If it was driver error alone, though, we would see it happening all across the industry, since many makers use the same systems now. When only one maker is having a large number of these incidents, the claim of human error becomes statistically unlikely. Case in point - there was a claim of sudden acceleration for the Toyota Tacoma as well. The NHTSA has over 30 complaints in a two year period (2006-2008) in its database while competing brands for similar trucks had one or none.
Huge isolated spikes like that are not going to be simply human error alone. Which is always what every maker blames it on when it does come to their attention. "Not our fault - must be human error". In every case, these complaints have occurred in vehicles with faulty drive-by-wire type components or malfunctioning cruise control systems. Both of which are computer-controlled in Toyotas.
Manuals never have this happen because they simply don't have these systems in place.(yet) As far as I can tell looking at the data myself, there have been no reported cases in the last decade of reported sudden acceleration for vehicles with a manual transmission in the NHTSA database. This also further reinforces it being statistically unlikely that it's simply human error and instead points to a design flaw that happens only in automatic equipped vehicles.
You can blame people all you want. I'll believe what the math says is likely instead.
I said human error plus scammers. How do you explain the steep drop in alleged sua's after a couple of the scammers were outed? Looks like the problem has gone away to me.
A lot of desperate people trying to make a few bucks, and of course, they are zeroed in on Toyota.
That's interesting, but to me the lack of data on manual transmission cars DOES suggest human error is the main cause.
Also, the fact that the UA reports do not cut across all makes of cars might suggest that the design and placement of the control pedals is a major factor in causing the human to err.
I would *like* to accept the "theory" of computer error, but applying occam's razor to the question suggests human error far more likely. Why? Because in many of these UA cases, the car runs away, is finally brought to a halt after a terrifying ride, and then the simultaneous failure of throttle control systems AND braking systems suddenly correct themselves and are never, ever replicated by third parties.
Now really, if I presented you with a similar case involving, say, your refrigerator and why the ice cream melted, might you not think that someone left the door open rather than a mysterious malfunction that only happens when one person is home?
Of course, with no hard data or access to the software coming from Toyota, we're essentially stuck
From a recent court order story:
"Both sides spent the better part of Thursday's hearing arguing about a proposed discovery plan. Galvin raised concern about the protective order surrounding Toyota's "source code," the software underlying its electronics system. Toyota considers the source code "the most valuable asset it has," he said.
Robinson told the judge that he thought that both sides had reached an agreement to provide the source code as long as it was in a secured and neutral building that has been used by the Department of Defense and the Central Intelligence Agency."
First the NHTSA called in NASA. Now the parties are using old DOD and CIA digs.
You'd think that Occam's Razor would apply, but the simple fact is that it is simply a garbage theory that is routinely touted by armchair pundits and popular culture that has no idea of the real limitations and reasons why it is nothing more than a cliche'
http://en.wikipedia.org/wiki/Occam%27s_razor Specifically, it is useful as a heuristic in developing models but is not valid as a method for proving them. Or disproving them, either. *quote* "It has been suggested that Occam’s razor is a widely accepted example of extraevidential consideration, even though it is entirely a metaphysical assumption. There is little empirical evidence that the world is actually simple or that simple accounts are more likely than complex ones to be true." -and- "As a methodological principle, the demand for simplicity suggested by Occam’s razor cannot be generally sustained. Occam’s razor cannot help toward a rational decision between competing explanations of the same empirical facts." **** "Of two equivalent theories or explanations, all other things being equal, the simpler one is to be preferred." - this is the commonly used example of Occam's Razor. You can't make the statement that "human error" is the simpler explanation in this case. In fact, the assumption that humans are automatically more fallible than machines is a classic case of poor reasoning that permeates our entire modern society.
The NHTSA database goes back for years. The general pattern of such claims always has two major factors.
1: It comes in large groups that affect one manufacturer at a time. In almost every case, it has (eventually) been proven to have been a defect that caused it. Most of the time it is a TPS or a cruise control system that has malfunctioned. Now, this alone doesn't mean anything concerning Toyota's claims, but yet again, we have a sudden spike in activity for one manufacturer and only on vehicles that have similar control systems in place.
2: While the propensity of humans to lie about something like some people are suggesting because it's in the news is a known factor, it's also been routinely shown that when such a series of claims is made, there also is a fair amount of bleed-over into other similar items or manufacturers. Because these cases have almost none of that, it points to something other than solely human error and greed. There is/was a legitimate spike in the number of cases reported, even if you discount all of the potentially questionable claims in their entirety. The few hundred that remain are alone enough to statistically prove that there still is an anomaly at work other than human error. Even if you factor in typical levels of human error in other similar cases, historically.
The judges and the Government realize this, of course, being well versed with statistics. Based upon Toyota's response to the normal by-the-book discovery process this week, they are hiding something. I know because I used to work in the e-discovery field. The only time anyone ever reacted like this was because they were guilty. In fact, the laws surrounding discovery are written with this in mind. Failure to undergo the process, destruction of evidence, or even failure to have backups of said evidence is considered grounds to simply lose the case in its entirey by default.
Toyota's claims are groundless. The legal discovery process is done every day and is extremely secure. They are playing games for a reason. The only reason we can't force them to comply is because they are technically a foreign corporation.(despite having plants in the U.S. and sub-corporation that runs them that's also based here)
No one said that picking the simpler solution over the complex one ever "proved" anything---who would say such a thing? The only person "suggesting" that is the one writing the wiki notation.
