Fooling a Mercedes Into Autonomous Driving With a Soda Can

New submitter Petrut Malaescu writes: Last year Mercedes introduced an intelligent Lane Assist system to its S-class, which is cataloged as a Level 1 "Function-specific Automation" system. In other words, hands and feet must always be on the controls. But a clever driver discovered that all it takes to keep the car in Lane Assist mode is a soda can taped to the steering wheel. It's enough to trigger the steering wheel sensor that's supposed to detect the driver's hands. Obviously, it's not a good idea to try this on a busy highway.

Comments

S-class driver with a soda can? Please...

[damn_registrars]

This must have been discovered by a Benz mechanic. Soda cans are far too proletariat for S-class owners.

This isn't new

[Anonymous Coward]

Ferris Bueller tricked a car into "autonomous" mode by putting a cement block on the accelerator--a sensor that is used to detect the pressure from a foot.

Sensors can be deliberately fooled with inanimate objects. News at 11.

Re:This isn't new

[ericloewe]

Who's to say they're not actually alive? At least a brick will never threaten to hurt you and, in fact, cannot speak.

Re:This isn't new

[Kjella]

Sensors can be deliberately fooled with inanimate objects. Film at 11.

FTFY

Hands and feet?

[swb]

What exactly is this automating? The whole point of cruise control is to not require your feet on the pedals.

My Volvo has distance sensing cruise control. It won't hold the lane for me but it doesn't turn off cruise when I take my hands off the wheel, either.

Re:Hands and feet?

[Anonymous Coward]

The Mercedes system drives the car for you, under heavy traffic conditions https://www.youtube.com/watch?v=AihC5flC-38

Re:Hands and feet?

[Anonymous Coward]

Its not cruise control, it's Lane Assist. It keeps you in your lane, essentially steering for you.

Re:Hands and feet?

[AvitarX]

That video claims to have nearly the same sensors, but doesn't seem be at all like the system mentioned.

Re:Hands and feet?

[tlhIngan]

What exactly is this automating? The whole point of cruise control is to not require your feet on the pedals.

My Volvo has distance sensing cruise control. It won't hold the lane for me but it doesn't turn off cruise when I take my hands off the wheel, either.

It's really more about how a modern car can actually "drive itself" in a limited way. It' snot a full autonomous car, but with what we have right now today, it's actually impressive.

Then again, I suspect he got the idea from a Hyundai commercial where a bunch of stunt drivers set up their vehicles and then exit them and having the cars drive around by themselves following a lead vehicle (still driven, of course).

Re:Hands and feet?

[AvitarX]

I'm quite surprised the automated driving will keep that close of a following distance (less than 1 second)

This looks like a ticket risk to me.

Re:Hands and feet?

[KamikazeSquid]

A computerized system can respond to a sudden stop much more quickly than a human driver. As human drivers, most of the distance we need to keep between our vehicle and the next is due to our delayed reaction time, and not the time it takes to actually stop the vehicle once the brakes are applied.

Re:Hands and feet?

[mjm1231]

Most cars have a braking distance of a bit over 100 feet or so at 50 MPH. If the car in front of you comes to a sudden stop by crashing into a stopped truck, a faster reaction time won't help.

Re:Hands and feet?

[AvitarX]

I said a ticket risk, not an accident one.

I'm assuming that because this system constantly requires human monitoring, the rules of a human driver apply.

The ticket risk is a cop giving you a following too close, because it was entirely too close.

Or was this ad with something not the same as what the production cars get?

Re:Hands and feet?

[m3000]

Which is fine until the car in front slams into a broke down semi trailer and the sliding distance turns to zero.

Re:Hands and feet?

[amicusNYCL]

Active Lane Assist is where the car corrects the steering to keep it in the lane. In the video in the article he looks like he's just going straight, but if you watch the wheel closely it will turn occasionally to keep the car in the lane. The driver never has his hands on the wheel. The Mercedes will use radar to sense and maintain the distance to the vehicle in front (not the best around motorcycles, or when going around turns), and lane assist will keep the car in the lane as well.

Re:Hands and feet?

[Maxo-Texas]

I wish that were the case in the states.

Here, the police no longer ticket for even aggressive tailgating.

Drivers here get angry if they are within 10 feet behind you and you don't speed up (and you are already over the speed limit by 5mph). And what you really need to do is to slow down to account for the extra stopping distance you need to keep them from plowing into you.

There are a lot of basic rules of the road that the police used to enforce and which made everyone better drivers.

Now, the police seem to focus only on high revenue tickets-- even blatantly showing up regularly 3-4 days at the end of the month to meet their quotas on certain stretches of the highway.

Re:Hands and feet?

[gnasher719]

Here, the police no longer ticket for even aggressive tailgating.

Interestingly, in Germany they have automatic cameras that don't measure your speed, but your distance from the car in front of you. The distance of 10 feet you mention is three meters. At 60 km/h (about 40 mph), the recommended distance is 30 meters. Less than 10 percent of the recommended distance is a major fine. At a higher speed (4 meters distance at 160 km/h = 100 mph), it has been ruled that a driver tipping on his brakes to turn the brake lights on, causing the following car to slam his brakes and crash, was acting in self defense.

Re:Hands and feet?

[Lehk228]

those systems demonstrated are not for automated driving, they are built-in safeties which help prevent accidents when the driver is (literally) asleep at the switch.

Re:Hands and feet?

[Neil Boekend]

I would love such cameras here.

Obvious

[crow]

They've had adaptive cruise control for a long time now that will slow you down so that you don't rear-end anyone in front of you. In theory, you can set it at your favorite speed, and then ignore the foot pedals until you reach your exit. I haven't used it, so I don't know if it handles stop-and-go traffic jams or things like that.

