Plug-and-Play for Automobile Embedded Systems

wskellenger writes "This article in the EE Times describes Autosar, a consortium of German automakers and suppliers that aims to standardize vehicle software infrastructure. In this way, vehicle software can be used in different ECUs, reducing complexity and development time for OEMs and suppliers."

Comments

ECU?

[Anonymous Coward]

ECU hasn't been used for years. You're thinking of the Euro.

Re:ECU?

[Anonymous Coward]

Check out the big brain on Anonymous Coward.... Is that a big kahuna burger there... can i taste it? .... bla bla.. allow me to retort...

Moron... you are so dumb it makes my head hurt to think you actually thought you were correcting him on the use of ECU!!!!

the article IS about Automotive ECU!!!

engine computers include; ECU, ECM, EEC, SMEC, SBEC units as well as both power and logic modules.

The future of car automation

[Matey-O]

Things have been going in this direction for quite awhile. The Corvette in 97 went to a serial communications protocol, talking to 14 different control units. It also had a throttle by wire system that eliminated a LOT of complexity in the traction control, cruse control and throttle applications. Active handling, a SIGNIFICANT feature, required a software change and two sensors.

The next step is to get sensors to talk this protocol. Rather than having a dumb sensor that goes to a control unit that interprets the message, you have a temp. sensor that the manufacturer doesn't have to worry about. They just need to look for a temperature unit reporting water temp, or another unit reporting vehicle speed on the wire.

Then the cruise control, the radio, the speedometer, etc all just have to listen for that packet that says 'wheel speed is 60 mph'.

the Cool thing is, the vendor that makes the Vehicle Speed Sensor might do it today for $50. Next year it may be $42, the year after, they might redesign it to have zero moving parts (optical) and with custom asics, make it a $12 part. Will that translate to a cheaper vehicle for you? probably not...but it'll make your car last longer, and will be easier to troubleshoot.

Re:The future of car automation

[2names]

Or on another level, could this mean that my cruise control/radio/whatever that is able to listen/speak this protocol in my truck can be easily pulled out and used in another vehicle? Modular components would be nice...

Re:The future of car automation

[bluGill]

Not exactly. The sensor says "tempature 200." Is that oil tempature, engine water inlet tempature, or engine water outlet tempature? You specified water, but the same sensor [electronics?] may be used in different parts. Where was it installed needs to be indicated somehow.

Some things can be infered if it is operating normally. Tempatures of 100 and 220 indicate inlet and outlet, respectivly (not sure if that is reasonable range, but you get the point). What if the sensor is broke though and is giving a wrong value. Somehow you need to know which one it is so you can replace it.

Re:The future of car automation

[grozzie2]

To really understand the future of automation as it applies to cars, you really need to understand the driving factors behind it. Necessity is the mother of invention, and Economics are it's father. You need both to spawn the end product.

To truely understand the process of moving from 'fully mechanical' to 'fully automated', just take a look at another industry that's much farther along that curve, and understand the reasons it is, and how it got where it is. Look at aviation, in particular, the large transport aircraft.

Early transport jet aircraft were huge and complex mechanical beasts. They had inefficient engines, that spewed a lot of smoke out the back, made a lot of noise, and produced thrust. They had only one re-deeming feature, they could operate at speeds much higher than propeller driven aircraft. If you stuck enough of them on a big plane, you could achieve reliability thru redundancy. This marked the beginning of the process. The 'necessity' originated in military mission specifications. Economics didn't play a big part, the money available to throw at development was virtually unlimited thanks to the military requirements.

When the engineering reached the point that the economics were within reach of the average airline, then the comet and subsequently the boeing 707/dc-8 era aircraft came to light. Completely mechanical, no automation, very inefficient (by todays standards), but they worked, and did the job.

At this point in the development curve, economics became the driving factor of development. the obvious area of development was fuel efficiency. For an aircraft this is MUCH more important than a car. Every pound of fuel you dont have to carry, is a pound of revenue payload you can carry. The engines, and aircraft systems went thru a large phase of refinement, until they reached physical limits of the mechanics, and, the next step was an obvious jump to electronic and automation. Again, early developments in this area were mostly driven by military mission requirements, where cost was no object. Only after military funding developed the concepts to a useable form, did the process of refinement begin to bring the economics down to a practical level for commercial application.

Fly by wire came in aviation for 2 reasons. In the military application, it allows for an aircraft to perform in a manner that true mechanical control cannot achieve, its physically impossible. Again, necessity mothered the invention, but before it came to the commercial world, it needed more necessity, and the economics. The move of large aircraft to fly by wire was driven by 2 factors, with weight savings being a huge driving factor. Replace a mile of cables and pulleys all operating under tension, with very high maintenance requirements, with a mile of wire, that's not moving all the time. Every pound saved, is another revenue producing pound of payload. on the engines side, electronic controls allowed for more fuel efficiency, which results in either more payload, or greater range, for the same airframe. Again, the economics in aviation are substantially different than in the automotive world. With items that sell in the hundreds of millions of dollars per each, and consume fuel in the hundreds of thousands of pounds per day, the incentive to reduce fuel consumption, and increase payload is huge. A few million dollars 'per each' is small cost comparatively.

Embedded control electronics with decent capability are commonplace cheap items today. Development costs have long been amortized, and they are just commodity items now. So, it's reached the point where, adding electronics for small gains on an automobile are cost effective. just put it in perspective. The microcontroller managing the engine in a modern automobile weighs in at a couple ounces, and has more capability than the navigation systems on an apollo spacecraft weighing in at 30+ pounds. That nav system got them to the moon and back. In the early 60's it was worth millions, today, it's worth

Re:The future of car automation

[wskellenger]

We're already there. High-end vehicles since the early '90s all communicate using some sort of network protocol. Now on the lowest end vehicles you'll find multiplexing in use. For late model vehicles, CAN is becoming the most popular. (I work in vehicle development for a supplier of electronic brake systems such as ABS)

Our ABS module reports vehicle speed (and actually four individual wheel speeds) that allow the 4WD system controller to do its job. Diconnect the ABS module and you also lose "automatic" 4WD capability. In some vehicles the spedometer will also drop to zero. The brake pedal switch is also interpreted by ABS it's status is reported on the CAN bus.

