Unintended Acceleration Update *UPDATED

*UPDATED: After writing this yesterday, I came upon an interesting blog in the Washington Post by Frank Ahrens entitled, “Why it’s so hard for Toyota to find out what’s wrong with its vehicles.” The piece has to do with how engineers try to understand why mechanical failures occur and how software glitches complicate things.

If you put a lot of parts together to form a complex electro-mechanical machine and make it talk to itself via software, it can behave, sometimes, in ways you cannot anticipate. It can fail for reasons you cannot anticipate.

We also got a lengthy comment on these problems from Randy, a regular reader who took time to discuss some of these issues and explore the relationship of driver error to unintended acceleration. I urge you to read his remarks.

I want to follow up on some assertions I made last week concerning electronics and the “untamed beast” that may be part of the modern automobile. The runaway acceleration issue is still on the front burner for the general media, and for those of us in the industry media it seems to be unfolding from two points of view.

First are those who claim merely to report the news—outfits like AutoWeek and the Detroit News (“recall issues persist”) that have a vested interest in keeping the stories alive—and that group includes Toyota’s competitors, of course. Ongoing congressional hearings and investigations by the National Highway Transportation Safety Administration (NHTSA) and others (Dimitrios Biller, the L.A. Times, et al.) provide constant fuel for the fires of the controversy.

Then we have the writers, this one included, who attempt to explore a situation and give it a perspective—or make a point. I tried to say last week that there have been and will always be issues with electronics in cars, problems that can’t be easily explained, and people will die because of them.

I just don’t see how one can dispute this. It is in the nature of complex machines to fail. We do all in our power to keep the failure rate at an acceptable, Six-Sigma level. The fact that bad, poorly trained drivers contribute far more accidents and deaths to the horrendous totals is quite another point. Dealing with that is a massive, separate problem from dealing with runaway acceleration caused by what appears to be the “ghost in the machine.”

Some car media people have tried to confuse the two, saying there is no comparison between the relatively few numbers killed in runaway cars and the highway carnage caused by bad driving. That seems to be the point made in this video.

If you got all the way through this, you learned that these two donkeys haven’t the sense to distinguish between what is human error and what may be a serious safety defect. It would seem the first job of Congress or any investigative agency to make that distinction. The Audi controversy dragged on for years because it couldn’t do so. This one may follow suit, though let’s hope not.

So far, NHTSA “has linked 52 deaths to Toyota’s unintended acceleration problems,” and the stories continue to be quoted in blogs and the major media. Despite the claims of Toyota and others that electronics cannot be at fault, they don’t refute the stories people tell about their out-of-control cars.

Don’t blame the media for pursuing these stories. Blame them for confusing the issue of who and what is at fault.

Tell us if you ever experienced runaway acceleration in a car (and lived to tell about it).


Find Used Cars in Your Area at CarGurus


  1. No easy answer for the Toyota problem
    By Jeremy Anwyl Tuesday, March 16, 2010
    The writer is chief executive of Edmunds.com, which recently announced a competition with a cash prize for anyone who can demonstrate in a verifiable manner the reason for unintended acceleration.
    [Included chart in W. Post hardcopy issue: (Toyota and the accelerating mystery; page A19)]
    Discrepancy in complaints
    Consumer reports of cases of unintended acceleration by manufacturer
    (Model years 2005 – Sept. 30, 2009)
    Sales (in millions) Complaints per 100,000 vehicles sold
    GM 16.5 0.81
    Toyota 11.0 4.81
    Ford 10.8 3.12
    Chrysler 9.1 1.72
    Honda 7.1 1.26
    Nissan 4.6 1.07
    SOURCE: Edmunds.com analysis of National Highway Traffic Safety Administration complaints

  2. Video: Consumer Reports makes the case for brake override for all
    by Chris Shunk (RSS feed) on Mar 25th 2010

    We’ve seen demonstrations that show how to stop a vehicle experiencing unintended acceleration before, but this newest video from Consumer Reports demonstrates just how crucial a brake override function can be in stopping a throttle gone wild. CR engineer Jake Fisher lines up a pre-recall 2010 Toyota Avalon and a post-repair Camry to compare and contrast the amount of time it takes to bring a vehicle at full throttle to a stop. Hint: There’s a big difference.

    The Avalon went first, and at 60 mph with the throttle open, it took over 500 feet to come to a complete stop – nearly four times the distance of a normally operating vehicle. Even worse, when the brakes were pumped, the ability to stop the Avalon diminished greatly. Pump the brake two or three times and CR shows that you might as well be driving downhill on a sheet of ice.

    Next comes the Camry, which has been retrofitted with a brake override system courtesy of Toyota’s recall. The Camry stops at wide open throttle as though the gas pedal is totally disconnected. The result of CR’s little video shows that vehicles equipped with brake override can quickly come to a stop even when the throttle is pegged, making unintended acceleration a non-issue. Hit the jump to watch the six-minute video for yourself and let us know if you think all automakers should adapt this technology in the post-jump comments.

    [Source : Consumer Reports]

  3. I do agree that any throttle by wire vehicle should have a very basic failsafe the drops the engine to idle when the brake is depressed, regardless of the throttle pedal position. I assume that this is what Toyota has been trying to install through a software change, and I’m appalled that their original software didn’t have this feature– If this is true, it shows that Toyota’s System Safety Program is a complete failure.

