of the drivers in the world.

Yes, one day there'll be driverless cars on the road, capable of handling any conditions without human intervention. But that day won't be any year soon.

Re:When they test these autonomous cars...

[0123456]

It sounds silly but that's exactly how autopilot and fly by wire systems work in Airbus and Boeing aircraft and they have hundreds of passengers at a time.

And pilots, with far more training and much more time available to determine what's happening and correct it... fly a perfectly good aircraft into the sea.

Autopilots have proven again and again that 'just dump the problem in the human's lap when the computer doesn't know what to do' is disastrous.

Re:When they test these autonomous cars...

[Bonobo_Unknown]

Just because it's been proven to be disastrous doesn't mean that it won't be the precedent that makes it okay in automobiles. I mean don't assume legislators will always make the most sane decisions.

Re:When they test these autonomous cars...

[R3d M3rcury]

Uh...weather information isn't quite that pervasive. That said, I don't think we even need to go THAT far.

Suppose you're driving along and you suddenly get hit by torrential snow or rain. What would you do?

Well, of course, the first thing you do is slow down. "Never outdrive your senses," to coin a phrase. So if rain or snow is affecting your visibility, you slow down. Why wouldn't a self-driving car do the same, while flashing a "Please take over driving" alert?

To me, a self-driving car is essentially a really clever cruise control. Which means it's great for the highway but not so good for crowded city streets. I don't see too many people using cruise control on crowded city streets and I'm pretty sure you won't see too many people using auto-drive in the same way.

I'll admit, my daily commute is 56 miles, of which 52 of them are on the freeway. I would buy a self-driving car in a heartbeat. I'd drive it to freeway and then say, "Okay, take me to my exit" and let the car drive the freeway with the stop and go traffic. When it got to my exit, I'd take over and drive the rest of the way home.

Re:When they test these autonomous cars...

[Firethorn]

But, yeah, in a city, forget it. When my 'driverless car' starts sliding sideways on ice, I don't want it to suddenly tell me to take over and then blame 'human error' when I crash.

I have the feeling that the standard would be 'I can't handle this, pull over to a safe stop; if you want to continue before conditions improve you do it' rather than just abruptly handing control over to you.

Indeed, it's my understanding that the collision avoidance detectors in many of these new vehicles are always on. So if you fall asleep or whatever it'll hit the brakes before you ram into something.

Re:When they test these autonomous cars...

[Firethorn]

It sounds silly but that's exactly how autopilot and fly by wire systems work in Airbus and Boeing aircraft and they have hundreds of passengers at a time.

Actually, modern aircraft have collision avoidance systems. While Boeing planes tend to default to 'trust the pilot', Airbus's planes will act to avoid collision no matter what the pilot does.

Re:When they test these autonomous cars...

[aberglas]

+1. Brilliant. You have the answer.

Dirverless cars just need somebody to walk ten paces in front of them ringing a bell. That will make them safe.

(Realistically, cars that can "only" drive down freeways on sunny days will be very useful and here very soon. Despite all the recalcitrants on slash dots.)

Re:When they test these autonomous cars...

[Anonymous Coward]

The horseless carriage was a novelty at first. I mean, who wants to ride in a machine that has no mind of its own? Something that can't even avoid minor obstacles on its own?

Clearly the solution is to allow cars to be driven by horses.

Re:When they test these autonomous cars...

[mjwx]

Stop being so extreme... I already have a car that will not automatically stop without hitting the car in front of me. Yes, in rare cases (extreme stop or someone cutting me off so close the car doesn't see it), I have to respond. But unless I see one of those two scenarios, I just cruise home while chatting hands-free with my wife. And it's wonderful.

And this is what the driver of the Ford Kuga (which has Automatic Emergency Braking) thought before she rammed into the back of my stationary Honda Integra.

A perfectly good sports car written off by some moron on the phone who thought technology would magically save her from harm.

Also there are more than two scenarios that require you to brake, almost all of them not programmed into AEB.