For traction control, we look at the actual engine torque, throttle position, gear position, etc. reported by the engine controller. To reduce engine torque we simply request a reduction via the CAN bus.

I think what is important here is the fact that if VDO suddenly wants to produce "smart brake lights" that increase intensity depending on driver brake effort, they can now simply drop their software into our ECU (as the brake module is best equipped to handle this function).

YOU FAIL IT

[Anonymous Coward]

Because he's an idiot! Like you!

Re:YOU FAIL IT

[Eric Ass Raymond]

Ditto!

Autosar participants...

[Anonymous Coward]

...express infectious optimism that, once embraced, their standard will become endemic.

Plug-And-Play President: +1, Patriotic

[Anonymous Coward]

This moron, when asked to summarize to date, responded with "I don't have any idea".

Brilliant. I guess Redubyacans enjoy stupidity.

Thanks and have a great weekend,
W00t

All this homogeneity

[pheared]

Can't wait for the first car worm.

Re:All this homogeneity

[twoslice]

I am going to write a worm that will roll down all of the car windows in the middle of winter when it hits -40.

Re:All this homogeneity

[PD]

Just make a worm that will do one of two things when it detects that the vehicle is in the left lane:

1) speed up, or
2) swerve to the right

That will fix the biggest problem on our expressways.

Re:All this homogeneity

[homer_ca]

Yeah, you could call the worm AutoSARS, har har.

Re:All this homogeneity

[Anonymous Coward]

Not enough accidents?

People who like to get good gas mileage/don't want to hear too much wind/engine/road noise but want better visibility than they can get behind that humungous truck?

What?

I have a better idea: network the cars and have them perform the optimal acceleration, braking, and turning so as to prevent accidents, improve timeliness, and conserve energy.

Re:All this homogeneity

[PD]

You must be one of them.

Get out of the left lane. It's not the smoking end of the high school. You do not just "hang out" in the left lane when there's nothing on TV to watch and you're bored.

Move over to the right. ALL the way to the right. If you have to pass, stay left long enough to pass, then move right again. ALL the way. The middle lane isn't right, it's middle. So if you're in the middle lane, keep moving to the right.

And don't get into the left lane, accelerate to .01 MPH faster than the slower lanes and claim that you're "passing". You're not passing. You're not fooling anybody. Passing is when you get around the other guy quickly and then move over to the right.

WinCE?

[Spillman]

The article says WindowsCE will be "taken into consideration" as part of the OS standard. I ask you do you really want a windows kernel controlling your anti-lock brakes?

Re:WinCE?

[Jason1729]

Reading the article, I just kept thinking of the "if Microsoft made cars" joke.

Jason
ProfQuotes

Bumble Bee?

[Telepathetic Man]

The cartoon lied to me!

Transformers aren't from another planet, they're from Germany.

So...

[Pig Hogger]

Does this means that to fix a flat, we'll have to swap tires until we find the faulty one???

Less software not more in Cars

[voss]

The ugly fact is all these computers have not made our cars more reliable. Im not against computers in cars but they need to have programs as simple as possible and these programs should be open source.

If my computer crashes because of a bug I can replace it.

If my car crashes because of a computer bug, me , someone I care about or someone who could sue me could be injured or killed.

Re:Less software not more in Cars

[NickDngr]

The ugly fact is all these computers have not made our cars more reliable.

Cars get far better gas milage and produce far fewer pollutants than they did before the introduction of electronics. I would argue that they are also more reliable. You can buy cars now that can be driven for 100,000 miles with only regular maintenance (oil, filters, etc). You don't even have to change the spark plugs.

Re:Less software not more in Cars

[battjt]

Huh? A 1960's car would not drive for 55,000 miles with zero problems like my 2000 has. I think the engine electronics are at least part of that reliability. New International trucks use an ESC for everything from emergency light control to boom control. Due to many fewer physical modifications, reliability is expected to increase. (no more fire engine company splicing truck company wires to make the headlight blink) Or are you talking about safety. Air bags have quite a bit of electronics associated with them. Joe

Re:Less software not more in Cars

[Fat Casper]

If my computer crashes because of a bug I can replace it.

If my car crashes because of a computer bug, me , someone I care about or someone who could sue me could be injured or killed.

We replace a lot of modules, sensors and the like. The customer complaint is either "the check engine light came on" or "it's running/shifting a little rough." Your car will only crash from a faulty computer if you ignore the lights and let it go until the computer falls out of the car. Similarly, you'll only go off the road with slick tires if you've been refusing to replace them, too.

Re:Less software not more in Cars

[cowlum1]

You are correct. Im a classic car enthusiast and have had much dealing with older vehicles. Older vehicles (even when new) definatly require(d) more tuning/tweaking in order to remain reliable. However because of their simplicity they tend to suffer less from age and so break down less often. Modern designs are definatly less reliable after the 200,000km mark - but many people argue a vehicle should be scrapped by this time anyway.

Its common to install fuel injection in older cars (cult cars) though, because fuel injection creates HUGE gains in reliabilty/efficency/performance.

Great

[Anonymous Coward]

I'm gonna get me a Series III Land Rover when I eventually get my licence. No unnecessary electronics to fail - not even spark plugs on the diesel version {known to run fine on cooking oil}. The engine's the wrong way around compared to a normal car {the cylinders go from front to back, not side to side}, but I guess turning it 90 degrees on must make it easier to get power to the back wheels.

Re:Great

[Anonymous Coward]

as long as it's a stick, buddy

Operating System

[ericspinder]

From their site:
The standardization of automotive operating systems is not regarded as an AUTOSAR goal but existing standards and products such as OSEK, VxWorks, Windows CE for automotive and their derivatives will be taken into consideration and used in AUTOSAR.

Adding in words like AUTOSAR interfaces, "AUTOSAR runtime environment (RTE)" and microprocessor abstraction layer, I'd say that they are taking a page out of Java's handbook.

imagine

[Anonymous Coward]

imagine a beowolf cluster of corvette!