    There are already quite a few proposed or in-production “active safety” systems, and these are things like adaptive cruise control, radar braking and active steering, that can actually mitigate driver errors. The main problem is the high expense of these systems, and the fact that you add yet another layer of complexity to the “beast”. Starting with the 2012 model year, all cars in the US will be required to have electronic stability systems, which also means they will have antilock brakes, but anyone who has spent a lot of track time with ESC vehicles understands that these systems help mitigate driver error but can’t prevent it. Ever wonder why NHTSA never required ABS systems on vehicles? Because their studies disclosed that drivers were never given the training to understand how the system works and how to use it properly, and therefore adopted a sense of “invincibility” where they felt their vehicle had a shorter stopping distance than a non-ABS vehicle. The result? More rear-enders. In fact, ABS systems DO NOT reduce stopping distance, they simply allow the driver to retain some ability to steer the vehicle during max braking events. I doubt that 1 out of 100 drivers with ABS understand this, or how to maximize braking with ABS. You can install an active braking system on a car but an untrained driver still won’t understand he or she can steer around an obstacle while the car is braking. We’re going to see this issue crop up again as all vehicles get stability systems, because drivers don’t understand what the system does and how to use it effectively.

  4. @ randy
    Thanks for your very thoughtful and detailed response, Randy. It’s great to have readers who will get into the weeds on a subject like this. There are indeed big, important issues at stake.

    1. I always thought the Audi problem was more than driver error, though that seemed to be the consensus after all was said and done. You’ve clarified some of that history for me.

    2. The engineering mentality: In my update note, I referred to a blog by Frank Ahrens in yesterday’s Washington Post which has an informative piece on engineering mentality and procedures for investigating failure. I think you’ll find that it aligns with your point of view.

    3: Airliners and systems interactions: This seems to be the heart of any solution, and several articles I’ve read have suggested an aeronautical systems approach might go far to solving these kinds of problems. But the interaction between driver and car seems to me the critical one to study, and I don’t know of many efforts to analyze and understand this. Do you? Which brings us to:

    4. Driver error. You say, “you really can’t separate the poor emergency driving skills of drivers from this problem.” But isn’t that exactly what has to be done? You can do the kind of engineering work required to (maybe) resolve the hardware/software/electro-mechanical aspects of the problem, though that will be difficult. But one must also account for the hows and whys of drivers making terrible errors—and that must be a separate, though related, investigation. In other words, the psychology of driver behavior ain’t the same as an engineering study.

    It would also be interesting to study why our society seems to devalue human life when it comes to driving cars. At the least, it speaks to the amazing predominance of the automobile in our culture, and our apparent willingness to “sacrifice” so much in its behalf. I was trying to suggest something on that order when I quoted Fuentes on the spirited and destructive nature of all machines.

  5. The Audi unintended acceleration problem was unique in many ways. It was a situation that was purely driver induced– Drivers pressed the gas pedal when they thought they were pressing the brake. The root cause, however, was unusual pedal placement, and incidents were biased heavily towards “secondary” drivers of the vehicles. That was the smoking gun that pointed towards the pedal placement. When the pedals were relocated, the problem stopped occuring. In many ways, Audi’s long insistence that there was no mechanical problem with the car was true, but as engineers often do, they were inflicted with tunnel vision because of their training and organizational structure. The engineers who designed the throttle and accelerator pedal system weren’t ergonomic engineers who analyze things like control placement.

    In many ways, Toyota is likely suffering from the same kind of tunnel vision that affected Audi engineers. The ECM hardware, throttle by wire and brake systems and associated software are likely designed and validated by a number of different engineering groups. And as with any big engineering organization, nobody wants the finger pointed at their work, especially when fatalities are involved. We could have hoped that an organization like NHTSA could have helped Toyota bridge the tunnel vision syndrome, except now we know that Toyota intentionally lobbied NHTSA to minimize defect reports. So NHTSA is no longer credible in this case, or any case as far as I’m concerned.

    In my own experience, I’ve never found ANY vehicle that could not be stopped by the brakes with the engine under full power. Not one, even a Corvette ZR-1. That’s a basic design principle used by automotive manufacturers. While you think a 500HP engine should be able to easily overpower any brakes, such vehicles always have brake systems that are sized to the performance of the vehicle.

    If Toyota has created designs that bypass this basic design principle and prevent the braking system from stopping a vehicle under full engine power, not only should that fact be exposed, but all of the cars potentially affected need to be removed from the road IMMEDIATELY.

    On the electronics front, aeronautical engineers can relate the difficulties of integrating layered electronics and software systems, but they’ve learned how to do it with great success. As complex as a car is, it’s relatively simple compared to a modern airliner, and automotive engineers need to follow the same design and validation techniques that are used in the aviation industry. Six Sigma is a commonly-used quality system, but it primarily addresses the hardware rather than the software and system interactions. This is probably where the problem lies with Toyota. They are champions at using SPC techniques to maximize component quality, but they may have met their match when dealing with the more subtle issues of systems interaction.

    And finally, you really can’t separate the poor emergency driving skills of drivers from this problem. This issue was highlighted during the Audi investigation and is a major contributing factor to the deaths and injuries in Toyota incidents. Any formal accident investigation MUST look at all the factors. In the case of vehicle control incidents, drivers are part of the chain of events and in most cases, the driver either propagates or mitigates the original failure.

    Our culture seems convinced that the enormous cost of hundreds of thousands of deaths and injuries is uncoupled from driver training, and that is why we continue to suffer. Is there anyone reading this that does not know of a family member or friend who has not died or been severely injured in a car crash? I doubt it.

Leave a Reply

Your email address will not be published. Required fields are marked *


This site uses Akismet to reduce spam. Learn how your comment data is processed.