cdynaco

So what happens when electro-magnetic interference from say cell towers, or wireless frequency overlap, or software glitches (those never happen right?), viruses that attack wireless sensors, or a terrorist EMP occurs? All of the world's computer driven vehicles will be worthless right?

Maybe the Cash for Clunkers thing pushed us closer to disaster... think I'll keep an old carb/non computer vehicle around jic...


Feb. 5, 2010, 2:36 p.m. EST
Your car computer may kill you

Commentary: Software has to be able to recognize a panic situation


By John C. Dvorak
BERKELEY, Calif. (MarketWatch) -- As more research is done into the recall of certain vehicles made by Toyota Motor Corp., the more likely it is that the sudden-acceleration phenomenon due to supposedly faulty gas pedals may actually be a software glitch.
In other words, an electromagnetic pulse outside the vehicle or a unusual series of commands to the onboard computer may trigger the hazardous condition.
Now we understand that the high-tech Ford Motor Co. (NYSE:F) models -- the Fusion and Milan, both hybrids -- have a software glitch that may indicate a brake failure when there isn't one. The cars will be brought back for a code patch.
When the concept of "fly-by-wire" cars (known in the auto industry as drive-by-wire, or DBW) became public, I was under the impression that some mechanical systems would remain in place, and that networks within the vehicle would be used for lights, instrumentation and engine fine-tuning.
Most people remember the 1988 incident at the Paris Air Show, where a fly-by-wire Airbus crashed into the trees during a demonstration. Although there is still a debate over the exact causes, some tech pundits are convinced that the computer, not liking the pilot input since it made no sense to the software, crashed the plane.
This very public accident is one of hundreds, if not thousands of software-related incidents, many grisly. Now we have the Toyota (NYSE:TM) "sticking" accelerator.
In a Prius, as well as many computer-controlled Toyotas, the accelerator pedal is more like a volume-control **** than anything else. In the olden days when you stepped on the pedal, it would be directly connected to, say, a carburetor, and open a valve mechanically as you pushed down.

This is now passé, as this activity is done electronically on the most modern cars by network signaling.
The original concept behind computer-networked cars began with engine-control units (referred to as car computers) in the 1980s, which handled the complicated fuel injection and other aspects of the modern engine. But car makers knew that the real benefit to them would to find some way to run a computer network throughout the car to get rid of the repair bugaboo and nightmarish cable harness/wiring assembly built into automobiles.
The cable harness had wires for everything electric within the car. A dedicated set of wires would go from the light switch, for example, to the headlamps and taillights. This was an expensive proposition when compared with a computer network with one wire looping around the entire car, telling every device what to do based on what switches (sensors) were turned on and off.
Because computer networks are easier to troubleshoot than a harness with a bunch of wires in it, cars got more reliable. So companies took the next money-saving step, which was to think seriously about a complete drive-by-wire car.
I can assure you that in the 1980s when this was first being discussed, this drive-by-wire idea was verboten . But hey, it works on airplanes, why not cars?
Of course, the software, if it is going to be used in the first place, should be coded to recognize a panic situation and shut down the engine or reset the gas pedal, or do something proper. Perhaps something as simple as a panic button or kill switch would suffice.
Unlike mechanical systems, software is nothing more than a lengthy series of instructions that require an inordinate amount of intense logical thought and second-guessing on the part of the software writer(s).
At the end of the day, you always get a bug. Apparently, at least for Toyota, the bug can be lethal.
This sort of problem can never be fully resolved or eliminated in complex systems. If we are going to continue to trust software and computers to run everything, disaster will forever occur. So get ready for this sort of problem to reemerge over and over in the decades to come.
The next step for the auto industry is drive-by-wire steering. I think we can already imagine how this is going to turn out.



http://www.marketwatch.com/story/your-car-computer-may-kill-you-2010-02-05?dist=afterbell

mustangGT90210

You've echoed my thoughts perfectly on the matter.

stangfoeva

I read earlier last week on autoblog or cnn (can't remember) that they can already stop car that use drive by wire using the methods you described above. They plan on using it to stop police pursuits and vehicle born IEDs with them...

cdynaco

Funny... I was rewatching Abyss the other night, and when the 'alien' ships came near they shut down all the human's electronics within a narrow range as they went by... and then the electronics were restored...

I thought what a potential military technology... limited range EMP that could be turned on and off... and how much more effective if the majority of vehicles were computer dependent so it doesn't just wipe your phone & computer, but keeps you from escaping in your car.

I'm not necessarily a conspiracy guy, but makes you wonder...

stangfoeva

Exactly. All I can tell you for sure is that the technology exists, and it is mature. If gets in the wrong hands.....lookout

ManEHawke

Carburetor with points ignition, and a CB radio. You are good to go.

houtex

"The Macintosh uses an experimental pointing device called a ‘mouse’. There is no evidence that people want to use these things." - John C. Dvorak.

So with that, I'm gonna just let you figure out what I think about Dvorak's musings, k?

Thanks to the computers in the cars, we're gettin' 300 HP outta a 281ci v8 at 17-25 MPG. That's not insignificant. And part of it is due to the disconnect of the pedal and throttle body.