Test Equipment

[repressitol]

This is interesting -- could this eventually do away with proprietary test computer equipment needed for each manufacturer? i.e. open it up into an environment where someone could extract this information via a some type of standard port to a laptop. Or would this lock it down further into more expensive (but standardized) equipment controlled by fewer providers? Anyone here have to deal with car problems & found you couldn't extract the details without an expensive "analysis" tool? (insert GM, Chrysler, BMW).

Re:Test Equipment

[Anonymous Coward]

AutoTap

Re:Test Equipment

[Anonymous Coward]

Actually I would say they are taking a page from SciTech on this one. Take a look at the SciTech SNAP White paper which outlines a common abstraction layer by which all hardware can be completely isolated from the surrounding OS environment.... SciTech today uses this in relation to graphics drivers but the SDK alludes to much more.

Re:Test Equipment

[garymcg]

This won't even come close to solving this problem. All of the major manufacturers use different protocols which are not compatible, that is why there is no standard one-tool-fits all diagnosticl tool. When they came out with the OBD (on board diagnostic) standard, the only thing that was specified was a standard connector, each manufacturer developed their own implementation using whatever protocol they thought best. It's chaos for engine diagnostic software developers. There was a time not long ago when different divsions of GM used different protocols, a Buick data reader wouldn't work on a Pontiac.

Re:Test Equipment

[Anonymous Coward]

"could this eventually do away with proprietary test computer equipment needed for each manufacturer? "

I think it's more likely to lock up test equipment in the DMCA envelope. Just because the components are standardized doesn't mean the manufacturers aren't free to fuck up the potential in the name of control.

I really don't get how "control of the consumer" leads to profits anyway, and I don't think it will yeild any benefits in the long run, and that gives me hope.

Short term, though, lots of industries seem to be all about control. But to what end? The reward for control is... control... not shareholder value.

Re:Test Equipment

[Anonymous Coward]

The reward for control is locking out unlicensed aftermarket equipment => more $$$ for the auto manufacturers. Oh, and don't forget you need at least a few hundred million $$$ to even think about starting a competing manufacturer (BMW is currently the *smallest* auto group). While it might at first seem alluring to take some other manufacturer's vehicles and replace all the proprietary ECU components, I doubt you could get the source cars and the upgrade parts cheaply enough to realistically compete. Most people initially won't know about the difference, and as long as the automakers don't go crazy with licensing fees can cause all aftermarket equipment to inflate to dealer prices, they won't really notice that it's a significant enough change to warrant looking for "open" features in their *nest* car (probably at lest 3 years, even for the average yuppie moron; more like 15 years anyway for the people who would be concerned about maintaining their cars).

Re:Test Equipment

[brainthought]

When Garymcg had the floor:

each manufacturer developed their own implementation using whatever protocol they thought best.

Yeah, and those dealers really want to standardize all the ODB diagnostics. Cause they don't want to charge you $100 just to run the diagnostic. GM only standardized because it was easier for them for two reasons, a)because all the cars were easier to manufacture if the only part that was different was the Chevy, Pontiac, or what have you emblem; and b) it was becoming a nightmare for the dealers who mostly sold in combinations (John Smith's Chevy-Cadillac-Olds Town... sort of stuff) to have to keep five or more testers around.

Listen people, no auto manufacture in the world is wanting to make things easy for you to fix or to work on. There will never be a standard between ODB systems or anything else for that matter! Your kidding yourself if you think that cars are gonna get more and more reliable to, because guess what folks their not. You think GM wants you to come in and buy a nice, easily reliable, cheaply fixable Cavalier (sorry Cavalier was discontinued this year, read Cobalt)? One that you can drive for 30 years and it always be dependable so you never have to come back for repairs again? No! Ford, Honda, GM, it doesnt matter. They want you to come on the lot, finance a car for $20k pay it off over 6~7 years then the thing starts failing... Plastic tensioner pulley dry-rots and the serpentine belt comes off, transmission needs servicing, plastic fluid reservoirs start breaking, seats start to brake, and the car rolls say 150,000 miles, so you come in, trade it in and finance another 20k for another 6~7 years... Hell, they don't really even want you changing your own oil. Most American auto companies, like Ford, with cars in the mid $20k's don't even want you to buy. They want you to finance. I was recently on a BMW/Mini lot and they wouldn't even talk to me about financing to buy. Companies like Kia and Hyundai have even gone so far as to basically make their cars disposable. $5K, you drive it for three years, then scrap it and get something else...

The cars basically are designed to have a shelf life of 10 years at best, and just to make sure the shade tree mechanic goes away, the newest auto manufacturing fad is to make the engine compartment just big enough that you drop the damn wrench every time you try to get to any component, and none of your tools fit anyways. You can't even get to the sparkplugs, which with a V6 or V8 half are now located behind the engine up against the firewall so they can mount the engine sideways where you can have front wheel drive so you get to replace your CV-Axles every 60k miles... and on and on. Or if it's a 4-cyl like the VW Beetle (new Beetle) the sparkplugs are wedged up in the engine housing so far you start thinking about dismantling the dash just to get to the things.

The other thing I am now enjoying is these new hybrid cars and alternate fuel cars... You know what, go onto a Honda dealership or Toyota and tell a salesperson you want a Hybrid-Civic or Prius and you want to be driving them 10 years from now... They'll tell you that you won't, cause in about 8 years the batteries stop holding charges (about $800 to replace them according to my local Honda dealer), and that's TODAY, who knows what it will cost in 2011 when they don't even make them anymore. Not to mention the whole plethora of new and special tools you need and the fun aspect of working on TWO engines in an engine bay that used to barely hold one.

Looking down the road to the future is just a bleak sadly, GM's skateboard concept vehicles [gm.com](or AUTOnomy) contain all the parts underneath the car in the chassis and EVERYTHING is drive-by-wire. So I guess you won't have to buy any more tools to work on the car in a few years.