Now they have automatic lane centering. The car uses cameras to read the paint stripes and keep it centered in the lane. Because it's not a general system for autonomous driving (and the obvious liability if it crashes), it shuts off if you let go of the steering wheel.

Combine the two, and you have fully autonomous highway driving under regular conditions. You just have to fool the sensor, and sensors are easy to fool.

What's interesting is to learn what conditions it won't handle.

Re:Obvious

[OzPeter]

Combine the two, and you have fully autonomous highway driving under regular conditions. You just have to fool the sensor, and sensors are easy to fool.

Yeah, but I'd be worried that the cruise control would punt the car into a corner at a rate at which the lane centering couldn't compensate. You really need a bit more smarts for simple autonomous driving scenarios.

Re:Obvious

[jklovanc]

conditions it won't handle
- changing lanes
- passing
- letting other drivers in
- avoiding debris on the road
- entering/leaving highway
- dealing with construction zones
- avoiding reckless drivers
- lane selection

Lane keeping and distance keeping are only a small part of driving.

Re:Obvious

[wisnoskij]

"fully autonomous highway driving under regular conditions."

I would argue it very much is not that. Autonomous driving is so so so much more than just not ramming into the car in front of you and not changing lanes. If it is 100% unable to react to what is going on to the sides and behind it is just slightly better cruise control.

Re:Obvious

[boristdog]

Unfortunately, lane keeping and distance keeping are skills that elude a lot of drivers.

Re:Obvious

[PRMan]

"They've had adaptive cruise control for a long time now that will slow you down so that you don't rear-end anyone in front of you. In theory, you can set it at your favorite speed, and then ignore the foot pedals until you reach your exit. I haven't used it, so I don't know if it handles stop-and-go traffic jams or things like that."

I have a 2014 CLA and it works. I have gotten on a freeway, set it to 80 and never touched the pedals for over 50 miles.

As far as what it won't handle, my car won't handle extreme braking, getting cut off badly or a car that is stopped completely (doesn't see it at all). Other than that, even in slow and go driving it works perfectly (if it stops completely you have to tap the gas to go again).

Re:Obvious

[PRMan]

Actually, I have the adaptive cruise control without the lane keeping and it knows to slow down when cornering.

Re:Obvious

[PRMan]

Actually, it handles letting other drivers in as long as they don't cut you off dangerously. It's actually quite good as slowing for reckless drivers as well (say they cut across 3 lanes in front of you, the car will slow down appropriately).

Re:Obvious

[crow]

So most of those won't be a problem when driving between cities. It's probably not great for daily commuters, but it's probably a lot safer than a sleepy driver on a rural highway.

Re:Obvious

[bill_mcgonigle]

or a car that is stopped completely (doesn't see it at all)

Ouch. This is rare, but I've seen it.

I'd be afraid if I was on a 50-mile stretch without having to think about speed my mind would wander, and I wouldn't notice this stopped car.

I'm the guy who never uses cruise control unless it's flat and empty for as far as the eye can see, though, so maybe I'm atypical.

Re:Obvious

[Em Adespoton]

You forgot -pulling off to the side of the road to let emergency vehicles pass

Re:Obvious

[parkinglot777]

Will it work on I-94 near downtown Chicago, where the lane markers have long been worn off?

If you read TFA (or look at the quoted below), you would be able to answer your own question. However this would require you to define the word wear off in your own definition...

Normally, ALA requires you to put a hand on the wheel after a certain amount of time, otherwise the system disengages. And it only works when lane markings are clear and conditions are clear enough for the sensors to see the road.

Re:Obvious

[mrchaotica]

Because it sees the corner coming or because it detects you turning the steering wheel?

Re:Obvious

[powerlord]

since he said "it slows down when cornering" it could be that a sensor detects the centripetal force of the cornering?

Re:Obvious

[rogoshen1]

exactly that, it seems rather dangerous to be on 'standby' for the better part of an hour, and still need to quickly to a deer or something wandering into the freeway.

Re:Obvious

[jklovanc]

Things that this can not handle in a rural setting;
- lack of lane marker on right
- animals on road
- people turning left from the opposite direction
- people entering from side roads
- intersections
- distinguishing between lane markers and tar strips
- lack of center line
- debris,gravel on road
- potholes
- speed changes due to corners
- narrow bridges

Rural roads are even harder to deal with than freeways.

Re:Obvious

[jklovanc]

Actually, it handles letting other drivers in as long as they don't cut you off dangerously.

Bumper to bumper traffic?

Re:Obvious

[OglinTatas]

What's interesting is to learn what conditions it won't handle.

The post-mortems may shed some light on that.

Re:Obvious

[crow]

I'm more thinking rural freeways like you have in the West. As long as you check for construction first and don't get unlucky with a deer, you're probably fine unless the paint goes wrong (as may be the case in post-construction sites).

Actually, there already are automatic braking systems for things like deer, and I would guess that that would be included.

One big point here is that we're a lot closer to autonomous driving that most people think.

Re:Obvious

[Ozymandias_KoK]

Pretty much any of the ones that would actually constitute autonomy, seems like.

Re:Obvious

[m3000]

This is why I'm frustrated we can't have a semi-autonomous car for highway driving already. 99% of long distance driving is just not hitting the car in front of you and staying in your lane, and having the car replace hours of monotonous highway driving would be awesome.

Re:Obvious

[Anonymous Coward]

I think we are further from autonomous cars than most people think. We may have 99% of cases covered but you can't just leave the 1% of cases that we haven't figured out off. In rush hour traffic you probably encounter a "situation" every few minutes. If only 1% of the time the car can't figure it out you will be in a wreck about every two weeks if your commute is 30 minutes. At a .5% improper situation handling rate you would be in a wreck approximately every month. To get down to once every 5 years (maybe acceptable to me?) the car can only screw up 0.0001% of "situations" encountered.

We are nowhere near this kind of precision.