Autonomous cars are better at driving than you not because they're good but because you're such a terrible driver.

Re:When they test these autonomous cars...

[swillden]

To me, a self-driving car is essentially a really clever cruise control. Which means it's great for the highway but not so good for crowded city streets.

It's worth noting that Google's current self-driving car platform is intended primarily for use on crowded city streets. They don't go very fast, and so aren't intended for highway driving at all. They do an outstanding job of dealing with pedestrians, cyclists and other automobiles.

They don't handle bad weather all that well yet, though.

Re:When they test these autonomous cars...

[coofercat]

...or how about the car works all by itself on major roads and highways. On crappy side roads that aren't properly documented in the GPS maps, let alone suitably paved, lit or maintained, it says "I'm sorry, I can't drive down this road, but you can". A solution like this would mean 80% of all journeys would be 100% autonomous and it would mean the vast majority of accidents would be avoided, as considerably few accidents happen on shitty side roads than they do on the major ones.

I doubt anyone would accept a car that might suddenly say "eek, I can't do this - please take over" while driving 50 down a windy road in the rain. I think we'd all expect it to find a safe place to stop and let us take over from there though.

Where does the hydrogen come from??

[Tokolosh]

Said it first!

Re:Where does the hydrogen come from??

[ShanghaiBill]

The nuclei were formed in the first few seconds after the big bang. Molecular hydrogen didn't form until about 380,000 years later, after the photon epoch.

Re:Where does the hydrogen come from??

[houstonbofh]

"Where does the hydrogen come from??"

Water, and dreams...

Feel good technology

[Dorianny]

The majority of hydrogen is produced from fossil fuels by steam reforming or partial oxidation of methane and coal gasification. Not exactly zero-emissions technology.

Re:Feel good technology

[bmajik]

The majority of electricity is produced by doing something that wrecks the planet somehow, according to somebody.

The point is that electricity is an interesting way to power a car because we can think of more ways of making electricity - both now and in the future - than we can of making gasoline.

The same is true of hydrogen. For instance, you can make hydrogen via electrolysis. If you are somewhere with abundant, cheap, clean electricity - like Iceland - then dumping off-peak electricity into electrolosys so you can store H-gas might be an interesting process.

In fact, I saw Shell stations in Iceland that were providing hydrogen refueling in the mid 2000s. I have no idea where they were getting the Hydrogen from, but experimenting with alternative energy/power production systems is a good idea.

Re:Feel good technology

[TWX]

That's not the only place that Hydrogen can come from though. There have been ways to crack Hydrogen from water using non-electric solar tech. If the production of Hydrogen from byproducts of fossil-fuel refiining for other purposes means less waste or allows for use of a technology as a migrating tech, good.

Re:Feel good technology

[ShanghaiBill]

The majority of hydrogen is produced from fossil fuels by steam reforming or partial oxidation of methane and coal gasification. Not exactly zero-emissions technology.

Hydrogen is mostly made from natural gas, with is cheap, plentiful (at least in North America), and domestically produced. The NG->H2->Fuel-Cell process is way more efficient than a gasoline ICE.

Re:Feel good technology

[houstonbofh]

But it is emissions in someone else's backyard, so it doesn't count...

Re:Feel good technology

[fustakrakich]

I am willing to sacrifice 'efficiency' for more flexible fuel choices. The infrastructure for hydrogen will always be far too fragile to be practical, unless somebody invents a portable home distillery version to produce the fuel, but even then there's too much black box technology involved. Diesels and jets running on algae based fuels is probably the path of least resistance from a practicality viewpoint. The typical short term thinking bean counters have a different opinion.

Re:Feel good technology

[drinkypoo]

That's not the only place that Hydrogen can come from though. There have been ways to crack Hydrogen from water using non-electric solar tech.

I hope you're planning to precede them with a distiller, because if you're putting anything other than clean water into that system, you're going to both be creating wastes and destroying electrodes, creating more waste. Now how much does your system cost to maintain? It's already spectacularly inefficient.