Now, I know a lot of you are gonna post back and whine in usual Slashdot-Automotive piece fashion, So what, I don't want the same car in five years

Re:Test Equipment

[mr.krabbs]

If you have access to a manual and a DVOM, you can see all the data you like without the expensive tools. I know this 'cause I was a Driveability Specialist (i.e. I diagnosed the electronic end of the automobile) before returning to school to become a 'Puter Engineer.

Integrated software?

[dustmote]

Software, schmoftware, I still want to be able to install a new radio without having to remove a series of plastic dashboard pieces larger than my torso. It's enough of a hassle that I end up paying people to do it for me. (Usually in beer, but the fact remains.)

Re:plug thine arse

[Trolling4Dollars]

I am gald to see the luminaries got early posts. Congrats to Sexual Asspussy, Eric Ass Raymond and Fecal Troll Matter. May you all live long and continue to post heartily. Props to Subject Line Troll as well. While I may not be a troll, I respect those who troll well.

ODB-II

[kevinmarsh]

This is good since we already have an open monitoring standard: ODB-II.

Make your car last longer ?!?!?

[Archfeld]

NOT EVEN CLOSE, it will allow them to completely control the the replacement time of your entire car. This will make engineered obsolecence (sp?) the current corporate darling even more controllable. The vendor can sunset the support for your car because of software/firmware upgrade issues...Think of M$'s OS upgrade cycle and how much a durable goods manufacturer would like to be able to emulate that kind of re-buy re-supply cycle. Of course with the DMCA, you will be forbidden to try and revers engineer parts, the warranty will be void if third party parts are used and the only place to buy certified parts is from the dealer at a 5000% mark-up. This spells COST COST COST COST for the consumer....

Re:Make your car last longer ?!?!?

[Matey-O]

You don't know much about cars. Do you?

Ever heard of SEMA? (Specialty Equipment Manufacturer's Association) They're a HUGE lobbying group for aftermarket suppliers of parts.

Microsoft perrenial upgrades? When your car is running okay, what do you put in it? Gas? Tires? What else? It won't automatically stop running if you don't pay your EA license.

Comparing the Auto industry to the Computer industry is just flat stupid and a non-starter.

Re:Make your car last longer ?!?!?

[sbaker]

If it can happen with something as mundane as printer ink - it can certainly happen to something like a car.

Dear Taco,

[Anonymous Coward]

This is getting really boring.
Plus, moderators are on crack
and the crack is only getting more powerful.
Funny stuff gets scored -1 Troll. Really
stupid, unfunny crap gets scored +5 Funny.
The moderators have a really warped sense
of humor, stuck in like 9th grade or something.
Only worse. Funny Slashdot != Real world funny.

Taco, fix your product or wind up having everybody
move to fark.com . Or watching TV. Or going back
to programming all day and no surfing.

There are already too many vehicle LAN standards

[Animats]

There's SAE J1850. There's SAE J1939. There's CANbus. There's IEEE1394. There's FlexRAY. There's TTA. And they're all incompatible. Currently, TTA has Audi, Delphi Automotive Systems, Honeywell, PSA Peugeot-Citroen, TTTech and Volkswagen. FlexRay has BMW, Bosch Automotive Group, DaimlerChrysler, Ford, General Motors, Philips Semiconductors, and Texas Instruments. The US heavy-truck people have J1939. Some European truck makers use CANbus. J1850 is commonly used today for auxiliary functions in US passenger cars. None of this stuff interoperates.

This new Autosar announcement is really a spec for an operating system. The companies pushing it don't want to say that, because that means taking on Microsoft. So they present this as a middleware layer. But it's really an operating system API that provides independence from the underlying OS. Think Netscape plug-ins.

turnkey vs a hackable platform...

[Anonymous Coward]

You bowtie guys just don't get it...

When Carroll Shelby beat the pants off the Italians, he was driving for F O R D!

Windows ?!

[danknight]

I was going to make a windows joke about this and then decided to actualy RTFA before posting and there it was, Windows CE! a few questions if automakers decide on windows... 1, will I have to agree to a Microsoft EULA in order to use the car? 2, will I need a separate Windows License for each Driver? 3, Will I be able to sell the car later or will I need to erase the operating system first ? 4, Will Windows DRM restrict what CDs I can play in the stereo? (will I even be able to install an aftermarket radio ?) 5, Finally will I crash if Windows Crashes while I'm crusing down the highway at 65MPH ?

Re:Windows ?!

[Baron_Yam]

You know, I thought that was kind of a stupid, not as funny as intended kind of post until you got to the bit about the radio.

I can just imagine MS putting code in that only permitted a DRM-compliant radio to work in a car - which would mean no home-burned compliations, legal or not, knowing the way DRM usually ends up 'functioning'.

Anyway, I think that level of problem is far enough off that one need not be paranoid quite yet.

Re:Windows ?!

[Anonymous Coward]

yes

Re:Windows ?!

[danknight]

Yeah, I was debating as to if I should even post and I was trying to come up with at least one point that was not to stupid. All joking aside, this seems to be an application that is screaming for an open source or at least a BSD style solution. In the automotive business even pennies count so why would an automotive manufacturer even consider a solution which would have licensing issues ?

Re:Windows ?!

[WheatWilton]

Buy an old Charger, Bird, Goat, Cuda, etc. No software to worry about. Just big, bad carbuerated engines belching smoke and fire... I mean, is there anything better than eight pistons beating in synchronization with your heart?

Please, oh please don't let it be windows based.

[snoopyjd]

I don't want to have to hit CTRL, HORN, HAZZARDS every time I want to change the radio station.

eaowg8u9h

[Anonymous Coward]

:SLJGsjgAS

So when I crash my next car...

[Ironica]

... The up side is, no body work, I'll just reboot it.

The down side is, it will happen several times per week, and usually right when I most need to get somewhere...

Next Logical Step

[Harley]

The auto industry has been headed in this direction for a while. The biggest step was probably in 1996 when all (I believe) vehicles sold in the US were required to be compliant with the ODBII (On-Board Diagnostics II) standard interface. The result was that, to troubleshoot problems with the vehicle's engine and transmission, a person could pick up an OBDII code reader and pull the codes from a Ford, Chevy, Toyota, etc. Foreign or domestic -- it's the same standard.