Re:Obvious

[lgw]

These systems are actually quite good at some of your list - you might surprised. What they can't do at all is predict the insanity of other drivers. Like the guy waiting to turn left who will just sit there until you get dangerously close, and then cross in front of you (why do so many people do that?). It's early days yet, but I fully expect software to pass average human driving skill in my lifetime.

Re:Obvious

[nospam007]

@Now they have automatic lane centering. The car uses cameras to read the paint stripes and keep it centered in the lane."

I already see the next article:

Youngsters 'hack' the street by spray painting lanes into the abyss to fool Mecedes S cars.

Re:Obvious

[jklovanc]

These systems are actually quite good at some of your list - you might surprised.

Remember the subject of this conversation; lane following and interval maintenance. While that assist driving they are very far from autonomous driving. Lane following is simple in that it uses two painted lines to figure out where the lane is and steers to stay between the lines. It does not figure out if the line curve s ahead and there needs to be a speed reduction to deal with it. Interval maintenance is simple because all it does is puts on the brakes if the interval gets below a minimum. Say you approaching a narrow bridge. The bridge has to be identified. How can you identify a bridge if all the information you have is the position of the left side of the lane, the position of the right side of the lane and the distance to the vehicle in front of you?

Re:Obvious

[steelfood]

What's interesting is to learn what conditions it won't handle.

When there are poor or no lane markers, especially when there's no double yellow in a two-lane, two-way local highway. Or when construction's shifted the lanes away from their original positions and the old lane markers haven't been erased so cleanly. Or when there are periodic potholes the size of half-basketballs in the most-used tire lanes (tire lanes being the path your car's tires take). Or when the lane is both narrow with inches to spare on either side, and shifts suddenly, and there's a H2 up ahead in the other lane going at half the speed limit. Or when a 45MPH highway has a sudden 20MPH curve, and the lanes are narrow to boot.

This kind of autonomous driving may work when both road and weather conditions are ideal, but something a little smarter would be necessary for even slightly harsher situations. In the extreme case, a lot of driving under extremely limited visibility is basically a high-stakes game of follow the leader. Essentially, it's not enough to be able to perceive the environment; driving under those conditions requires perceiving the actions of other like actors (and relying on the assumption that those actors are sane).

Re:Obvious

[mrprogrammerman]

I wonder how it deals with construction or snow. What happens when the lines aren't there?

Re:Obvious

[lgw]

. Lane following is simple in that it uses two painted lines to figure out where the lane is and steers to stay between the lines.

My car does much better than that. I've been surprised at how little visual information it needs to determine where the lane is. I does sometimes get confused by zebra crossings, however. It doesn't brake for curves, but it does look ahead and understand curves - if the car "ahead" of me is actually in a different lane, for example, it figures that out and doesn't panic (the first gen system from 10 years ago had problems with that).

Say you approaching a narrow bridge. The bridge has to be identified. How can you identify a bridge if all the information you have is the position of the left side of the lane, the position of the right side of the lane and the distance to the vehicle in front of you?

My car has a variety of sensors, including a camera built into the rearview mirror assembly (so, better visibility than my eyes). It lacks the software to deal with e.g. sharp curves ahead, but the raw data is already available.

Re:Obvious

[jklovanc]

My car does much better than that.

What else does you car do?

but it does look ahead and understand curves

What does it do when it "understands a curve"?

but the raw data is already available.

The fact that the raw data is there and the computer having the ability to interpret that raw data in any meaningful way is oceans apart. Comparing lane following to object identification is like comparing the game of x's and o's to chess. The complexity of the problems are several orders of magnitude apart.

Re:Obvious

[Bob_Geldof]

I have a 2014 Subaru Forester with EyeSight (their adaptive (stereoscopic based) cruise control with lane departure warning system). The adaptive cruise control is amazing. I like using it in traffic on I-5 in Seattle (I said Subaru, didn't I?) and it works really well at dealing with people cutting in/stopped cars in front of me. My one complaint is when you come to a complete stop it times out and disables the system. You now have a vehicle that is idling in drive. There is an audible beep to warn you that the system is turning off, but I think I would prefer the ability to control the time out period to something more than 5 seconds.

Re:Obvious

[Andrio]

Adaptive cruise control + lane assist = poor man's self driving car :)

Re:Obvious

[redback]

sounds annoying.

nowhere I would drive with cruise on has corners tight enough to slow down for

Re:Obvious

[umafuckit]

I agree with you, although in many places in the US there is an over-abundance of stop signs. Many stop signs should actually be a yield. It's also damn stupid placing a 4-way stop sign on a larger road, where it crosses with a more minor road. The more minor road should always yield to the larger road. Either that or add a roundabout. Other countries do this and it works much better: traffic flows faster, everyone knows who has the right of way, and people on the more major road don't have to be on the look out for randomly placed stop signs.

Re:Obvious

[StikyPad]

I'm the guy who never uses cruise control unless it's flat and empty for as far as the eye can see

I don't think you're atypical at all, but... why? CC guards against unintended acceleration, as well as unintended deceleration (a phenomenon I call "tidal lock" or "flocking" with the cars next to us, depending how nerdy my conversation partner is.) Unintentional acceleration risks a ticket, while unintentional deceleration causes traffic jams (not to mention adding time to your trip). With CC, I can spend less time monitoring my own speed and more time looking at the road. The fact that I had to acclimate to using the hand controls was a minor inconvenience for the benefits achieved.

FWIW, I live in a large metropolitan area with heavy traffic, and I'm not the guy anyone is waiting to pass -- at least not for more than a few seconds when I get over to let them be my rabbit.

Boo

[timeOday]

Cars are generally not designed to be resistant to 'hacking' by their owner/operators, and should not be. Yes, you can drive without a seatbelt if you snip the little blue wire. You can disconnect your airbags. You can cause your tires to explode just by letting out most of the air and driving on the freeway.