When applied to diagnostics, it is simply common sense to eliminate the simplest explanations before become embroiled in complex ones.
Nothing "pundit" about it. That's a great way to fix cars, and it works pretty darn well---not as a sure-fire way to find an quick answer, but as a very reliable way not to waste time and money.
e.g., if a car doesn't start, you don't replace the computer before you check the battery cables. That's Occam's razor.
That's all that Occam's Razor implies to me. I can't speak for all the people on wikipedia who want to fog the idea over.
A far more common human error is when one assumes that the answer to a baffling mystery is always complicated. FINDING it might be complicated, but often the answer is not.
You know, if I was forced to choose sides between a person proposing the simplest answer or the one proposing conspiracies, I'd tend to give more credibility to the simplest answer. Conspiracy theories are, in a sense, a clever way of giving up rational inquiry without seeming to.
Something I am involved with through work may help shed some light on the difficulties involved in troubleshooting errors that occur very infrequently.
I have been working on a system for over 6 years now. We have been through the (what some may call) the engineering phase, and iterated the design for a deliverable build. The complete system was integrated last February (2009) and has been undergoing various testing procedures that culminated in the system being delivered in August of this year.
In February of 2010, one of the assemblies exhibited some anomalous behavior that would be consistent with a loss of power or a reset occurrence. Note that I said "consistent with", because to declare either as the root cause would have been premature.
The anomalous behavior did not show up again until early May, when we saw 2 occurrences over several days. Then, nothing more. So here we have 3 "failures" over several thousand hours of operation or so. After some head scratching, we removed the offending unit from the system and replaced it with a spare. The system completed the remainder of it's testing error free and, like I said, was shipped near the end of August.
The removed, suspect unit was placed on a test set and run continuously, 24 hours a day. It has exhibited two similar failures over that period of time. On the first our instrumentation did not "trigger" properly, so we did not catch the parameters we were looking for. The second occurrence we did capture the parameters of interest and are in the process of refining our probe location on the hardware to try and narrow the problem down to one of the several power supplies involved.
So, these kinds of infrequently occurring problems can be very, VERY difficult to 1) duplicate, and 2) track down. Tracking the problem down to it's root cause is usually not that tough, once you can come with a test arrangement that causes the problem to occur on a more regular basis than what is seen in normal use. And that is the problem with the UAE incidents - nothing that has been done has been able to, to my knowledge, increase the rate at which the UAEs occur.
That's very interesting stuff. So you don't think that by testing Priuses on public roads for a couple million miles with multiple drivers under all types of driving conditions--that this approach would "prove" or disprove something? I mean, how much better a test than that, which is bound to duplicate the mysterious parameters somewhere, somehow, sometime.
If such testing showed nothing, there are actually two conclusions: 1) driver error or 2) a level of risk so minute and unlikely that no legal liabilities could be applied -- e.g., a law saying you can't throw anything overboard on a cruise ship at night in the middle of the ocean because you might hit a shipwreck survivor in a rowboat on the head.
SECOND QUESTION: Has anyone ever seen a report of a UA incident happening to the same vehicle TWICE?
No one said that picking the simpler solution over the complex one ever "proved" anything---who would say such a thing?
Actually, you inferred as much by claiming that Occam's Razor was a valid reason to chose human error over mechanical or software error.
But as anyone who works with designing and troubleshooting software and/or complex machinery will tell you, it's only a very basic method that's used to narrow down choices. It cannot be used in cases like this, as you are presupposing that human error is the "simpler" explanation. Many young engineers and university students unfortunately latch onto this idea and don't realize that it also can be a trap. IMO, there should be a name for this scenario of being overly enamored with simplistic theories and ideas that are taught in college. Generation after generation learning the same nonsense that has to be un-learned all too often.
Occam's Razor has many limitations, and in this case, it can be shown via statistical analysis of the data that there are other likely factors at work besides (or in addition to) human error. Remember the "All other factors being equal" part of it? Well, if there is any sort of statistical anomaly AT ALL, then it cannot be used until that anomaly is dealt with and the data is normalized.
You're flat out wrong here. Simple as that. Until we also have a thorough analysis of the hardware and software that's involved, we cannot make any statement either way as to whether human error is a more valid explanation.
But the second issue is more telling, IMO.
Of course, Toyota is stonewalling, which is a good indicator that they know something that they don't want investigators to know about. This is always worrisome as it points to them trying to skew the data to their benefit. Perhaps this is typical corporate posturing by them, but given the number of similar actions by other companies in the past leading it to being a legitimate case of hiding something, one has to wonder what's really going on. As I pointed out, the legal presumption is that failure to comply with a request for discovery is seen as automatically hiding something(as well as breaking the law), because the only times where it happens has been when they are actively trying to obstruct justice. It's also why the judges refuse to let it die and Congress is getting more and more angry about it(almost half of Congress is made up of lawyers who take a dim view of refusing to comply with the law)
Simply put, if they had nothing to hide, they would show that they have nothing to hide. I think the FBI and CIA can manage to keep their "data" secure.
Nay, nay----I'm not even "implying" any such thing for Occam's razor. I'm simply stating that the simplest solution is the one to look at first. I really don't know how to make this any clearer. Maybe I'm just not very good at words today.
So let's unpack what is known so far:
On the one hand, we have some UAs, all with automatic transmission cars, and many of which were later examined with the UA cause neither detected nor ever reproduced in, I believe, 100% of the cases.