Granted, it a standard for pulling codes rather than controlling functions within the vehicle, but it was certainly a step in the right direction.

The auto makers will get there once there's either a de-facto standard that's proven to work, or there's government regulation requiring it.

Wonder which will happen first?

Standardization Always

[MhzJnky]

Most problems I've had with my two cars have all been either sensor or controller issues. Seems the actual physical parts of today's cars work rather well, and don't ware out that fast. However the computer bits seem to be not as reliable.

These parts are expensive because they are proprietary. Only a Ford controller can be fitted to a ford car. Doesn't matter who makes it, it has to be made specific for a given type of car no matter what. Standardizing would help bring these costs down a great deal.

Proprietary hardware and OS's in this day and age are redundant. No longer should this be seen as an advantage, because its not. Commodity hardware and software packages are tested more than an individual vendor could ever dream of doing. They are more reliable and cheaper than anything a small group could build themselves. Not that GM or Ford is a small group, but the more people using it and testing it, the better it will be.

After all, isn't that what made linux what it is today?

Re:Standardization Always

[Matey-O]

It's worse than that. Currently, only a Corvette Left Door Control Module can be used as a Corvette Left Door Control Module...

What we need is a standardised spec with standardised connectors. That way you take your Napa Gold System controller, tell it it's a Corvette Left Door Control Module and go to town.

(The aforementioned LDCM has a sensor to tell door state, relays to control door lock solenoid actuation, and relays to handle the power window. Concievably you'd buy a Napa Gold 4 port controller as it'd have enough relays to control what a LDCM would deal with. Or even better, it's a three part unit. You attach an X channel input, a Brain of a certain power, and a Y channel output controller.)

Man, too much thinking for a Friday afternoon.

Re:Standardization Always

[ltkije]

Proprietary hardware and OS's in this day and age are redundant. No longer should this be seen as an advantage, because its not.

There are some very powerful disincentives to introducing standardized (i.e. generic) hardware or software into any automobile. Start with the notion expressed in the EE Times article that software from one manufacturer would run on another's hardware. Today the auto industry treats software as if it were free, and pays only for metal boxes with wires coming out of them. The European consortium will have to propose a viable way for software-only vendors to make a profit.

Second, the auto manufacturers, in Detroit at least, are obsessive about keeping control over every aspect of what goes into their vehicles. The incompatible controllers are no accident. A lot of it has to do with parts cost -- at a minimum run of, say, 500,000 for any controller, the auto company can afford to cut pennies by making each one fully customized. (Remember, software is free!)

Third, keep following the money. There is a deliberate effort to get you to maintain and repair your new car at the dealer where you bought it. Cash flow has a huge influence on auto company decisions.

Fourth, between the time a new model rolls off the line and the time your corner garage can get hold of the diagnostic codes for it, two or three years will have passed and at least one company will have made money off the slow proprietary-to-public transition.

Starting to get the picture? The right (cynical?) way to look at the consortium is as a way to drive down the cost of their proprietary products, not as a way to get open software or even software standards into cars. MISRA C is standard, OSEK compliance is becoming standard, but that's about it.

Now, having said all this, I would also guess that the auto industry has recognized the problems of software complexity and development cost. Anyone who's been around, though, recognizes that the reality of software components has never lived up to the hype, and the cost-cutting mentality reigns supreme among the gearheads.

Interoperability is great

[Anonymous Coward]

As long as I can play nethack using the car's computer system.

AUTOSAR related to HOMSAR?

[Atario]

Maybe?

Just strip the OEM electronics out

[xtal]

There are many aftermarket ECU manufacturers out there. A number of open source projects are getting close. It's only a matter of time before someone releases completely open hardware and software that can replace your car's ECU. Any modern RTOS is more than capable (QNX, for example).

Unfortunately, the above is technically illegal in a lot of places because it lets you bypass emissions controls. Not where I live, though. :)

Re:There are already too many vehicle LAN standard

[BWJones]

This new Autosar announcement is really a spec for an operating system. The companies pushing it don't want to say that, because that means taking on Microsoft. So they present this as a middleware layer. But it's really an operating system API that provides independence from the underlying OS.

This is important as Microsoft has expressed an interest in automotive control systems. In fact, I recently found out that the iDrive system in BMW 7 series automobiles is Windows CE based. (No wonder I hated that monstrosity so much. Commonly accessed functions are buried several menus deep requiring you to absolutely take your eyes off the road for long periods of time.)

QNX in Automotive

[Kilkonie]

The Ottawa Citizen newspaper just had a profile on QNX as part of their semi-annual high-tech review. It goes into quite a bit of detail about QNX's recent move into the automotive space. Ottawa Citizen Link: QNX operating system revs up for the road

Bug or Lemon

[OYAHHH]

So,

At what point does all the computer bugs in your car create a point where you can legitimately invoke the lemon law provisions?

On a side note I started trailing a lady in a brand new BMW 7 series a couple of days ago. The car's emergency flashers were on and at the leisurely pace she was taking things I knew she wasn't aware of it.

So I pulled alongside at a redlight, fortunately she had her drivers side window down, so I shouted to her that her emergency flashers were on.

She looked really surprised and muttered something to the effect of "Oh really". Not a doubting oh really, but a surprised oh really.

Apparently there was no indicator inside the car telling her what was happening with her lights.

If I'm not mistaken the BMW 7 series has a Windows CE O/S? I've heard the 5 series does.

I know I'd be incredibly irritated to spend the kind of money she had in that BMW only to find it riddled with computer bugs.

Lastly, isn't it the law in the US that car makers have to "support" the vehicles they sell for 8 years?

Will MS still be willing to issue a BMW patch 8 years from now? They've certainly seemed to be trying to reduce the amount of time they support a particular version of their O/Ses.

user-accessible diagnostics would be nice

[egburr]

What I would like is the ability to run a cable (USB/ serial/ firewire/ ethernet/ whatever) out to my car and fire up some open source diagnostic software on my computer to get some hints at what's wrong *before* I haul it in to the service center.

