SUDDEN UNINTENDED ACCELERATION (SUA)

But do you and your wife “left foot brake”? If so....

 

I don't know any better than anyone here what happened more than the OP. However the incident happened, there is one part of the description that bothers me. That is that the OP, who was the passenger, claims to have been observing the foot on the correct pedal during the entire duration of the event. Really? This goes against all instincts to not ever look out the windshield to see what you might hit or crash into.

Mike

 

the chip accident was due to piling floor mats up where the gas pedal got stuck on top of the top floor mat.

in that model, the brake did not negate the gas as they do now. it had nothing to do with how much pressure he was able to put on the brake. and iirc, he was driving a powerful lexus, not a wimpy prius.

 

I think I agree.
Was just thinking the other day, how when I was growing up, you spent most of the entire day, evening, in the summer, outside playing, if you could. My parents usually had to call me inside for dinner, and you negotiated for as much time as possible outside playing. My childhood, saw the birth of the video game...as primitive as Pong-Atari...but I saw things change.

There's construction workers, drivers, farmers....but the number of jobs and careers that place you outside I would agree has dwindled.

I see people in scenic area's and parks, on smart phones.
I know more and more people that work from home.
I think I have to agree.

 

we were the same way, and i would still rather be outside than in. that's why we go to florida every winter.
when i was a kid, all our neighbors were outside much of the time, doing yard work, chatting with others, working on and washing their cars, playing with the kids and what have you. i don't see that anymore around here.
today, i see people, neighbors and friends, who never come out of the house. they get in their car and drive to the mall, supermarket, movies, children activities and etc.

i didn't read the article myself, but it does have a ring of plausibility.

 

There's something about this thread that's starting to bother me.

I really like my Prius (2010). I really liked my 2001. These are the first and only Toyotas I've owned, and I've had great experiences so far with each one, better reliability than anything else I've driven. I bought one for my Mom. So I'm about as much of a "satisfied customer" and Toyota booster as you're likely to find.

But there seems to be a giant pattern in this thread of posts dismissing any possibility of UA, seemingly bending over backwards to avoid mentioning Bookout v. Toyota or any of what the Barr Group found in the Camry firmware.

I can be a satisfied customer without having to pretend that every incident was shown to be user error or that there was never any kind of a problem.

Software-controlled stuff is one of the things I've been paid for, and I shared a lab for a while with a guy who earned his Ph.D. building a tool for worst-case stack depth analysis of embedded code (I don't know whether it was exactly what the Barr Group used in their work). That is to say, the Barr Group findings are kind of written in my language, and as much as I'd love to say they're nonsense, they ain't.

Now, there are some things that make a Prius different from a Camry. In the Camry, the tasks for running the engine and for monitoring the pedal positions ran in the same ECU. In a Prius, the ECM still has responsibility for running the engine, but the separate ("HV" or "Power Management", depending on generation) ECU is responsible for watching the accelerator pedal. That software might have a lot of differences from the Task X in the Camry ECM, so it might not have identical bugs; it might have different ones.

It's good even for me to go back and re-read that report from time to time. For one thing, it's a reminder of why fuzzy1's advice in #9 is pertinent:

It's not obvious why that would matter, but the Barr Group found that if, at the moment Task X happens to die, the brake pedal is already depressed even slightly, then until that status changes (that is, goes back to completely released, and then pressed again), it's as if the brake press hasn't been seen.

It's a lot like those web sites I visit with javascript turned off, and I select an option from some dropdown menu and nothing happens, and I think "oh, right" and click javascript on, but the site still won't go where I told it until I deliberately set the same menu to some different option and then back to the one I want.

One thing a Prius has going for it is the lack of vacuum-powered brake assist. In a vacuum-boosted car, you're in deep trouble if you pump the brakes during a UA, because an open-throttle engine isn't supplying vacuum, and the booster only stores enough for 2 or 3 pumps. After that, you're fighting an engine using unassisted brakes.

The Prius, by contrast, has the electric-pumped pressurized fluid accumulator, generally with enough fluid stored for 20 or 30 pedal strokes, so you won't be losing boost by letting up once on the pedal to make sure the ECU sees you press it again.

So, why am I still driving a car that I can perfectly well believe could have such issues in the firmware? It's mainly that I understand what it means for the combination of conditions to trigger a bug to be really, really rare, and I don't have much confidence that the embedded code in other cars or any of the other things that could kill me is much better. As long as stuff in the world is running on unreviewed, undisclosed, closed source software, my default assumption is that "IP protection" isn't the only reason the developers don't want you to see it.

(It is kind of funny, the lengths Toyota went to just to get particular facts redacted from all the trial transcripts, and then what wasn't redacted. They absolutely couldn't allow you to know how many milliseconds an ECM reset takes, but then a little further down the page you can see it's 11 feet at 60 mph. Hmm. The courtroom had to be cleared for anybody to mention how many tasks run on the Camry ECM, but it was ok to say there are 16 million combinations of which ones might be stopped. I notice the "final, redacted" Barr PDF now openly says 24 tasks; somebody must have decided what power of 2 is around 16 million wasn't as much of an impenetrable mystery as they'd hoped.)

Walking around the modern world with any kind of awareness of typical embedded code quality, or computer security, or both, is pretty much like having a sort of X-ray vision constantly showing you how close everything around you is to falling over on you. You keep walking around anyway. What else can you do?

If I were ever to experience such an incident in my Prius, I don't know how I'd feel about trusting the car again. The emotional part is hard to predict. The engineer part would probably still assume that the probability of another event was neither higher nor lower than it was for the first.

I do hope that, in the various firmware updates coming along from Toyota for these cars, as many bugs as have been identified are getting fixed. I can't expect that to mean "all bugs", and I also can't expect that the updates never introduce new ones.

There's another Barr Group pdf that's more a general presentation on risks in embedded systems. One of its first examples is the Therac-25, also the first such case study I was taught. The first real break in figuring out what was going on there was after the same operator had been at the controls for two of the accidents, and she actually remembered the order of her keypresses from the earlier time.

The current story in the news about the Hart eSlate voting machines in Texas also seems to come down to the order and timing of the user working the control wheel and button. Nothing new under the sun....

On the bright side, perhaps, given the way other risks are multiplying week on week in these times, in relative terms, my car is getting safer and safer with no effort at all....

-Chap

Edit: I also wonder if, after the recent expansion of the DMCA exemption for land vehicle firmware for repair, maybe some more insights into these unreviewed, closed-source masses of code might start to become available.

 

i think there would be less doubters if the claimants followed through more thoroughly.

personally, i'm not saying it couldn't happen, and i'm not saying it didn't happen. but without some type of confirmation, i'm allowing for other possibilities

 

I try to both be critical...and keep an open mind.
In this case the OP has disappeared and really hasn't answered any questions about how they plan to proceed or feel about their Prius moving forward.
So I guess you just say best of luck with "whatever"....

 

Nice!

 

Probably means Saturday the 3rd or Monday the 5th.

Looking forward to the update Monday at the latest .

 

I'm just curious as to your expectations and criteria for personal resolution of the issue.
Do you plan to keep the Prius?
If I really believed a vehicle had an episode of SUA, it would be hard for me to want to ever be in it again, or trust it again.

 

what do you do with a vehicle you believe could kill someone

 

maybe part it out (sans computers)

 

just read the green juice post. interesting that there was no conclusion, but he did buy another prius.

 

Keep us informed of your unfolding challenge/saga.