Hello Micheal

[portwojc]

They should have had David Hasselhoff as the driver.

Re:Mirror Inside

[TWX]

It looks like a minivan and the Trimaxion Drone Ship from Flight of the Navigator had a drunken night together...

Re:Getting more reasonable

[TWX]

The over/under for range for me is 150 miles. That number is based on the sheer size of my municipality and a round-trip to the furthest destination that I routinely visit, plus 50%. It's also about half of the range that a single tank of gasoline will allow cars to reach in city/mixed driving.

Not Content To Merely Drive Like an Asshole

[l0ungeb0y]

Now I shall a rear bumper with an LED display to tell other drivers what to do! Now excuse me while I execute 1000 consecutive lane changes in rush hour traffic -- I've got important places to be

Re:Not Content To Merely Drive Like an Asshole

[drinkypoo]

Now I shall a rear bumper

You accidentally the whole thing

with an LED display to tell other drivers what to do! Now excuse me while I execute 1000 consecutive lane changes in rush hour traffic -- I've got important places to be

There's no room to make 1000 lane changes if people are paying attention. Which self-driving cars will do, unlike the typical jerkoffs driving about.

Need the Concept Bus

[gurps_npc]

Concepts cars are worthless, most never get made.

You want to convince me you are SERIOUS about getting into the driverless car? Then build a Concept Bus - or Concept Garbage Truck.

Those are large vehicles that honestly do not need drivers. They are expect to drive slow, not fast and usually travel set routes. Small cities can easily afford to self-insure them, and they won't have to worry quite so much about the stupid technology ignorant laws, as they will be purchased by the people that enforce, if not write the laws. Finally they are already expensive and the cities pay large salaries to people to drive them.

They will in all probability be the very first driverless vehicles we actually see on the road [as soon as we 1) convince the unions to let us and 2) actually get them to work.]

So forget about concept 'cars' and show me a concept bus or concept garbage truck.

Re:Need the Concept Bus

[Toshito]

I don't know where you live but here in Canada garbage trucks are run by private companies, the contract being awarded by the city to the lowest bidder.

So good luck having those private companies invest in anything else than trucks that are crapped out and barely legal on the road, and the lowest wage they can give to any driver which has a driving permit.

Re:Need the Concept Bus

[gurps_npc]

I live in New York City. The city of New York owns over 2,000 collection trucks, not to mention street sweepers, salt spreaders, front end loaders, and support vehicles. Yes, we subcontract out some of the work, but a large mega city like NYC owns and operates it's own trucks. Also, last september, Chicago owns 600 trucks. Both of us are unionized.

But in fact, a private company is a lot EAISER to convince to buy such a device, because they are more focused on the bottom line, rather than appealing to Local 831, USA of the Teamsters (a worthy organization, but they have to put their own interests above those of the city - that's what their members pay them to do.)

A new Garbage Truck can easily cast $500,000. They are large and expensive devices and they typically make over $40,000 a year. For a system that costs $30,000, a small company would easily come out ahead after just one year. If it costs $100k, it takes 3 years.

(Note, you still need employees to LOAD the garbage, you just need 2 men per truck as opposed to 3).

Re:Need the Concept Bus

[colesw]

Really, 3 people per truck? I remember in the 90's seeing 2 people per truck here in Ottawa. The last time I remember seeing a garbage truck it only had 1 person, he would drive on the right hand side, and hop down to throw the garbage in, then hop back in.

Re:Need the Concept Bus

[Harlequin80]

More than 1 person? The driver gets out of the truck? I haven't seen that here in 20 years! Here the truck drives along the street, has a big mechanical arm that reaches out, grabs the bin and empties it into the truck.

As I said it has been like that for at least 20 years.

Re:Need the Concept Bus

[radarskiy]

Komatsu Autonomous Haulage System: https://www.youtube.com/watch?...