Re:user-accessible diagnostics would be nice

[grammaticaster]

Then get a Volkswagon. You can run hook your laptop up to the computer in a VW (or Audi, Seat, or Skoda) and get all sorts of nifty information (http://www.ross-tech.com/vag-com/), and even adjust the performance and emissions characteristics of the car. The open-source part is the trick, though. Aside from the expensive connector, you have to pay another couple hundred bucks for the software. I'd be very happy if somebody wrote some freeware similar to VAG-COM.

Re:user-accessible diagnostics would be nice

[sbaker]

I'm pretty sure you can already do this.

The diagnostic interface to all cars that are sold in the USA *is* standardized. Look under the dash somewhere close to the steering wheel and you should see a socket (probably covered with a snap-shut cover). This must conform to the 'OBDII' standard.

You can buy the interface pack to hang a laptop or PDA onto the car:

eg: http://www.elmelectronics.com/

There are also circuit diagrams out there to let you build your own if you want to (they interface into the serial port). The codes for the 'typical' problem reports are standardized - and lots of the non-standard ones are well documented out on the web.

The software to read out those codes - and to reset them once you've fixed the car yourself must be out there - and if it isn't, it can't be more than a page of code if you don't want some kind of fancy GUI.

Re:user-accessible diagnostics would be nice

[Anonymous Coward]

Ah, $199 gets you the interface (what you call an "expensive connector") and the full-function software.

If you supply your own interface, there's a free (but somewhat limited function) version of the software, and registration for the full version costs $99.

-Uwe-
VAG-COM Author
Owner, Ross-Tech LLC

GM, the leader in innovation

[SuperBanana]

Things have been going in this direction for quite awhile. The Corvette in 97 went to a serial communications protocol, talking to 14 different control units.

Meanwhile, the same technology was in German and Japanese cars at least 2-3 years before, depending upon the manufacturer. Who do you think GM got the idea from?

Vehicle software

[TennesseeJeff]

I believe this will have a positive impact on the cost and maintenance of my future car. Lowering the cost from the current $100 to a future of $25 for one sensor/device may not have much of an effect, but for 100 different sensors it should have a very positive effect. And with the many manufacturere producing standards compliant sensors/devices the reliability of our vehicles is also bound to increase. I jut have a few questions here:

1. How will the DMCA affect this? Will we still be able to work on our own cars or will that be circumvention subject to DMCA? How about modifying the software (hacking in the pure sense) the software that runs my car - Will I be able to do that legally?

2. What about the black boxes that are now standard in cars? What data will it now store? Who owns that data? Who controls that data? Can I erase it? Can I refuse to let it be recorded? What happens when GPS is integrated?

3. What happens when I sell my car? Is the software included like it is now? Or does it have to be relicensed like when I sell a computer?

Re:Vehicle software

[Ironica]

What about the black boxes that are now standard in cars?

According to the link you posted, these aren't standard... the standard for them is being developed. At least, it was as of April 2002, and it seems unlikely that in less than a year it's gone from concept to standard equipment.

However, standardization of car electronics interface would probably make the project you've pointed out quite a bit easier.

Re:Vehicle software

[Anonymous Coward]

DMCA = Digital Milennium Copyright Act.

Unless manufactures include some binary code that is copyrighted, so that code has to be present for the OS to operate it (ala Lexmark printers), then you don't have a thing to worry about.

About the other things: cars are comodities. You don't liscense gold, food or other things... You don't liscense the operating system in equipment, unless it's specifically mentioned in a contract the user must agree to. There's your answer. Nothing has changed.

Next.

reliability, reliability, reliability.

[www.sorehands.com]

That is a question that must be answered. How reliable would the new standard enforce. Issues regarding noise on the bus, or component failure causing the ECU to go nuts, etc.

It is a nice idea to have plug and play components and are able to access the control system for tweeking, diagnostics, etc. But, you also have to consider how much control should a car owner/user be permitted to do.

Re:reliability, reliability, reliability.

[sbaker]

I agree - and it's not just electrical noise (which is a big problem with all of those sparks going off at a few tens of kilohertz - and the starter motor) - but there are also thermal issues (it gets pretty hot under the hood) and vibration too.

It's not going to be easy to physically manufacture a workable/reliable ECU.

Then you can bet there are going to be liability issues in car wrecks.

It really doesn't bear thinking about - yet it's something I think NEEDS to be thought about. OpenSource software under the hood of my car would be a very good thing.

Standard audio controls/system

[Faeton]

While you're at it, please standardize audio components. Make all audio decks the same size (DIN would do nicely), have standardized connections (so I don't have to use a harness).
Standardize the on-the-wheel audio control interface so I can use it to do simple things like control volume, skip/search track/station.

If someone could do this, we would all save hundreds by not getting the overpriced audio options the car dealer has. Whoops, doesn't sound like it'll happen anytime soon.

More software, not less, in Cars

[TimTheFoolMan]

So how many crashes can you attribute to closed-source embedded-system failures? You surely love the shorter stopping distances of ABS brakes. You doubtless like the fuel efficiency of advanced ignition timing that's possible now that engine control computers have replaced the old "points & condenser" systems of years ago.

Ever wonder about the gibberish that your local import tuners (hot rodders with Hondas, Subaru's, etc) keep spewing about ECUs? They're reprogramming the air-fuel mixture at various points in the RPM curve. More accurately, they're gaining HP and performance by programming tables in an EPROM. Less software means less configurability.

Furthermore, if you think that a safety-related computer system doesn't get seriously tested, you obviously haven't tried to get a life-safety product UL-listed before. The system isn't perfect, but it's annoyingly comprehensive. Likewise, all safety-related systems for consumer products go through all sorts of validation steps AT THE DESIGN PHASE!

It's simply ignorant to think that your life isn't constantly being protected by numerous software systems. Product liability law being what it currently is (and it's not ALL bad), don't expect this to change anytime soon.