Presenting this as some sort of coup fosters the notion that he system ought to be idiot-proof. No sudo rm -fR / for you! We'll put a thousand annoying and ultimately useless obstacles in the way to doing any little thing!

Don't blame the car for not protecting itself from you.

Re:Boo

[bluefoxlucid]

I've driven over 800 miles across six months on a tire that was completely flat, on the front wheels of a front wheel drive car, and then put air in it and driven off. It wasn't a run-flat, and wasn't inflated between. It was a Dunlop Signature Sport stock dealer tire (never buy these! They suck!).

I drove from Baltimore to DC and back with a rear tire flat the whole time (a Goodyear Assurance TripleTred, something actually useful), and then put air in it when I noticed it was flat. That's like 300 miles in one day.

I had a tire explode on me once. It wasn't low, and hadn't been thusly abused; it had about 12,000 miles on it. I didn't realize it had exploded; I felt the car start to go thump-thump-thump and knew one of the tires probably had gone flat or something, so pulled off the expressway and found four pieces of tire loosely held together by some sort of nylon mesh wrapped round my wheel. Apparently my car doesn't go spinning out of control when the front passenger tire explodes at 80mph, either. I fucking love this car.

As far as I can tell, tires just blow up when they feel like it. Ridiculous abuse hasn't failed my tires, but normal driving with 35-40psi in a 50psi rated tire has.

Re:Boo

[i kan reed]

Your comment has enlightened me.

I've resolved to take all American drivers out of my sudoers file.

Re:Boo

[silfen]

Cars are generally not designed to be resistant to 'hacking' by their owner/operators, and should not be

How low we have fallen. People used to build entire cars, and they used to maintain and repair their cars too. And now people get pushed out of shape about, gasp, using a lane following system to follow lanes!

Re:Boo

[drinkypoo]

I drove from Baltimore to DC and back with a rear tire flat the whole time (a Goodyear Assurance TripleTred, something actually useful), and then put air in it when I noticed it was flat.

How do you know how long it was "flat" before you noticed it?

Re:Boo

[cpotoso]

I once put 70 psi on a tire rated for 45 (and designed to be inflated at 24 for normal ops.). Only noticed that a few days later (hey the car is really bumpy, mother said). Never had any other issue :)

Re:Boo

[Tapewolf]

As far as I can tell, tires just blow up when they feel like it. Ridiculous abuse hasn't failed my tires, but normal driving with 35-40psi in a 50psi rated tire has.

...fifty pounds each didn't seem to help with the cornering, so I went back a few hours later and told him I wanted to try seventy-five. He shook his head nervously. "Not me," he said, handing me the air-hose. Here. They're your tires, you do it."

"What's wrong?" I asked. "You think they can't take seventy-five?"

He nodded, moving away as I stooped to deal with the left front. "You're damn right," he said. "Those tires want twenty-eight in the front and thirty-two in the rear. Fifty's dangerous, but seventy-five is crazy. They'll explode!"

"I told you," I said, "Sandoz laboratories designed those tires. They're special. I could load them up to a hundred."

"God almighty!" he groaned. "Don't do that here."

"Not today," I replied, "I want to see how they corner at seventy-five."

He chuckled. "You won't even get to the corner, Mister."

"We'll see," I said, moving around to the rear with the air-hose. In truth, I was nervous. The two front ones were tighter than snare-drums; they felt like teak wood when I tapped on them with the rod. But what the hell? I thought. If they explode, so what? It's not often a man gets a chance to run terminal experiments on a virgin Cadillac and four brand-new $80 tires.

--Hunter S Thompson, 'Fear and Loathing in Las Vegas', 1971

Re:Boo

[timeOday]

It is better to err on the side of over-inflation. The centrifugal forces of high-speed driving are pretty extreme (even 100 mph) and you'd have to over-inflate a LOT to replicate them, so they'll take a lot before bursting from pressure alone. The problem with under-inflation is the tire is racked by vibrations that cause extreme forces, and also cause it to overheat.

Here is a cool page, evidently 225 mph at 20 psi below the optimum is a bad idea.

Re:Boo

[amicusNYCL]

I drove from Baltimore to DC and back with a rear tire flat the whole time (a Goodyear Assurance TripleTred, something actually useful), and then put air in it when I noticed it was flat.

You can drive from Baltimore to DC and back without noticing that you have a flat tire?

I've driven over 800 miles across six months on a tire that was completely flat

You never notice the tires when you're getting into or out of a car? What the hell?

I didn't realize it had exploded

I am Jack's complete lack of surprise.

As far as I can tell, tires just blow up when they feel like it.

How do you know that? I'm going to take a wild guess and assume that you did not inspect the tire to see what kind of condition it was in before it blew, and probably not at any point for months before that.

Re:Boo

[TheSync]

You can drive from Baltimore to DC and back without noticing that you have a flat tire?

Actually I have done precisely that, but it was a rear tire on a front-wheel drive. Only noticed a problem when listening to AM radio and hearing a "click-click-click" when going at low speeds. That was the nail in my tire hitting the road and shorting out the static building up from the tire rubbing.

Re:Boo

[WoOS]

Or to say it in a different way:

The hands-on detection on the steering wheel is there for a reason. The reason is that drivers might not read the manual telling them about all the dangers of letting the car drive without human supervision. Reading the manual is not something the car manufacturer can force people to do.

But by manipulating the hands-on detection the driver shows that he had understood the car's restrictions but willingly circumvented it. If there is an accident, it will be thus the driver who will be charged with negligent homicide, not the car manufacturer.