Theory 1: driver error
Theory 2: defective floor mats
Theory 3: an unexplained, undetectable undefined computer glitch.
Why should I believe #3 "more than" 2 or 1? Shouldn't they have equal weight at least?
Of the three theories, driver error actually explains what was observed better than the other two.
Is it "true"? We don't know (yet).
If I'm reading your ideas correctly, your defense of theory #3 is based on a suspicion that Toyota is withholding something, and that "something" holds the answer.
That feels very unsatisfactory to me.
Or are you saying that if an answer is never found, that doesn't mean theory #3 isn't right?
However, if an answer is never found, what good is theory #3?
1 - Defective floor mats is a tall claim since these same mats are not present in vehicles like the Tacoma. It can be ruled out the easiest. 2- Human error may be a factor. It may not. 3 - Mechanical failure cannot be ruled upon at all, since we don't have any data yet. Basically we're trying to make a decision between two ideas with only one set of data.
Simply put, the test is only halfway done and we can't test the hypothesis yet. Talking as if it's already been decided is premature. Whether or not other evidence has been found or not is irrelevant. We have a competing theory that is still awaiting data to come in.
As for the legal process, I worked for five years in this exact aspect of the law. Discovery is a by-the-numbers process that's well understood by any company today. If you are failing to comply with it, you are assumed to be doing it willingly and the chance that it's a mistake is a statistical impossibility if your company has even a single lawyer on staff. Short of your data center burning down and your backups also burning up at the same time, it never happens unless you are hiding evidence that you know will incriminate you.
It's such a strong statistical improbability that failure to comply (if you are a U.S. corporation)pretty much automatically looses you the case. You might not like it, but it's how the law works in the U.S. as well as most of Europe. You either comply or your are judged to be willingly obstructing justice. Telling the FBI, CIA, the Justice Department *and* Congress "no" would be insanity for any U.S. corporation to contemplate.
But Toyota is not a U.S. corporation and is using that to play games with established legal procedures. If it was me, I'd threaten to embargo their cars and shut down their factories in the U.S. until they comply. But then again, I doubt if our current Congress has the ability to play hard ball with them like they should.
Having a bigger data base to work with is always better, IMO. The more vehicles, the more drivers, the more miles driven, can be a help in ascertaining what exactly is going on.
But, on the other hand, the problem with real world faults(?) is that the conditions under which they occur are uncontrolled. What was the weather? Was the vehicle going past a power station? Was the driver on the cell phone? Was he moving the windows up and down, and adjusting the power seat?
Along with this, people in general are notoriously unreliable when, after the fact, recounting exactly what transpired. How many stories have we all heard about eyewitness accounts being shown to be inconsistent between multiple eyewitnesses, or the account being discredited when an unbiased source such as a surveillance camera show that something different occurred?
"You say you were reaching over to adjust the volume of the radio when the car suddenly accelerated. Did the car start accelerating before of after you turned the volume control?" "Well, I'm not sure, I think it was after, but it could have been before".
I am in agreement with your comment about the level of risk being so small. Even if the problem does turn out to be a defect (using the term loosely), with billions of vehicle miles and billions of hours of operation to only have a relatively few number of UAEs, I would proffer that such does not rise to the level of negligence. People expect technology and things they use, whether they be autos or passenger planes, to be 100% safe 100% of the time. Such expectations are unrealistic, and need to be adjusted.
Robinson told the judge that he thought that both sides had reached an agreement to provide the source code as long as it was in a secured and neutral building that has been used by the Department of Defense and the Central Intelligence Agency."
That should make Toyota feel safe about their code in light of Wikileaks. Then what auto maker would want to steal defective code??? :shades:
People expect technology and things they use, whether they be autos or passenger planes, to be 100% safe 100% of the time. Such expectations are unrealistic, and need to be adjusted.
Brake override systems should help with the safety issue. Looks like most all the automakers are installing them.
That's how I feel about it and how I think it will all turn out. Even if, someday, in a galaxy far away, someone does figure out that there is this computer glitch, perhaps combined with atmospheric or human influences, could cause a UA 1/1 zillionith of a time, this does not constitute any legal negligence. The risk factor is far too small.
To rule otherwise would essentially paralyze the very basis of production of goods. This is *especially* true with those products that humans interact upon.
If you are afraid of toyota products, then don't buy one.
Since you have zero evidence for your claim of "mechanical failure" you really have no justifiable basis for your claim.
You might just as well add "what they ate for breakfast" as a possible cause also, since there is zero evidence for that as well...and would be extremely hard to prove it is not the cause.
Considering that we *can't* get at the computer code so far to even analyze it for faults, it's also true that we can't rule it out either. Any competent person looking at a scenario where you are unable to analyze the software controlling a piece of machinery would of course wait until they had the ability to do so before passing judgment.
I think the real problem here is that you and many other people don't really properly conceptualize that that it's not "mechanical" and "software", but more of a robotic system that functions together. So far, in essence, we've only been able to poke at the exterior and look at the shiny wires and actuators. Our job of even ruling out the "mechanical" aspect of the system is actually only halfway completed.
It seems perfectly sane to want to complete this part of the examination, doesn't it?
*BUT* Non-human faults have always been a component in cases like this historically. With every previous set of major incidents to date being related to a defect, it's something that we cannot overlook and just say "there's zero evidence so far". Especially when Toyota is refusing to give access to investigators.