As for making such code open source... that's just a band-aid recommendation from someone who assumes that software engineers aren't capable of creating embedded systems without flaws. On the other hand, if WinCE or Embedded XP start seeing more applications in this space, you may have a point... <g>

Tim

Re:More software, not less, in Cars

[SonicBurst]

Less software means less configurability
Yeah, because pulling the block, boring it 30 over, putting in new headers, manifold, crank, cam, heads, pistons, rings, 4 bbl carbs (perhaps a predator tri-pak, and electronic ignition are oh so non-configurable!
You haven't seen configurable unless you've seen a fully tricked out 30's street rod cruising town. Now that's a car.

Re:More software, not less, in Cars

[TimTheFoolMan]

Wow... I never thought of that. (insert sound of smacking forehead here)

Oops! That's because most of those mods have either been done by the manufacturer of a modern car already (upgraded internals and better breathing), or the manufacturer's standard config (with things like fuel injection and distributorless ignition) is better. (Don't bother arguing that a 4 bbl carb is better than computer controlled port injection. Simpler, yes, but not better.)

For example, take the F20C out of my car (Honda S2000), and see what you can improve. Outside of ECU reprogramming and forced injection, you're not going to do significantly better in terms of breathing and internals than the factory (120 normally aspirated horsepower per liter, and a 9000 RPM redline).

I've seen configurable (not just my own cars). Cars of today are not less so, it just takes more of a geek, and less of a grease monkey than it used to, because most of those changes have already been made, and by people who are smarter than you and I.

Now we can all pine for the good old days, when engines were simpler, and Goober and Gomer could fix the problem without knowing the OBDII codes. Unfortunately, we'll have to leave behind the dramatically improved spark advance curve, variable valve timing, and longer-lasting mechanicals of the modern era.

Tim

The BMW/MINI Cooper saga.

[sbaker]

The new BMW MINI Cooper has an ECU built by Siemens and programmed by yet a third company. BMW claim that they don't even have a copy of the source code for their own car! The same ECU is used in a variety of engines. So in order to have the code optimised for a particular car, the software "learns" - over several tankfuls of gas - how best to drive the car. Since cars change over time, it continually re-learns. If you add a new air-filter for example - the effects of doing it only gradually appear - over about three tankfuls - as the ECU learns to adjust the fuel/air mixture again.

This has consequences. Firstly, when you buy a new MINI Cooper, you get really poor gas milage for the first few tanks of gas - but gradually (as the ECU learns), it gets better.

So far, so good.

But the MINI has a problem (known as the 'stumble' amongst owners) - it's a software bug that appeared in 2003 model-year cars - older cars don't have it unless they upgrade to the 3.3.x version of the ECU software for some reason.

Under the special combinations of high air temperatures (and perhaps only in low humidity) in the summer in the southern USA - and with 'Reformulated Gasoline' that we get here in Texas and in Florida - the car sometimes stalls out at dangerous times. (eg You pull out into traffic - and the car stalls halfway across the road).

The stumble was VERY hard to diagnose - both because BMW couldn't reproduce the problem - it took a lot of MINI enthusiasts across the US to finally figure it out.

We (within the owner's community) decided that this couldn't possibly be temperature related - because the car would still stumble in the cool of early morning. We decided it couldn't be reformulated gas because we could drive to Oklahoma - buy a tankful of "the good stuff" - and still experience the stumble.
During a heatwave in Washington (who also have reformulated gas) - there were no reports of 'stumbles'.

These were cases where diagnosis was made almost impossible because the ECU had *learned* to stumble - and needed either cool temps or better gas for THREE TANKFULS in order to recover from it. People who experienced a short heatwave - or who bought only one or two tankfuls of reformulated gas didn't see the problem.

In consequence, it's taken over a year to convince BMW that there really is a problem and to find out what it is. However, BMW themselves can't fix it. They have to work through Siemens to get to the third company who programed the ECU so it could be fixed - and those guys didn't want to just fix it "the easy way" because it would have the potential to screw up performance in other kinds of car that use the same ECU software.

We are promised a fix for the stumbles - sometime in December.

This is all VERY yukky and unsatisfactory.

The thought of trying to write OpenSource ECU software came to mind - and there are some projects out there to do just that. This ECU has reloadable software - using a serial port connection that appears just under the steering wheel (used for emissions control stuff too). You can buy a cable to adapt the car's serial port to that of your laptop or PDA - and there is even software to let you read out and reset the engine management error codes in the comfort of your own driveway.

Armed with a laptop, your car dealership can upload new software into your car in about 20 minutes.

However, attempts to do this ourselves resulted in a fascinating inside into what the world of Palladium/DRM. When you tell the MINI "Please accept a new software load" - it sends you back a 16 bit random number. You are supposed to execute some predefined math operations on that number and send back the result as another 16 bit number. If you get the answer wrong, the car completely shuts down for 3 HOURS! You can't even start it under those circumstances - let alone try again with the software download. Obviously, the math operations you have to evaluate to solve this challenge/response scheme are secret.

So - welcome to the world of the future. For some of us it's already here!

Re:The BMW/MINI Cooper saga.

[hopbine]

Thank heavens I have a real mini cooper. (1973)

Re:The BMW/MINI Cooper saga.

[chiph]

I haven't had the stumble or "yo-yo" as it's also known in my September 02 build MINI Cooper S ('03 model year). Could be that the right conditions just haven't happened for me. The Raleigh area doesn't use the reformulated (i.e. crappy) gasoline, although we do get high temps in the summer (95F and up).

I am selling my car, however. The MINI is a fantastic car, but I want an IS-300 instead.

Chip H.

Re:The BMW/MINI Cooper saga.

[sbaker]

I also have an 02-build/03-model-year MINI - and I don't have the problem.

So long as you don't get the ECU software upgraded to the 3.2, 3.3, 3.4 or 3.5 ECU software - you should be fine. The 3.6 version is supposed to have the fix - and versions prior to 3.2 don't seem to exhibit the problem.