Re:Boo

[cpotoso]

Let's see... a tire with a radius of ~0.3 m, in a car traveling at 100 mph ( 45 m/s) experiences a maximum centripetal acceleration of ~6700 m/s^2). A 1 in^2*0.5 in piece of rubber (assuming density to be ~ 1 g/cm^3) experiences a "centrifugal foce" of ONLY ~4.7 pounds! That is only 4.7 psi. So over-inflating by 20 psi MAY BE A PROBLEM.

Re:Boo

[timeOday]

Well I must admit, I calculated PSI using Goodyear's figures from my link, which explicitly states the tread weight, and assuming a width of 9" and I got... 4.7 psi.

The forces are 4x at twice the speed, so you'd have to go a little over 200 mph to generate an extra 20 psi.

So I must agree, 20 psi over-inflation is quite bad, don't do that.

Re:Boo

[bluefoxlucid]

My TPMS had been alarming at me, but I ignored it. It turns out I knocked a nail out of the tire when pulling out of the driveway, causing it to rapidly deflate. I assumed it was around 28PSI and not critical.

Re:Boo

[bluefoxlucid]

My tires are rated for 50 and I run them about 5 below. The car suggests 32 for load, as this will keep normal tires from ballooning (high spot in the middle, accelerated tread wear). My tires retain shape with or without load, so this is not a problem; they perform better with a more round profile.

Re:Boo

[bluefoxlucid]

If you have tires rated to handle 100PSI without fault, you don't have $80 tires.

Lorry tires are run at 125PSI fairly often. They're built that way. They're also large, handle long-distance driving fine, and cost $2000+.

The high pressure helps with puncture resistance and braking. Not so sure about cornering; it's not a single-track vehicle, and they're not hilariously cambered for lapping track. It's a road vehicle for driving on normal highway at proscribed speeds.

Re:Boo

[bluefoxlucid]

The tire that blew hadn't been subjected to abuse. I replaced the ones that had been abused with better tires, because I wanted an upgrade; I hated the stock tires. Got rid of the one I'd ridden flat in the transaction.

It was, at the time, fully inflated and in good condition. My TPMS wasn't alarming at me. The tire just blew, and the car started going thud-thud-thud-thud while driving; I assumed I had a flat tire, and found instead what looked like it may have once been a tire. TPMS is marketed as a safety feature to prevent driving on a low tire, because driving on a low tire can cause a blowout, which supposedly causes the car to spin out of control; I could have readily continued at 80mph on that tire until it ripped off the wheel, had I not been concerned for the welfare of the vehicle, so of course I didn't guess the tire blew apart into small, loosely-attached fragments.

So yeah. Tires ridden flat on the drive wheels at expressway speeds for hundreds of miles? No problem. Tires not subject to abuse? Explode. Obviously abusing tires doesn't extend their life; I can only assume they blow up pretty much when they care to.

Re:Boo

[Tapewolf]

If you have tires rated to handle 100PSI without fault, you don't have $80 tires.

Adjusting for inflation it's more like $460 in present-day money - one reason I added the publication year after the citation. Even so, I doubt stock tires on a 1971 Cadillac were rated that hight. Of course, how much of that was actually autobiographical and how much of it is sheer fiction is open to debate.

Re:S-class driver with a soda can? Please...

[Anonymous Coward]

Except for all the driver services that use S-class.

Answer to that age old problem...

[__aaclcg7560]

My older brother used to drive around with an opened beer can between his legs in the 1970's. I wonder why he never thought to duct tape the beer can to the steering wheel and drink from a straw.

Re:Answer to that age old problem...

[smooth wombat]

I wonder why he never thought to duct tape the beer can to the steering wheel and drink from a straw.

Probably because when he made sharp a turn the beer would spill out.

Re:Answer to that age old problem...

[Anonymous Coward]

Turn the wheel fast enough and the beer stays in the can!

The real reason is that the beer foams too much when sucked through a straw.

The real solution is to have a keg and CO2 system to deliver beer under pressure through smooth-bore lines to a vacuum actuated valve much like a scuba mouth-piece.

Re:Answer to that age old problem...

[Anonymous Coward]

Because drinking from a straw is a crime against beer?

Because he'd have to untape it to replace the empty can?

Because you shouldn't even drink beer when you're driving?

Re:Answer to that age old problem...

[thegarbz]

Probably because when he made sharp a turn the beer would spill out.

He should wear a beer hat.

Re:Answer to that age old problem...

[PPH]

Because he'd have to hang his ball sack on the steering wheel to keep it cool.

Re:Answer to that age old problem...

[__aaclcg7560]

Nah... He has ball sack hanging off the truck hitch.

Dumn Idea Stories

[tiberus]

"This is, without a doubt, a really stupid thing to actually try. So don't."

Hmm, wow. Nope the really stupid idea is posting a story on the InterWeb about a really stupid idea and warning us that it's "a really stupid idea". Road & Track should be ashamed that many Slashdoters are now searching E-Bay, CarMax and the trades for an S-Class to try this out in or texting their friends (hopefully not while driving to see them) with S-Class' to try this out. Responsible media, right! Telling geeks about a hack, is like giving crack to a junkie. Tomorrow's lead, dozens die recreating S-Class hack.

Oh, yeah, please PM me your findings.

Re:Dumn Idea Stories

[93 Escort Wagon]

Responsible media, right! Telling geeks about a hack, is like giving crack to a junkie. Tomorrow's lead, dozens die recreating S-Class hack.

And thus is the human gene pool improved by some small amount.

Re:Dumn Idea Stories

[sadness203]

Yeah, there's going to be like, 3907823789 deads tomorrow because of that. :)

Link to original article

[danbert8]

So instead of linking to the original Jalopnik article, you post a copy on Road and Track?

Re:S-class driver with a soda can? Please...

[CanHasDIY]

The actual article says "soda bottle" rather than can... perhaps it was Dom rather than Tab...