IMO, that's a giant red flag. Toyota could comply but simply is making up excuses at this point. If this was a criminal case, the DA would add a charge of obstructing justice(at the least - they love to tack on extra charges). It's exactly like the following scenario: "So where is the gun that you have registered in your name?" "I don't know and I'm not telling you." (cue some very odd looks from the prosecutor and judge)
or: "So where is the computer code that controls the system?" "We're not going to let you have it." ???
My opinion is that if all the king's horses and all the king's men couldn't reach a determination using field testing then they certainly wouldn't find it examining minutiae. All you'll get from digging into computer chips is a 500 page report of computer jargon with not a scrape of a definitive answer (my prediction). It would be full of "mays" and "could haves".
If nobody can make a Toyota do a UA, other than the owner of the vehicle, well then, I'd say that Toyota doesn't owe anybody anything.
"Mr. Plekto, your gun is defective. I shot myself with it"
"Okay, Mr shiftright, show me how it is defective"
"Well I can't make it happen again, Mr. Plekto but we need to know all your metallurgical formulas so that we can melt it down and examine the basic structure to see if certain thermodynamic and climactic forces might have caused a slight distortion in the firing pin under laboratory conditions, thereby possibly causing a delay in the strike against the bullet's primer, and also possibly a slight deflection of the pin's path if the gun is held in a 45 degree position after removal from a warm holster".
Ok, I'll bite. Why wouldn't a brake override system prevent UA situations?
Toyota says "When the vehicle throttle is opened beyond the idle position, at speeds greater than five miles per hour and then the brakes are firmly applied for longer than one-half second, the override feature will reduce engine output to the idle position, allowing greater braking performance."
I think a lot of the sua's are caused when the driver tries to hit the brakes but hits the gas instead and never touches the brakes. The vehicle only stops when it hits something. The override would not help in those situations.
Then the driver would claim that the override did not work and here we go again !!
The same thing has happened to me more than once. If you are alert and paying attention you can tell what's going on but big boots and driving just do not mix. In that situation the override would be effective.
Plekto, you can say that again, why is it that only toyota is facing many lawsuits abt these matters, why are Nissan, Hyundai, Honda, Ford not sued ?
So only toyota drivers once in a while wear big boots and stepped on the wrong pedal ? So Hyundai, Nissan, Honda drivers don't wear big boots ? So only toyota old drivers screw up ? So there are no old drivers driving Hyundais ? Nissans ? Give me a break !
Look at the statistics. Hyundai, Honda, Nissan customers are no more loyal or fanatical to their cars than toyota owners. If Nissan, Hyundai etc cars have the same kind of problems that toyota have today, trust me, they WILL sue !
Actually, you just made my point, even if you didn't realize it at first. Or maybe you were being sarcastic...
Owners, if they are doing this just to create a lot of press and false claims, would also be attacking other manufactures. Because it does happen from time to time with other makers as well.(usually a defective TPS unit) We should see a similar number of false claims from idiots who can't figure out what pedal is which(because they almost always won't admit making such a mistake).
But they are not.
The sheer number of claims cannot be all lies and human error. There is a verifiable spike in the data coming from some factor other than human error. What that points to, we don't know yet. But Toyota sure is acting like they'll lose their case immediately if anyone examines the code running it. To the point of making up inane excuses and telling Congress and even the FBI "no". Only someone who is paranoid or a criminal would take such an action.
My educated guess is that a lot of what's driving this is that large Japanese corporations never, ever admit to anything. Their image at home is essentially inviolate and it is verboten to go against it at any level as an employee. Even in the press and society problems just aren't mentioned unless it's unavoidable.
The awful reality is that Japanese vehicles aren't that much more or less reliable than their counterparts so much as that Japanese companies never recall or admit to anything unless they have no choice. And to make it worse, the Japanese government is complicit in all of this by enabling them in their quest for a "perfect" image.
The sheer number of claims cannot be all lies and human error
Why not? Why can't it be the same as UFO reports or visions of God on the side of buildings?
As for why it's mostly about Toyotas, that seems to be a classic case of media-fueled auto-suggestion (no pun intended)....as in "Toyotas have UA....I have a Toyota....I have UA"
Last of all, the big boot theory makes perfect sense if one automakers pedals are designed differently than anothers---which they are. Pedals vary considerably in size and spacing from car to car, as well as how far they sit off the floor and how they are hinged.
I'd be convinced that toyota was as fault if someone showed me a video of a Toyota running away with both the driver's feet off the pedals and the cruise control light off.
In order to have a "failsafe" brake override system it cannot rely on any component "suspect" to be the causative factor, or contributor to the SUA incidents. For instance simply reprogramming the engine/transaxle ECU firmware might not work if the ECU firmware is at fault, or even partially so, for SUA.
So I would add the simplest, least complex, "add-on" circuit possible. Monitor the TPS, Throttle Position Sensor, and the brake light sensor and if the TPS doesn't quickly return to idle with brake application use a simple latching relay to open the fuel injector circuit.
I have come to notice recently that when moving my right foot from the gas pedal to the brakes I oftentimes move my right foot just to the right side of the brake pedal. If I then end up having to seriously apply braking then the right side of my foot depresses the accelerator along with the brake.
So I am "relearning", teaching, my right foot to move more agressively to the center of the brake pedal. Or I suppose I could saw an inch or so off the right side of the brake pedal.
Note that the above is really only possible with an automatic transmission.
I have assumed the CC itself might be at fault, the firmware code anyway.
My guess is that the CC code is getting locked into a "deadly embrace", executing ONLY a tight loop of instructions to the detriment of ALL other engine/transaxle ECU control functions, and thereby continuously executing ONLY the CC "accel" mode.