Re:Need the Concept Bus

[mjwx]

(Note, you still need employees to LOAD the garbage, you just need 2 men per truck as opposed to 3).

The loading of garbage onto trucks was something Australia automated well over a decade ago.

In fact, seeing a person lifting a garbage bin into a truck is a rare sight, so rare that when I went overseas to Thailand last week, I was surprised to see actual garbagemen.

Re:Need the Concept Bus

[swillden]

(Note, you still need employees to LOAD the garbage, you just need 2 men per truck as opposed to 3).

All the trucks in my area are automated, meaning there's a hydraulic arm that picks up the cans and dumps them. They're slower and more expensive than non-automated trucks, but require only a single operator (the driver). I have a brother-in-law who drives such a truck and in discussions with him I've become skeptical that they'll replace the driver in such trucks any time soon, because he encounters enough bizarre situations on a regular basis that having a human on board is essential.

In situations where the garbage is thrown by hand, I can see the driver's job being automated away. The loaders would be around to deal with the unexpected.

Re:Need the Concept Bus

[AmiMoJo]

The problem is that for the foreseeable future driverless cars will still require human occupants at all times. Otherwise it would be pretty easy to screw with them, from a simple "denial of service" attack where you pen it in with traffic cones for lulz to hijacking driveless long haul goods vehicles. Sometimes the computer just can't figure out what to do and needs help.

Basically, if something goes wrong a human needs to step in, often immediately. Eventually this may change, but notice how even this concept car still has controls for a human driver.

Re:Need the Concept Bus

[RyoShin]

Plus, with buses, they can make smaller buses that have more routes and can also respond to demand to better stick to schedules. (If 13 people "check in" at Bus Stop B but no one at Bus Stop A, then the bus can take a shortcut that avoids A and goes to B sooner, to better handle the larger amount of people waiting to get on.)

1) convince the unions to let us

Even if the tech is viewed as very mature by every automotive professional, Average Joe will still view it with heavy skepticism. Having a human who could take control in the event of a bad situation will alleviate a lot of concern, regardless if the human could not realistically do anything useful even with very early warnings and a sharp eye. Public trust will probably grow at the same rate as bus drivers quit/retire, which brings us to the union solution: Attrition. Everyone keeps their job and rate but does less. Make them drive the bus in and out of the storage facility to keep their driving skills up and give them busy-work. No new blood, raises are probably capped off, but the drivers can keep working until they quit/retire or a specific amount of time (like 5-10 years) has passed.

While I've not dealt with a union directly (only felt some indirect effects, like not being able to move my own desk), I think that would be acceptable. As an added (evil) bonus, the driver can be a scapegoat if a bad accident occurs.

(I wonder if buggy whip makers had a union that had to deal with this...)

Re:Need the Concept Bus

[lsatenstein]

Concepts cars are worthless, most never get made.

You want to convince me you are SERIOUS about getting into the driverless car? Then build a Concept Bus - or Concept Garbage Truck.

Those are large vehicles that honestly do not need drivers. They are expect to drive slow, not fast and usually travel set routes. Small cities can easily afford to self-insure them, and they won't have to worry quite so much about the stupid technology ignorant laws, as they will be purchased by the people that enforce, if not write the laws. Finally they are already expensive and the cities pay large salaries to people to drive them.

They will in all probability be the very first driverless vehicles we actually see on the road [as soon as we 1) convince the unions to let us and 2) actually get them to work.]

So forget about concept 'cars' and show me a concept bus or concept garbage truck.

Regarding unions. Their raison-d'être is job creation, protection, and union dues. The problem that many municipalities have is union agreements and idiotic outdated rules. For example of such a rule is , "a driver drives, and does not get out of the truck to pick up or deliver". Another example, "The electrician is responsible for the breakers. We had a swimming pool front door attendant plug in a kettle from home, and with it and the other electrical stuff on the circuit, caused the breaker to trip. The attendant unplugged the kettle, but had to call the electrician to reset the breaker. (Job protection). Gosh, we stopped using glass fuses eons ago.