Similar to No Hands Across America in 1995

[awtbfb]

This actually isn't that big of a leap from a technical difficulty level. A pair of Carnegie Mellon researchers drove across the country in 1995 using a forward camera based system. 98.2% of the trip was autonomous. The non-autonomous parts of the NHAA drive are the same which would be needed under this approach.

Re:S-class driver with a soda can? Please...

[sinij]

Poster above is absolutely correct, I have my butler tape monocles and glasses of chardonnay to my S-class steering wheel.

Re:Soda can...

[AvitarX]

I'm quite surprised the automated driving will keep that close of a following distance (less than 1 second), looks like a ticket risk to me.

Re:S-class driver with a soda can? Please...

[Anonymous Coward]

Regardless of the origin of the driver's nose, the soda can will be disastrous when/if the airbag deploys.

Fully-autonomous or bust, because

[purplie]

"Pseudo-autonomy" is where the driver is expected to be alert and ready to take over. Therefore,

Autonomous car is to Chauffeur
as
Pseudo-autonomous car is to Student Driver

Ever chaperoned a student driver? Nerve-wracking, and harder than just driving the car yourself. Forget it.

Re:Fully-autonomous or bust, because

[Kjella]

A "pseudo-autonomous" car will probably never fail the basic operations on a road with regular markings and road signs, do everything by the book and pay full attention all around it all the time and it'll never panic, fumble or road rage. I think it will very quickly lull you into a false security where you're wondering why exactly you're babysitting this car because it's driving far more consistent and correct than you would.

The problem is when something unexpected happens and the car fails to recognize it or do something reasonable - that's a very fuzzy definition but everybody who's programmed computer software knows what I mean, no matter how many sanity checks and errors and exceptions you catch something unexpected happens and the software tends to fail spectacularly. I expect that at this point the "driver" will be totally blindsided and useless.

Re:Soda can...

[lgw]

Where I drive, you simply can't leave any more distance when traffic is heavy: if you leave reasonable space between you and the car in front of you, someone will pull in. It's a bit nuts.

But the great thing about this tech is that, unlike me, it has the reflexes to always react safely and the ability to maintain that focus indefinitely. I rely on "looking upstream" to predict changes in traffic flow, and that works well enough, but it doesn't help with drivers who are just crazy, lose a tire, or other such unpredictable events. Now, I'm not sure what scope of events the car can react to, as it's early days yet for self-driving, but in principle it's great.

How close you drive to the car in front of you is a matter of reaction time. I expect we'll no longer be bound by the limits of the human nervous system, soon enough.

Awesome

[Anonymous Coward]

I go through a lot of soda and I use duct for everything so I have plenty of that as well!! All I need now is an Mercedes S class.

Re:Soda can...

[lgw]

Ah, you don't get it - I'm guessing you drive someplace more sane. You cannot leave a safe following distance ahead under some traffic conditions. You could try, but there will be a continuous stream of cars pulling into the space you're trying to leave in front of you, and if you slow by too much to try to maintain that space, now you've become a hazard to navigation, endangering everyone else.

Re:Soda can...

[thegarbz]

Where I drive, you simply can't leave any more distance when traffic is heavy: if you leave reasonable space between you and the car in front of you, someone will pull in. It's a bit nuts.

I've heard this before, but in my experience there's actually very little to it. The people inclined to pull in front of you just because your lane is slightly faster than the other are also those who are inclined to pull out again the first chance they get. Or people who actually want to be in your lane frequently do so because they need to turn off anyway. In either case you're exactly where you were to being with.

Now if they don't do that, how worse off are you? Say 20 cars pull in front of you in your short trip, you're now 20 car lengths back from where you would have been. At 60km/h it means it'll take you an additional whopping 10 additional seconds to get to your destination.

People are a horrendous judge of risk vs reward, especially on the road.

Re:Soda can...

[nitehawk214]

What it means is you will be constantly cut off by other drivers, at least on every single multilane highway in the US that I have been on.

Even if you are in the slow lane you can expect to continually be cut off. They even do it to trucks in heavy traffic which is just freaking suicidal.

Nomination

[pbjones]

A Darwin Award.

Re: Soda can...

[jd2112]

Particularly since tailgating laws are STRICTLY enforced on the autobaun. Nothing like a year end collision at 150 mph to ruin your day.

Can a soda can drive a car?

[marciot]

Yes, a soda can can; that's why it a soda can and not a soda can't!

Re:Soda can...

[SillyHamster]

Try keeping that distance without driving significantly slower than the flow of traffic - unless people pulling in front of you is rare, you can't. Driving at the wrong speed, especially in "bumper to bumper at 70 MPH" traffic creates a significant traffic hazard.

To maintain the same distance in front of you, you are driving the same speed as the car in front of you. How is that the wrong speed?

If there is a constant stream of people merging into your lane - your lane must be faster, or it's their exit. If it's not their exit, there is no reason for them to merge into your lane when you're slower.

The only way your lane is faster while you are going slower is if there is a giant gap in front of you - in which case they should be merging into your lane to distribute the traffic load.

Re:Soda can...

[tlambert]

The primary reason for traffic slowdowns on these highways was rear-end collisions blocking traffic.

You've apparently never driven on U.S. 101 in the SF Bay Area; the primary reasons for traffic slowdowns are:

(1) Auuuuuuuugh! There's a huge ball of light up in the sky! We fears it, my precious!

(2) Look! An accident! Is there blood? Hey, Bill, can you see any blood?!?!

(3) I must get in the fast lane because it is the "fast" lane, even though I'm coming up on my exit!

(4) I must get from the fast lane all the way over to the exit lane, but it's OK if this takes forever, I was in the fast lane for 50 feet, dammit!

(5) Yes, I know it's after 3 PM and before 7PM! What do you mean, the lane to the left of mine is "The Car Pool Lane"? I'm driving slow in the middle lane; if you want to pass, you should get into the car pool lane and pass, then get back into this lane; you probably won't get a ticket anyway...