Possible, possible, I could buy that....BUT...I still can't buy the total simultaneous brake failure---I am especially skeptical about UA incidents from a stand-still position. On a freeway incident, I could see that if you were WOT, it would be hard to stop the car---but you should be able to slow it down enough (considerably) to give you time to turn off the ignition. On some cars, like a Prius, I don't see why you couldn't stop it from WOT--I believe Toyota actually demonstrated that.
If I might bring up the dreaded Occam's Razor again: as proponents of the electronic glitch theory are forced to come up with ever more complex explanations for simultaneous systems failures, this suggests (to me) a solution in search of facts, rather than the other way 'round.
First, no one, at least none of the 4 or 5 well documented instances, has even implied "total simultaneous brake failure". And none of "these" seem to have occurred, began, from a standstill position.
Has anyone, even Toyota, checked to see if the braking capability is enough to override the engine at WOT on a FWD vehicle..? The evidence to date, however meager, indicates NOT.
Toyota demostrated that a normally operating Prius could be brought to a quick stop "from" WOT but that was becuase the Prius goes automatically into regen mode (throttle "override..") when the brakes are used.
"..proponents of the electronic glitch theory are FORCED to come up with ever more COMPLEX explanations for simultaneous systems failures.."
Complex..??, Forced...?, not at all.
Having been in the industry, specifically the REAL-TIME process control industry, since '72 I can tell you that while software "deadly embrace" scenarios are VERY rare, they do happen, definitely.
How many of you out that have, at one time or another, have had to cycle the power off and back on in order to recover it to operational status with your PC or Mac..??
Why....no one knows or really cares.
But make use of commercial off-the-shelf PC hardware in a REAL-TIME environment...?
It reminds me of how early astronomers came up with complex and ingenious epicycle theories to explain Mars retrograde. All of which were of course barking up the wrong tree at the time, even though they did sort of explain it.
You can see how this goes:
"Well X caused the cruise control to go to WOT"
What about the brake failure
"Well, X + Y in conjunction with Z would explain that"
Well, what about the claim that putting the car in neutral didn't work?
"Well, X + Y + Z + A + B + C, if D=E, and if F, at certain times, defaulted to G---then that might explain that"
Yeah, okay, but....but.....it still doesn't explain the inability of the driver to stop the car with the brakes, though, nor the inability to put the car in neutral and kill the engine. CC can't work without ignition no matter how deadly its embrace on the throttle.
Nor does it really explain how, after the car is brought to a stop, no one, EVER, can replicate the problem.
So what you really need is the "CC-Deadly Embrace-Hydraulic Depressurizer-Ignition Cutoff-Gearshift-Locker--Disappearing Software Bug Theory"
I have no doubt that the inability to stop a car, FWD especially, already doing 60+ and WOT, using braking just simply doesn't work.
You can't put ANY modern day car into neutral unless the engine/transaxle controlling ECU is "paying attention" to the shifter gear ratio selector SWITCH group/set. Not locked in a deadly embrace, repetively executing only a narrow sequence of instructions, in this case most likely the CC "accel" instruction sequence.
The same applies to the PB start/stop function, and maybe even the ones with ignition keys/tumblers.
I have a senior, top level, programer working on such a problem at this very minute, and has been for about 4 weeks now. The problem "crops up" at customer sites around the world just barely often enough that it MUST be addressed.
We have been able to replicate the problem in our offices only once or twice in those past weeks and it seems that anything we change, instruction trace, say, causes the problem to abate.
Comments
2013 LX 570 2016 LS 460
Its in the flow,but without me ,you cannot go.
(For the archives, the forum was down over a day.... :sick: )
2013 LX 570 2016 LS 460
Huge isolated spikes like that are not going to be simply human error alone. Which is always what every maker blames it on when it does come to their attention. "Not our fault - must be human error". In every case, these complaints have occurred in vehicles with faulty drive-by-wire type components or malfunctioning cruise control systems. Both of which are computer-controlled in Toyotas.
Manuals never have this happen because they simply don't have these systems in place.(yet) As far as I can tell looking at the data myself, there have been no reported cases in the last decade of reported sudden acceleration for vehicles with a manual transmission in the NHTSA database. This also further reinforces it being statistically unlikely that it's simply human error and instead points to a design flaw that happens only in automatic equipped vehicles.
You can blame people all you want. I'll believe what the math says is likely instead.
A lot of desperate people trying to make a few bucks, and of course, they are zeroed in on Toyota.
2013 LX 570 2016 LS 460
Also, the fact that the UA reports do not cut across all makes of cars might suggest that the design and placement of the control pedals is a major factor in causing the human to err.
I would *like* to accept the "theory" of computer error, but applying occam's razor to the question suggests human error far more likely. Why? Because in many of these UA cases, the car runs away, is finally brought to a halt after a terrifying ride, and then the simultaneous failure of throttle control systems AND braking systems suddenly correct themselves and are never, ever replicated by third parties.
Now really, if I presented you with a similar case involving, say, your refrigerator and why the ice cream melted, might you not think that someone left the door open rather than a mysterious malfunction that only happens when one person is home?
From a recent court order story:
"Both sides spent the better part of Thursday's hearing arguing about a proposed discovery plan. Galvin raised concern about the protective order surrounding Toyota's "source code," the software underlying its electronics system. Toyota considers the source code "the most valuable asset it has," he said.
