Software Glitches Stall Toyota Prius 560
t35t0r writes "CNN/Money/Tech reports that 2004 and early 2005 Toyota Prius models have software bugs that cause them to stall while traveling at highway speeds. While no accidents were reported to have been caused by the software glitch, could we be heading into an era where our automobiles will require software updates and fixes to keep them from literally 'crashing'?"
Failover (Score:5, Insightful)
Re:Failover (Score:3, Insightful)
IIRC, some of the stealth bombers will fall apart in less than a second if the computers go.
If the fuel injection is gone because of the computer crash, what do you fail over to?
There is a reason VW Beetle (Score:2, Insightful)
is still the world most reliable car
it has nothing to do with electronics
Car Firmware Upgrades and Rebooting on the Road (Score:3, Insightful)
Back in the 80s, I had an old beater 1971 Chevy Van with the usual Weird Chevy Electrical Problems. Every once in a while the engine would stop running while I was driving down the road (which is a problem for power steering...), so I'd put it in neutral and reboot, which would usually work. My current van is a 1987 Chevy, with a new engine installed about 5 years ago. The engine's not quite identical to the original, and every once in a while the monitoring system decides something's wrong and turns on the "Service Engine Soon" light, typically when I accelerate to pass somebody while going uphill on a freeway. There's no harm done, as long as that's the cause (as opposed to something actually being wrong with it), but to turn the light off you also have to reboot the car.
Updating software (Score:5, Insightful)
Without putting too fine a point on it, yes! But there is no reason to go all chicken little. Standards of reliability for automotive software are generally much higher than for desktop PC software. No EULAs and auto manufacturers generally can not disclaim warranties. If a car breaks down due to crappy software, Consumer Reports will put out a report and people won't buy it. Additionally there are Lemon Laws and lots of eager lawyers to protect consumers. Unlike PCs where we have been trained to expect crashing software, people don't put up with that in cars, especially since there is the potential for physical harm when hurtling down the road at 80mph.
Re:BMW?? (Score:5, Insightful)
Re:Failover (Score:3, Insightful)
How is this a big problem? Have you never had a car stall and these things fail on you before? It's no big deal. You push the pedal a little harder and you put a little more effort into steering.
Perspective (Score:5, Insightful)
Re:Failover (Score:5, Insightful)
It did. At least based on the anecdotes posted at edmunds.com by the drivers. The engine shut off, the dashboard lit up like a Christmas tree and the battery continued to power the car. Not surprising that you might conclude total failure from the
Guess what folks; you are expected to be capable of coping with vehicle problems while traveling at the phenomenal rate of "highway speed". Tires blow, people fuck up, things fly off randomly; deal with it.
Re:BMW?? (Score:3, Insightful)
But most of us assume that part of the extremely large cost of those planes is in both more reliable technology and increased redundancy. I think the systems of a Boeing 777 are probably held to a higher standard than a Mazda or even a BMW...mostly due to the more catastrophic nature of a failure.
Doesn't mean we're right...maybe the systems on a BMW are every bit as reliable as on a plane. But it would still explain this reaction.
Engineering philosophy (Score:3, Insightful)
Regular devourers of world news will recall that a few years ago, Bridgestone/Firestone got sued for producing tires with a propensity for exploding. A few years before then, there were horror stories of malfunctioning cruise control that would activate itself due to a short-circuit, with no way to switch it off.
Actually, a similar fault to that last one even appeared on the Space Shuttle - the last launch window was scrubbed when it was realized that the attitude rockets could fire themselves, even when the power was switched off.
Engineering to build fault-tolerent systems (ie: systems that will still behave sensibly, even when something goes wrong) is expensive, difficult, time-consuming and requires enormous resources to cover every possible aspect.
Even when faced with the prospect of multi-million dollar lawsuits for death/injury, it is often cheaper to simply let people die a torturous, firey death in agony than to prevent such incidents from arising. Because we live in a competitive world, where success is measured in dollars, there is simply no incentive to get things right. Getting things affordably wrong is a far more profitable approach.
It would be possible to build a car that can do 100 miles to the gallon, be able to keep the occupants intact after a 150 mph head-on collision (F1 monocoques can handles 240 mph collisions) and have software driving every aspect of the system that is not only 100% free of bugs but is able to adapt to handle the natural degredation of the hardware. Such a car would cost about as much as a NASA Space Shuttle and don't expect the insurance to be any less, simply because of the theft value.
A company producing such a car might sell as many as one. The McLaren F1 road car would be much more affordable but is wtill somewhere in the low double-digit sales, and was reportedly still in single-figure sales at the end of the first year.
Having said that, I think that it should be mandatory that car companies produce the very best they can. Failure is not only an option, it's often so cheap that it's the best option. That should not be the case, ever. Bugs in software and failures of hardware are going to happen in the Real World, but they should not be encouraged. Good practices, good designs and thorough design reviews should be the norm.
Re:Cars already need this.. (Score:3, Insightful)
I call BS on this one.
Re:Irresponsible article! (Score:2, Insightful)
Re:BMW?? (Score:2, Insightful)
Yuck! A better solution is to stop issuing crackerjack licenses to the idiots.
I for one do not want my car second-guessing or overriding my control inputs in a vain attempt to keep a potential idiot, who shouldn't be driving anyway, from being idiotic.
Larger picture (Score:5, Insightful)
Electronic systems are, in general, more reliable, with lower failure rates, than the mechanical systems they replace. They are also easier to service. (Though the repair bill may very well be higher, and specialized equipment may be necessary.)
This "software", as others have said, are not the same as the software we run on our PCs. The software quality standards are higher, and the testing is far more intense.