In small communities, with the introduction of vehicle automation, the driver can be outside the vehicle, can be helping the crew, and when the truck has to be displaced a few feet, can use a remote control to advance the vehicle. Nothing wrong for the gardener or plumber or electrician to handle several kinds of jobs.

Had a pleasure to see early self-driving footage

[quax]

That was in 1997 when I worked at what later became the KIT.

Back then they tested an early artificial neural net controller under real life conditions on the Autobahn A8. The driver just sat with his arms folded behind the wheel.

This technology has been a long time coming and still lawmakers haven't caught on to it.

Re:Mirror Inside

[nospam007]

"Any airbags would have to be integrated into the seats."

I have 3 airbags in my seat right now and I don't even have a Benz.

Oxymoron

[wonkey_monkey]

Mercedes-Benz's Self-Driving Concept Car Is Here

When are concept cars every really "here," in any practical sense of the term?

Re:Oxymoron

[aitikin]

Mercedes-Benz's Self-Driving Concept Car Is Here

When are concept cars every really "here," in any practical sense of the term?

About 3-4 years in the case of the NSX, but that's definitely the exception, not the rule.

omfg, take the hint!

[ihtoit]

NOBODY wants a self driving car!

But then the second user joined

[raymorris]

> There was a time when Slashdot was full of people who'd think more than two seconds about their brave new economic ideas, rather than just demanding a pony.

Yep, that time was October 5th, 1997. Then the second user joined the site, and the idiocy began.

Mine did. Denver in December

[raymorris]

Icy roads? Yeah, when I was 16 I took my driving test in Denver Colorado, in December. So several feet of snow on each side of the road and plenty of ice around. Come to think of it, that was kind of dumb.

Re:Elon Musk made a comment about this

[prefec2]

The car is a technology demonstrator. However, they plan to sell fuel-cell cars in 2017 and Toyota wants to sell them 2020. Depending on the availability of hydrogen filling stations this technology could be near you in a decade. The interesting thing of the concept car is that they combined hydrogen fuel-cell and batteries allowing to bridge the gap between hydrogen filling stations with the battery. In addition they had their self driving technology already integrated in a normal E or S class car which is able to drive around on some roads in Germany.

Concept cars are there to show (and show off) what is technological possible. Normally the same technology ends up in the real thing.

Hopefully this still gets read...

[Bourdain]

To address the two common themes I see here:

(1) Make no mistake, semi-autonomous cars are useless, but meaningful collision avoidance systems are useful and that's the first stepping stone in the process

(2) Autonomous cars are still decades away from any sort of real adoption and automobile manufacturers should (I suspect they are...) develop them in the context of a shared usage vehicle given their much higher utilization than a regular car (an autonomous car could be in use 100% of the time as opposed to how most cars sit parked most of their time). As such, most users of such vehicles will use them like taxi's but they would cost much less, be safer, and would be available anywhere and for any length trip unlike just metropolitan areas.

Re:Had a pleasure to see early self-driving footag

[pipedwho]

And before your KIT, there was the KITT. It even had a voice synthesiser that sounded just like that guy from St Elsewhere. I'm still waiting for Trans-Am to release that model.

Re:Had a pleasure to see early self-driving footag

[quax]

KITT, not invented in Germany but certainly embraced like nowhere else by an entire generation of German pre-teen boys.

Probably lead directly to this prototype.

Hello, it emits water and energy.

[thisisauniqueid]

Just sayin'.

Flying Car

[abhishek765]

Wait for some more years, our Engineers will be ready to surprise us with a Flying car.

self-driving car vs auto-pilot plane

[dslmodem]

auto-pilot plane was available several decades ago because the complexity of flying is significantly lower than that of driving. if i can, i would bet on auto-pilot flying machine.

Re:raymorris, ac replies from you fool yourself

[badkarmadayaccount]

Get a girlfriend dude. And medication.