Robinson told the judge that he thought that both sides had reached an agreement to provide the source code as long as it was in a secured and neutral building that has been used by the Department of Defense and the Central Intelligence Agency."
First the NHTSA called in NASA. Now the parties are using old DOD and CIA digs.
Law.com
http://en.wikipedia.org/wiki/Occam%27s_razor
Specifically, it is useful as a heuristic in developing models but is not valid as a method for proving them. Or disproving them, either.
*quote*
"It has been suggested that Occam’s razor is a widely accepted example of extraevidential consideration, even though it is entirely a metaphysical assumption. There is little empirical evidence that the world is actually simple or that simple accounts are more likely than complex ones to be true."
-and-
"As a methodological principle, the demand for simplicity suggested by Occam’s razor cannot be generally sustained. Occam’s razor cannot help toward a rational decision between competing explanations of the same empirical facts."
****
"Of two equivalent theories or explanations, all other things being equal, the simpler one is to be preferred." - this is the commonly used example of Occam's Razor. You can't make the statement that "human error" is the simpler explanation in this case. In fact, the assumption that humans are automatically more fallible than machines is a classic case of poor reasoning that permeates our entire modern society.
The NHTSA database goes back for years. The general pattern of such claims always has two major factors.
1: It comes in large groups that affect one manufacturer at a time. In almost every case, it has (eventually) been proven to have been a defect that caused it. Most of the time it is a TPS or a cruise control system that has malfunctioned. Now, this alone doesn't mean anything concerning Toyota's claims, but yet again, we have a sudden spike in activity for one manufacturer and only on vehicles that have similar control systems in place.
2: While the propensity of humans to lie about something like some people are suggesting because it's in the news is a known factor, it's also been routinely shown that when such a series of claims is made, there also is a fair amount of bleed-over into other similar items or manufacturers. Because these cases have almost none of that, it points to something other than solely human error and greed. There is/was a legitimate spike in the number of cases reported, even if you discount all of the potentially questionable claims in their entirety. The few hundred that remain are alone enough to statistically prove that there still is an anomaly at work other than human error. Even if you factor in typical levels of human error in other similar cases, historically.
The judges and the Government realize this, of course, being well versed with statistics. Based upon Toyota's response to the normal by-the-book discovery process this week, they are hiding something. I know because I used to work in the e-discovery field. The only time anyone ever reacted like this was because they were guilty. In fact, the laws surrounding discovery are written with this in mind. Failure to undergo the process, destruction of evidence, or even failure to have backups of said evidence is considered grounds to simply lose the case in its entirey by default.
Toyota's claims are groundless. The legal discovery process is done every day and is extremely secure. They are playing games for a reason. The only reason we can't force them to comply is because they are technically a foreign corporation.(despite having plants in the U.S. and sub-corporation that runs them that's also based here)
"suggesting" that is the one writing the wiki notation.
When applied to diagnostics, it is simply common sense to eliminate the simplest explanations before become embroiled in complex ones.
Nothing "pundit" about it. That's a great way to fix cars, and it works pretty darn well---not as a sure-fire way to find an quick answer, but as a very reliable way not to waste time and money.
e.g., if a car doesn't start, you don't replace the computer before you check the battery cables. That's Occam's razor.
That's all that Occam's Razor implies to me. I can't speak for all the people on wikipedia who want to fog the idea over.
A far more common human error is when one assumes that the answer to a baffling mystery is always complicated. FINDING it might be complicated, but often the answer is not.
You know, if I was forced to choose sides between a person proposing the simplest answer or the one proposing conspiracies, I'd tend to give more credibility to the simplest answer. Conspiracy theories are, in a sense, a clever way of giving up rational inquiry without seeming to.
I have been working on a system for over 6 years now. We have been through the (what some may call) the engineering phase, and iterated the design for a deliverable build. The complete system was integrated last February (2009) and has been undergoing various testing procedures that culminated in the system being delivered in August of this year.
In February of 2010, one of the assemblies exhibited some anomalous behavior that would be consistent with a loss of power or a reset occurrence. Note that I said "consistent with", because to declare either as the root cause would have been premature.
The anomalous behavior did not show up again until early May, when we saw 2 occurrences over several days. Then, nothing more. So here we have 3 "failures" over several thousand hours of operation or so. After some head scratching, we removed the offending unit from the system and replaced it with a spare. The system completed the remainder of it's testing error free and, like I said, was shipped near the end of August.
The removed, suspect unit was placed on a test set and run continuously, 24 hours a day. It has exhibited two similar failures over that period of time. On the first our instrumentation did not "trigger" properly, so we did not catch the parameters we were looking for. The second occurrence we did capture the parameters of interest and are in the process of refining our probe location on the hardware to try and narrow the problem down to one of the several power supplies involved.
So, these kinds of infrequently occurring problems can be very, VERY difficult to 1) duplicate, and 2) track down. Tracking the problem down to it's root cause is usually not that tough, once you can come with a test arrangement that causes the problem to occur on a more regular basis than what is seen in normal use. And that is the problem with the UAE incidents - nothing that has been done has been able to, to my knowledge, increase the rate at which the UAEs occur.
If such testing showed nothing, there are actually two conclusions: 1) driver error or 2) a level of risk so minute and unlikely that no legal liabilities could be applied -- e.g., a law saying you can't throw anything overboard on a cruise ship at night in the middle of the ocean because you might hit a shipwreck survivor in a rowboat on the head.