People lament the loss of simpler mechanical systems that can be fixed with know-how and a socket set. We publicize every example of a system failure we hear of. But the numbers don't lie: a 2005 model with a half-dozen embedded computers has a far lower incidence of problems than a corresponding 1970 model when it was new. You are far less likely to ever have to call a tow truck in your lifetime than your father/grandfather was.
Sensationalism is so much more fun than fact, though.
Re:BMW?? (Score:3, Insightful)
Re:Failover (Score:1, Insightful)
If this is the case, you're driving way too fast for the conditions. You should be able to react to anything happening on the road quickly enough to avoid death and destruction.
Re:Failover (Score:1, Insightful)
Wrong again. ABS often increases stopping distance. The advantage you gain is that it allows you to steer.
Get your friends to push you faster and you can steer GTO just fine.
Just another problem... (Score:2, Insightful)
Isn't that what recalls and engine checks have been doing all along, except for physical parts?
Re:Shouldn't have stolen that code not informative (Score:5, Insightful)
Whenever I get a 'new' car, I run down to the nearby college at night and find an empty lot and slide around a bit, and see what happens when I turn the engine off and if I can turn the key back and have it start magically, aka, a push start, which is incredibly useful if your car stalls while you're driving down the highway. (The other option being a normal start in neutral, but that takes much longer. And wouldn't work if your battery was dead, but that's a rather worse-case scenerio.)
Then I come back and do it again when it's raining, solely for seeing how it skids.
And if I have a car I've never tried it on, and I'm on a completely empty and straight stretch of highway, I kill the engine there, too, to see if it does something different at high speeds. (That's probably a traffic violation, but if a cop appeared out of the blue, I'd just say I stalled for some reason.)
I will admit I've never tried to solve a hypothetical 'stuck pedal', but, OTOH, the parking lots aren't really big enough for that. It's a good idea, though. I know I can shift into neutral at any speed, but I agree that cutting the engine is better...for one thing, it should let the engine slow down the car. I'll have to figure out some way to test that.
Do people really drive around in a ton of metal and not know in advance how it operates when bad things happen to it? When, exactly, are they planning on learning? The time to learn what happens when you slam on the brakes on a puddle of water is not in the middle of traffic. I once had an early antilock system that pulsed the brakes really oddly...there was a lag between losing traction and the unlocking of the brake, or something, I never really figured it out.
I mean, there are somethings you can't learn until they happen, for example, if you really need to stop the car, you can switch into park when you're going 20 mph, but you'd obviously never want to do that unless you had to. But what happens when your engine cuts off, or if you hit a patch of water while turning? Everyone should test that.
Re:Shouldn't have stolen that code... (Score:5, Insightful)
I don't know why so many moderators thought this was funny, but you likely have it exactly right.
It happened to me once on an icy road when my car started drifting. I thought I had my foot off the accelerator and on the brakes, but didn't realize why the anti-lock system wasn't working and the engine was making so much noise until I was sliding into a ditch. There was no damage and I was able to drive out, but at that moment I knew exactly how people can believe they had their foot on the brake.
Unfortunately, my mother wasn't so lucky. She got the pedals mixed up while manuvering in the driveway behind the house and ended up parking in the neighbor's bedroom (fortunately, no one was home). When my father ran outside and shut down the ignition, she was dazed from the impact, but her foot was still jammed on the accelerator.
Re:Shouldn't have stolen that code... (Score:4, Insightful)
Re:Larger picture (Score:4, Insightful)
I've spent a lot of time looking Consumer Reports' car reviews, and frankly, their reviews have no basis in reality. Cars that were rated as extremely noisy, may be quieter than cars rated as quiet. I know this from experience. Go read their reviews of a few cars you have owned, and see if they match reality.
Now, you can certainly attribute that to different reviewers having different standards, or to individuals' biases, but it seems to be very widespread, and seems obvious to me that it's always in favor of one brand or another. I can't prove it, but it looks very much like certain companies (eg. Toyota) are getting far more favorable reviews than the best cars of other makes, even on their poorest offerings.
I've come to this conclusion long before I came to this thread, so the fact that the car in this case happens to be a Toyota is coincidence.
That's really not true. I've never had a mechnical cooling fan fail on me, while I've had a few electric fans fail. Which would you will find in trucks? Not electric.
Now that's just blatantly wrong. The only problems that are easy to fix on computers are the problems introduced by the computer. For instance when the warning light comes on, sensors go crazy, etc. You may also have heard of cases where certain models of cars will drive fine for 80,000 miles, then like clockwork, start running too rich/lean/etc. It's debatable whether car manufacturers are intentionally inducing these faults, but it is not debatable that these faults are there, and tricking the computer can commonly fix problems the computer has caused.
The numbers don't lie, but you sure do. Those numbers are certainly due to improvments in engineering vehicles, (manufacturing/design) improvements to mechanical parts, improvements in fluids, materials, etc. It's only common-sense that those numbers would improve.
Also, the
Those numbers are also deceptive, because people don't complain much about computer problems with new cars. Give me numbers when those cars are 5 years old, the we'll see.
I'm tired, so I'll say the one thing that will end this arguement instantly: New BMWs
Re:BMW?? (Score:3, Insightful)
It can also be the other way around. Take an example where someone is driving the speed limit in the left lane of a major urban expressway. On most of these roads, when traffic permits, the left lane moves at least 10 mph faster than the speed limit. Someone driving the speed limit, obeying the law, will cause drivers behind them to back up and try and go around on the right side, which creates a hazard.
What you say about traffic/accidents being caused by one person separate from the "herd" is correct.. it's just that the "herd" may not necessarily be the ones following the law.