(6) Let me race up in this lane that I need to be out of before too long, rather than getting over now, even though I see barricades ahead, because I know some dumbass will let me in, right? Right? Hey, dumbass, I'm talking to you!

(7) I want to get on one of the bridges, but I don't want to wait behind all the people who also want to get on one of these bridges, so I'm going to block the next lane over until someone lets me in just to punish everyone else... if I have to wait, then everybody else damn well has to wait, too.

That probably should have been a countdown; fast lane discipline while car pool hours are in effect is probably the number one cause of traffic slowdowns, followed by "I'm too stupid to get over ahead of time", with "Auuuugh! Ball of fire!" in third place...

Re:Soda can...

[gregor-e]

From what I've seen of the HOV lanes in the Bay Area, the signs designating it as such are just there for decoration. This suggests a potentially lucrative opportunity:
(1) Position cameras to catch jerks abusing the HOV lane
(2) Look up their address from DMV license plate records
(3) Mail incriminating photos to jerks, informing them of your mailing service, which promises to forward the evidence to the DMV for free, but with the option of custom-mailing it to an alternative address of their choice for only $100
(4) Profit for the rest of your short and violence-filled life

Re:Soda can...

[thegarbz]

Despite what you may think cars are not governed by fluid mechanics. My earlier comment stands, there are reasons people will pull in front of you and also reasons for them not pulling out from you. Cars don't magically change lanes for no reason because there's a gap. Keeping a proper following distance does not mean you're travelling slower continuously, only for half a second while you're adjusting your safety distance.

If driving 70mph, or slowing to 69mph to open a small gap in front of you is the difference between you getting rear ended then I recommend taking the getting rear ended approach and then getting out of the car and bitch slapping the drive who hit you so hard he gets taken out of the gene pool.

A real disparity between speeds in different lanes is a hazard, as is following too closely. If you think otherwise maybe you should hand in your licence.

MB need to implement captcha dead man system...

[marlaogame]

Because captcha seems to be the solution to everything..

Re: Soda can...

[fisted]

As a native German, most drivers actually yield to faster traffic (,,Stay-right-except-when-overtaking''). I drive on the Autobahn every day (especially the notorious A555); on the rare occasions where I do see someone tailgating, the person being tailgated is usually unnecessarily blocking the lane and hence causing a more fundamental problem to begin with.
While that doesn't justify tailgating, it's questionable whether that's worse than overtaking on the right, or similarly suicidal stuff.

As for enforcement, we don't live in a police state (anymore), so more often than not, there's simply no police around to notice someone tailgating, but if there is, and they do then you're in for a MASSIVE ticket, potentially including having to turn in your license. In this context, it is quite strictly enforced, indeed.

Just for what it's worth Entirely legal no-tailgating rather clean high speed run (~250 km/h)

Re:S-class driver with a soda can? Please...

[mgcarley]

This is an S-Class, not a lowly C or E class.

Why on earth would anybody other than the chauffeur be touching the steering wheel anyway, and why would a chauffeur be trying to fool the car which he's being paid to drive with a can? Sheesh.

Re:Soda can...

[lgw]

To maintain the same distance in front of you, you are driving the same speed as the car in front of you. How is that the wrong speed?

You leave some space in front of you. Someone pulls in. Now what? Do you slow down and leave some space in front, or accept your fate? If you slow down a bit, get some space, then someone new pulls in front. Keep trying to leave space, and you're now going slower than traffic. Get it?

Re:Soda can...

[lgw]

Fun fact about Bay Area freeways: the right lane is reliably faster in heavy traffic. Everyone crams the "fast" lane, and everyone is oblivious to the reality that it's not faster. Moving as far to the left as possible despite objective evidence that it's harmful - bet you never saw that coming in California.

Re:Soda can...

[lgw]

Cars don't magically change lanes for no reason because there's a gap.

Your opinion diverges from the data. I've seen people switch to an obviously slower lane just to fill a gap, because the habit is just that ingrained. Nature's a whore with a vacuum, or something like that.

If driving 70mph, or slowing to 69mph to open a small gap in front of you is the difference between you getting rear ended then I recommend taking the getting rear ended approach

Again, the data does not support your position. Slowing to leave a gap in front of you fails to achieve that goal. The space you create will be filled as fast as you create it, unless you drive significantly slower than the flow of traffic (since people will then leave space in front of you to tailgate 2 inches off the back bumper of the next car - I swear, it's like some guys want to ride in my trunk!).

What happened at the road exit?

[SpaghettiPattern]

If you watch the navigation screen you see the guy approaching a an exit and the video stopping right there. What happened there?

A few observations more.

1. Considerable effort went into the work around. A redundant device was prepared as cold standby.
2. Why was the driver listening to Bavarian (Bavaria is the B in BMW) radio?

Re:Soda can...

[thegarbz]

Your opinion diverges from the data. I've seen people switch to an obviously slower lane just to fill a gap, because the habit is just that ingrained. Nature's a whore with a vacuum, or something like that.

And what happens with the gap they opened up? Or does their car magically split in half? People jump a lane over because they think its faster. If it is great, if it's not they switch back. In either case how many cars are going to jump infront of you during your daily commute? 120? You're still only 1 minute worse off. Hardly worth risking life, injury, insurance fights, roadrage etc.

If driving 70mph, or slowing to 69mph to open a small gap in front of you is the difference between you getting rear ended then I recommend taking the getting rear ended approach

Again, the data does not support your position. Slowing to leave a gap in front of you fails to achieve that goal. The space you create will be filled as fast as you create it, unless you drive significantly slower than the flow of traffic (since people will then leave space in front of you to tailgate 2 inches off the back bumper of the next car - I swear, it's like some guys want to ride in my trunk!).