SECOND QUESTION: Has anyone ever seen a report of a UA incident happening to the same vehicle TWICE?
(Well, you didn't say it had to be unexplained...).
I wonder why the UAs have suddenly stopped? Weird, huh?
Actually, you inferred as much by claiming that Occam's Razor was a valid reason to chose human error over mechanical or software error.
But as anyone who works with designing and troubleshooting software and/or complex machinery will tell you, it's only a very basic method that's used to narrow down choices. It cannot be used in cases like this, as you are presupposing that human error is the "simpler" explanation. Many young engineers and university students unfortunately latch onto this idea and don't realize that it also can be a trap. IMO, there should be a name for this scenario of being overly enamored with simplistic theories and ideas that are taught in college. Generation after generation learning the same nonsense that has to be un-learned all too often.
Occam's Razor has many limitations, and in this case, it can be shown via statistical analysis of the data that there are other likely factors at work besides (or in addition to) human error. Remember the "All other factors being equal" part of it? Well, if there is any sort of statistical anomaly AT ALL, then it cannot be used until that anomaly is dealt with and the data is normalized.
You're flat out wrong here. Simple as that. Until we also have a thorough analysis of the hardware and software that's involved, we cannot make any statement either way as to whether human error is a more valid explanation.
But the second issue is more telling, IMO.
Of course, Toyota is stonewalling, which is a good indicator that they know something that they don't want investigators to know about. This is always worrisome as it points to them trying to skew the data to their benefit. Perhaps this is typical corporate posturing by them, but given the number of similar actions by other companies in the past leading it to being a legitimate case of hiding something, one has to wonder what's really going on. As I pointed out, the legal presumption is that failure to comply with a request for discovery is seen as automatically hiding something(as well as breaking the law), because the only times where it happens has been when they are actively trying to obstruct justice. It's also why the judges refuse to let it die and Congress is getting more and more angry about it(almost half of Congress is made up of lawyers who take a dim view of refusing to comply with the law)
Simply put, if they had nothing to hide, they would show that they have nothing to hide. I think the FBI and CIA can manage to keep their "data" secure.
So let's unpack what is known so far:
On the one hand, we have some UAs, all with automatic transmission cars, and many of which were later examined with the UA cause neither detected nor ever reproduced in, I believe, 100% of the cases.
Theory 1: driver error
Theory 2: defective floor mats
Theory 3: an unexplained, undetectable undefined computer glitch.
Why should I believe #3 "more than" 2 or 1? Shouldn't they have equal weight at least?
Of the three theories, driver error actually explains what was observed better than the other two.
Is it "true"? We don't know (yet).
If I'm reading your ideas correctly, your defense of theory #3 is based on a suspicion that Toyota is withholding something, and that "something" holds the answer.
That feels very unsatisfactory to me.
Or are you saying that if an answer is never found, that doesn't mean theory #3 isn't right?
However, if an answer is never found, what good is theory #3?
2- Human error may be a factor. It may not.
3 - Mechanical failure cannot be ruled upon at all, since we don't have any data yet. Basically we're trying to make a decision between two ideas with only one set of data.
Simply put, the test is only halfway done and we can't test the hypothesis yet. Talking as if it's already been decided is premature. Whether or not other evidence has been found or not is irrelevant. We have a competing theory that is still awaiting data to come in.
As for the legal process, I worked for five years in this exact aspect of the law. Discovery is a by-the-numbers process that's well understood by any company today. If you are failing to comply with it, you are assumed to be doing it willingly and the chance that it's a mistake is a statistical impossibility if your company has even a single lawyer on staff. Short of your data center burning down and your backups also burning up at the same time, it never happens unless you are hiding evidence that you know will incriminate you.
It's such a strong statistical improbability that failure to comply (if you are a U.S. corporation)pretty much automatically looses you the case. You might not like it, but it's how the law works in the U.S. as well as most of Europe. You either comply or your are judged to be willingly obstructing justice. Telling the FBI, CIA, the Justice Department *and* Congress "no" would be insanity for any U.S. corporation to contemplate.
But Toyota is not a U.S. corporation and is using that to play games with established legal procedures. If it was me, I'd threaten to embargo their cars and shut down their factories in the U.S. until they comply. But then again, I doubt if our current Congress has the ability to play hard ball with them like they should.
But, on the other hand, the problem with real world faults(?) is that the conditions under which they occur are uncontrolled. What was the weather? Was the vehicle going past a power station? Was the driver on the cell phone? Was he moving the windows up and down, and adjusting the power seat?
Along with this, people in general are notoriously unreliable when, after the fact, recounting exactly what transpired. How many stories have we all heard about eyewitness accounts being shown to be inconsistent between multiple eyewitnesses, or the account being discredited when an unbiased source such as a surveillance camera show that something different occurred?
"You say you were reaching over to adjust the volume of the radio when the car suddenly accelerated. Did the car start accelerating before of after you turned the volume control?" "Well, I'm not sure, I think it was after, but it could have been before".
I am in agreement with your comment about the level of risk being so small. Even if the problem does turn out to be a defect (using the term loosely), with billions of vehicle miles and billions of hours of operation to only have a relatively few number of UAEs, I would proffer that such does not rise to the level of negligence. People expect technology and things they use, whether they be autos or passenger planes, to be 100% safe 100% of the time. Such expectations are unrealistic, and need to be adjusted.
Like the saying goes, sometime it does happen!
That should make Toyota feel safe about their code in light of Wikileaks. Then what auto maker would want to steal defective code??? :shades:
Brake override systems should help with the safety issue. Looks like most all the automakers are installing them.