I will admit, as a former software engineer, mistakes can be made. Humans are anything but perfect. But thinking that a mechanically linked throttle is better... my dad ran a 1966 Tbird through the back of the garage once due to the mechanical linkage gettin' bound up.

Oh, and there are MUCH better ways to make you not able to drive the car... stop letting oil show up, for one, and that hits you all, not just us 'puter cars.

/only scratched the Tbird a little... tough ol' bird that car.

cdynaco

Well at least we know who's gas to syphon first

While I understand there are many ways to cripple society, since I live well away from a 7.11, I need to have some way in reserve to get me closer to the chaos.

As for me, I've never had a manual throttle stick and the spring is default to close. In contrast how many Toy throttle's have stuck thus far?? And now "brake by wire" problems with the Prius?? And now Ford needs a software bump to correct its hybrid brakes???

So... I'm in my new '11 Diaper Stang 5.0 with electric power steering, and the DBW throttle sticks... so I turn the key off to save my bacon... and hopefully the brakes still work (since the Mustang's aren't electric yet...), but what about the steering since I turned the electricity off??

houtex

Most people assume that the power steering stops when the motor stops, and to that extent, they're right... but the steering can be made to work, it just takes more effort, helped by vehicle motion. I know... my 1966 Thunderbird had lost all it's power steering one time... and it still could be steered... it sucked a lot, but it never "lost" steering.

That car also lost it's power brake booster. I could still stop the car, it just took a lot of forethought.

Oh, and ask my dad about the time in Hawaii where he decided it'd be nifty to turn off the car and coast down hill... and couldn't stop the car very well without the power brakes... but the steering worked just fine, allowing us to narrowly miss the fence and he was able to stop the car... a lot later than he'd've liked... Which is vaguely reminiscent of a throttle getting stuck wide open and needing to stop the car with the engine off... nah, couldn't be...

Also, parking brakes, while not terribly awesome to stop, from a coast they can be applied and stop the car. Done it.

Your EPAS can go out, and the rack and pinion can still be made to work without it. Much like my '66 Tbird, it will suck, but it won't totally stop working if the car's moving. And that would be exactly like the hydraulic versions, by the way, so...

It's not drive by wire, and it's not very likely to ever be so, ditto with the brakes on a gas car. The throttle is the one thing I think they needed to do for efficiency purposes, otherwise, they would have probably left well enough alone.

Brakes on a hybrid for the regeneration? Hm... I'm thinkin' they're regular with the added bonus the motors reversing and becoming generators... and the drag from that is what's being fiddled with on the brakes... but if the car is off, the brakes will still work normallyish as above.

cdynaco

My winter car is a Nopar and its an 08 with all the computer goodies... still uses a cable throttle...

StangMahn

I like old school stuff, not for any conspiracy ideas or w/e, I just hate all the computerized stuff. Give me power nothing (other than brakes and steering, lets be reasonable), no computers or sensors or automatic thingamajigs. I want a car that you have to know what you're doing to be able to drive it. People are becoming worse and worse drivers all the time because cars are becoming easier to drive, and that sucks.

Slims00ls1z28

It's just a Toyota thing. All GM LS engines since 05 are DBW. Never had any probs out of mine yet. F-16's been using it for a while for every function too. Now if the EMP hit's I'm screwed however I'm sure computer failure would be a bigger stopper than the throttle body.

hi5.0

Got to be software issues - something Toyota would probably be loathe to admit. Would rather blame mechanical failure on a parts supplier. Even if they were supposedly built to Toyota's specs and "passed" Q.C.

houtex

Probably because they're a little behind in gettin' to that model of engine/car/computer, and/or it's efficient enough, so what's the point? I'm no engineer though.

Glenn

I have always thought the yota thing is more than a faulty pedal.

1chocophile

I agree, Glenn. If the toy throttle is drive by wire, it seems the computer should be able to detect a throttle error (e.g., throttle and brakes applied at the same time) and compensate for or override any mechanical failure.

Antigini-GT/CS

This conversation will be interesting in a few years when BMW goes to brake by wire.

bob

Man, if you guys are looking for the ultimate in reliability, then its mechanical injection diesel and run flat tires all the way, no electronics at all!

Now I'm not saying I'd given this some serious thought, but I remeber reading about EMP devices being tested for police use years ago and for whatever reason the conversation just naturally gravitated to "well if I robbed a bank and needed to make sure I could absolutely get away"

However, to don a tinfoil hat for a second, I'm more suspect about OnStar and similar services where the doors and ignition can be remotely controlled on the off chance I might be in the middle of a Ruby Ridge scenario.

Slims00ls1z28

Lol ya but bump the mechanical injection slap a dominator on it!

I remember seeing a show on A&E that they had a remote controlled car with 2 prongs on it that would short a car as soon as both hit the underside of the vehicle. Supposedly in a chase cops could drive it under the fleeing car to stop it. Don't remember the specifics of it because it was a long time ago and before I knew much about cars.

stangfoeva

I found it!

http://www.engadget.com/2010/01/20/e...marines-video/