Without reformulated gas, I doubt you'd ever see the problem - and even if you happened to visit a place that has it and bought just one tankful, I don't think that's enough to throw the ECU's algorithms off enough to show the problem.

I think you're OK.

Re:There are already too many vehicle LAN standard

[mhifoe]

Indeed there are a lot of different standards.
J1939 is even subdivided into several different sub-protocols. The heavy vehicle stuff (also used by the military) is J1939/71. There are also different versions for factory automation and agricultural vehicles. What's really needed is a CAN type standard (either new or an extension of a current one) to allow standard parts to be hung on a car's bus and used by all manufacturers.

I'm not sure if a fixed OS structure will be particuarly helpful. Different controllers in a car have different requirements. An engine controller is likely to be use a microcontroller rather than a processor which may well be completely incompatible.

The quantities involved mean that throwing in some more RAM is not acceptable. You can save enormous amounts of money by making something run with minimal resources. Consider that something like an MPC56x which is used for many engine controllers. If you can have your software running using just the 32k of onboard RAM you could save many million of dollars. Smaller items such as light clusters or airbags can have 128 bytes of memory and 1K of flash. A fixed OS structure is not going to help here.

Re:There are already too many vehicle LAN standard

[Jeff DeMaagd]

Does Microsoft really have much of a market in vehicle systems? I know they wanted to go there but in order to be there someone has to buy it and put it in their systems.

I think legally automobile components have to be available for a decade. This means that MS has to allow licence sales ten years after replacing the product, and support the product through whatever fixes may be needed, not just support and sell for maybe five years as they does now.

Plug N Play??

[Anonymous Coward]

Linux user, "Plug N' Play, what's that?"

You forgot AVC-lan!

[Medievalist]

My car has over 56 independent nodes running on three separate LAN protocols in a star topology centered on a multi-protocol gateway. Runs great, extremely reliable.

OT: Oxycontin Addiction--It Could Be You.

[istartedi]

Because he's an idiot

That's a very short sighted statement. Rush was prescribed Oxycontin for post-operative pain. Learn more about Oxycontin. Notice, it does not appear that anybody has kept statistics on what percentage of patients prescribed Oxycontin by doctors went on to become addicts. However, the marketers of this product (oxycontin is not a new drug, it's a different form of delivery for a synthetic opiate) have been criticized for over aggressive marketing to physicians (who really ought not to be swayed by marketing, but they're human too).

The bottom line? You too may someday be prescribed an opiate, get "used to it" and then find yourself begging for a "script" or seeking out a corrupt doctor who will "cash your script".

Don't forget: "Pride comes before a fall".

So, treat these drugs with the caution they deserve. Don't take them unless you absolutely have to. Start at the lowest dose possible. Always follow prescribed dosing, etc...

Of course you would never become an addict. You don't write code with bugs in it either. :)

RS-485

[mwilliamson]

I think a well-defined time-tested industry standard protocol like RS485 should be considered. RS-485 is a serial protocol that uses balanced signaling (good for noisy enviroments) and can handle multiple devices on the same wire.

What does my engine light mean?

[Anonymous Coward]

All I want is a readout of the error code instead of a generic "Check Engine Light". Why should I have to pay $70 to an auto mechanic to diagnose the problem when the car can tell me what's wrong?

Re:What does my engine light mean?

[sirsex]

Autozone will read the code for you for free.

Plug 'n' play

[tgraupmann]

What happens when the MSFT OS detects that your steering wheel, breaks, and airbag have all been ejected, when you fully know they are still plugged in!

Re:Plug 'n' play

[Anonymous Coward]

If it's a linux system embedded and the brakes (not 'breaks') go out, you just re-compile the kernel with the right options while doing 93 mph. And if you can't do that, then you're an idiot and you deserve to die.

Aftermarket Radios (wasRe:Windows ?!)

[akamoe]

will I even be able to install an aftermarket radio?

Hmm. depends. Some cars you can't now -- you have to keep your existing radio, or the car's computer will screw up. Like my Oldsmobile. Apparently there's some way of doing it that involves a re-location kit to mount it in the trunk, but that exceeds my attention span.

Think how this relates to the Internet

[RT Alec]

Standards for communication protocols. A variety of components from different manufacturers, all interoperating because everyone follows the standards. "Embrace and extend" is fatal-- your components won't be chosen by the end user (in this case, the car manufacturers).

This is a true driver of innovation (that word I always hear from one of our beloved software manufacturers). You know your product will compete on its merits, because any manufacturer (or, I suppose, car owner/tinkerer) can truly plug it in to almost any car on the planet.

"Reliable alternative!"

[Tokerat]

I prefer to use Homsar, the captain of the gravy train. It's a song form the sixties.

This would be a great venue for OSS

[chemindefer]

I have somewhere in my archives an article about a woman having to pay her Volvo dealer close to $100 per download for software downloads as part of the procedure to replace the electronic keys on her car after she misplaced the keys. Not only didn't the ~$500 set of keys and resetting of the computer in her car not work very well, but her old keys, which were supposed to be invalidated by the downloaded codes, worked perfectly when she found them. If more repairs are done electronically, and the software used is proprietary, it would seem to me that the dealers and manufacturers will be able to shut out the home mechanic, and force exhorbitant charges as described above. If a standard is adopted, it would greatly empower the owner to have it be open source, right? Will there be a day when the car comes with the equivalent of Windows on it, but there will be a way to install the equivalent of Linux on it, and then hack the prefs so the window buttons turn up the stereo volume? Or, more seriously, perhaps one could set the FI and ignition parameters on the fly, the way older cars had spark advance and choke levers? I know people offer chips for this, but it would be great if one could tune and set things with a laptop and a Firewire connection, or perhaps, ethernet (you would have to log in and sudo).

Re:This would be a great venue for OSS

[gimpyben]

I would assume that when the day comes that you can upgrade the OS on your car's ECU in your driveway, that some of the paperwork you sign for the warranty on your new car is going to get a lot more intricate. I think it would be completely reasonable to have a separate shop-rate for computer-related issues, if it were easy enough for the owner/end-user to change settings.