I think you're going to have to actually provide some data. Right now my anecdotal evidence does not agree with your anecdotal evidence which makes me think you just have a jaded view or are otherwise trying to justify yourself when I call your behaviour dangerous. Or maybe that 1 minute really does matter so much to you and we're back to being a really poor judge of risk.

I agree there's some arseholes on the road. But they are the far minority. Also textbook driving would be if you're being tailgated to double the safety distance from the car in-front of you to minimise the chance you need to break suddenly. An expected bonus of this is that also pisses off the idiot tailgater behind you.

Re: Soda can...

[astar]

Hmm. I used to leve big gaps in stalled traffic. I can avoid stop and go and then the pole behind get to also. In principle the jam is eliminated. Also consider the case where you need to merge right couple miles out. The guy to my left can see what i am doing and hang on my back bumper screwing the assholes who try to get to the front of the line and so block smooth merges. Now tell me about why assholery is optimal behavior at a group level.

Re:Soda can...

[lgw]

otherwise trying to justify yourself when I call your behaviour dangerous

I give approximately 0 fucks about your opinion on my driving. But of course our anecdotes are different, as presumably we live in different places. Why you have trouble accepting that the culture and behavior of drivers in different places around the world might be different, I'm not sure. Watch some traffic camera videos or dashcam videos from various places on Youtube some time (and what a boring place Earth would be if we were all the same).

Re:Soda can...

[rioki]

This was obviously filmed in Germany, as the radio (Bayern 3) tells us. He is driving in the right lane, behind a truck, which puts him around 110 km/h (~70 mph). (This is rather odd for a Mercedes driver, but OK...) The 1s - 1.5s distance, although more on the low end, is rather normal in Germany, everything higher will be treated as a merging slot.

In the US traffic tends be something +/-10% of the speed limit. Which results to people not giving a fuck about only passing on the left hand side, at least in Texas. As a result, in dense traffic the left hand lanes will only move marginally faster than right lane.

On the other hand Germany the speed difference can be anywhere withing 200 km/h (125mph) and 90km/h (55 mph). In optimal conditions traffic is evenly sorted along the three lanes. Since the driver going 160 km/h (being a nice person), will let the driver doing 180 km/h pass, he will merge to the middle lane and back to the left lane. The result is that you see way more lane changes on an Autobahn than a highway. So unless you have really dense traffic the flow of traffic is very irregular or you are tailgating someone, there will probably always be sufficient space to merge in front of you; if you keep 2s or more separation.

Re: Soda can...

[rioki]

Add to the fact that the guy was not tailgating by German law. If I recall correctly tailgating starts at a 0.7s separation. The guy was following at something like 1s - 1.5s separation. This separation distance is normal in Germany. I learned 2s separation Texas and my friends in Germany think I am crazy by leaving such a large separation... Then again German drivers are way more disciplined than Texan drivers...

Re:S-class driver with a soda can? Please...

[rioki]

I know where this joke is going. Then again the car costs a "mere" 100000 EUR, something that a high education / good credit rating person can do in Germany. In addition there are many who buy the car new and resell it after a year, is apparently cheaper than leasing. With some tax trickery even quite feasible. If you "need" the car for representational purposes, like you are a sales person, it is quite plausible that a "soda drinking" person may drive a S-Class.

Re:S-class driver with a soda can? Please...

[mgcarley]

100k EUR for the cheapest one, yes. Significantly more with all the mod-cons and larger back seat.

Either way, I have a feeling that people with 6-figure salaries (save yahoos/yuppies and such) - and especially Germans with this income level - tend toward consuming healthier things than soda, instead probably opting for Perrier water or something - at least I've never any of my friends parents (who are the type who would own an S-Class) drink anything resembling a Coke.

Hell, even regular Germans tip the beer in to a glass rather than drink straight out of the can (save maybe university students, but that kind of reinforces the point).

But yes, the post was tongue in cheek... a Maybach would probably be more representational of the Chauffer-driven types ;)

Re:Soda can...

[Sciath]

I'm not convinced more lanes are any long-term solution to traffic problems. My experience with paving more and widening highways is that creates an incentive for more business development along said roads. Which in turn generates more traffic in a given period of time. Which in turn results in more traffic congestion. My philosophy would be to promote development in areas away from high traffic areas rather than concentrations of businesses. Spread the traffic out more evenly between business areas and even limit the use of certain areas for development to be designated "natural reserves" even if that meant spreading businesses out over a larger area. Then concentrated on the development of fast, comfortable and reliable public transportation,

Re: Soda can...

[fisted]

No, 0.7s is clearly insane.
The German law doesn't fix it to a number, IIRC it's more a "you've got to leave enough distance that you can safely come to a stop if the vehicle in front of you suddenly brakes hard (a.k.a. the 'safety distance')". A general rule of thumb for good conditions (dry road, clear view) is to leave half the speedometer reading, in meters, for a distance. That roughly equals 2s.

Re:Soda can...

[SillyHamster]

You leave some space in front of you. Someone pulls in. Now what? Do you slow down and leave some space in front, or accept your fate? If you slow down a bit, get some space, then someone new pulls in front. Keep trying to leave space, and you're now going slower than traffic. Get it?

Are you trying to thought experiment this driving style, or have you actually tried it?

If you left a car-sized gap in front of you - you don't have to slow down at all for someone to merge in!

Now, the driving style would rebuild that gap - but you don't have to do it instantly - you let off the accelerator and build a new gap over 5~10 seconds, which limits how soon another car can "cut in". If someone wants to merge in every single time you leave a gap, your lane is faster than the adjacent lanes - which indicates you are *not* slower than traffic.

"people merge every time I have a gap" and "slower than traffic" do not mix. Either you are slower than traffic and people stop merging, or people are constantly merging and you're faster than (overall) traffic.