To rule otherwise would essentially paralyze the very basis of production of goods. This is *especially* true with those products that humans interact upon.
If you are afraid of toyota products, then don't buy one.
You might just as well add "what they ate for breakfast" as a possible cause also, since there is zero evidence for that as well...and would be extremely hard to prove it is not the cause.
2013 LX 570 2016 LS 460
I think the real problem here is that you and many other people don't really properly conceptualize that that it's not "mechanical" and "software", but more of a robotic system that functions together. So far, in essence, we've only been able to poke at the exterior and look at the shiny wires and actuators. Our job of even ruling out the "mechanical" aspect of the system is actually only halfway completed.
It seems perfectly sane to want to complete this part of the examination, doesn't it?
*BUT*
Non-human faults have always been a component in cases like this historically. With every previous set of major incidents to date being related to a defect, it's something that we cannot overlook and just say "there's zero evidence so far". Especially when Toyota is refusing to give access to investigators.
IMO, that's a giant red flag. Toyota could comply but simply is making up excuses at this point. If this was a criminal case, the DA would add a charge of obstructing justice(at the least - they love to tack on extra charges). It's exactly like the following scenario:
"So where is the gun that you have registered in your name?"
"I don't know and I'm not telling you."
(cue some very odd looks from the prosecutor and judge)
or:
"So where is the computer code that controls the system?"
"We're not going to let you have it."
???
On the contrary, the only other incident of this kind that I remember was the alleged Audi sua fiasco, and I believe this was decided as driver error.
2013 LX 570 2016 LS 460
If nobody can make a Toyota do a UA, other than the owner of the vehicle, well then, I'd say that Toyota doesn't owe anybody anything.
"Mr. Plekto, your gun is defective. I shot myself with it"
"Okay, Mr shiftright, show me how it is defective"
"Well I can't make it happen again, Mr. Plekto but we need to know all your metallurgical formulas so that we can melt it down and examine the basic structure to see if certain thermodynamic and climactic forces might have caused a slight distortion in the firing pin under laboratory conditions, thereby possibly causing a delay in the strike against the bullet's primer, and also possibly a slight deflection of the pin's path if the gun is held in a 45 degree position after removal from a warm holster".
Only in/by (minority??) public opinion.
Public opinion formed/formulated/molded by you know who.
On the other hand, there is considerable US debt that is held hostage..
Toyota says "When the vehicle throttle is opened beyond the idle position, at speeds greater than five miles per hour and then the brakes are firmly applied for longer than one-half second, the override feature will reduce engine output to the idle position, allowing greater braking performance."
link
Then the driver would claim that the override did not work and here we go again !!
2013 LX 570 2016 LS 460
(no kidding--I almost smashed into the garage door one evening. I would have sworn my foot was on the brake---and it was---and on the gas, too).
2013 LX 570 2016 LS 460
So only toyota drivers once in a while wear big boots and stepped on the wrong pedal ? So Hyundai, Nissan, Honda drivers don't wear big boots ? So only toyota old drivers screw up ? So there are no old drivers driving Hyundais ? Nissans ? Give me a break !
Look at the statistics. Hyundai, Honda, Nissan customers are no more loyal or fanatical to their cars than toyota owners. If Nissan, Hyundai etc cars have the same kind of problems that toyota have today, trust me, they WILL sue !
Anybody want to adress this fact ?
Owners, if they are doing this just to create a lot of press and false claims, would also be attacking other manufactures. Because it does happen from time to time with other makers as well.(usually a defective TPS unit) We should see a similar number of false claims from idiots who can't figure out what pedal is which(because they almost always won't admit making such a mistake).
But they are not.
The sheer number of claims cannot be all lies and human error. There is a verifiable spike in the data coming from some factor other than human error. What that points to, we don't know yet. But Toyota sure is acting like they'll lose their case immediately if anyone examines the code running it. To the point of making up inane excuses and telling Congress and even the FBI "no". Only someone who is paranoid or a criminal would take such an action.
My educated guess is that a lot of what's driving this is that large Japanese corporations never, ever admit to anything. Their image at home is essentially inviolate and it is verboten to go against it at any level as an employee. Even in the press and society problems just aren't mentioned unless it's unavoidable.
The awful reality is that Japanese vehicles aren't that much more or less reliable than their counterparts so much as that Japanese companies never recall or admit to anything unless they have no choice. And to make it worse, the Japanese government is complicit in all of this by enabling them in their quest for a "perfect" image.
Why not? Why can't it be the same as UFO reports or visions of God on the side of buildings?
As for why it's mostly about Toyotas, that seems to be a classic case of media-fueled auto-suggestion (no pun intended)....as in "Toyotas have UA....I have a Toyota....I have UA"
Last of all, the big boot theory makes perfect sense if one automakers pedals are designed differently than anothers---which they are. Pedals vary considerably in size and spacing from car to car, as well as how far they sit off the floor and how they are hinged.
I'd be convinced that toyota was as fault if someone showed me a video of a Toyota running away with both the driver's feet off the pedals and the cruise control light off.
So I would add the simplest, least complex, "add-on" circuit possible. Monitor the TPS, Throttle Position Sensor, and the brake light sensor and if the TPS doesn't quickly return to idle with brake application use a simple latching relay to open the fuel injector circuit.
I have come to notice recently that when moving my right foot from the gas pedal to the brakes I oftentimes move my right foot just to the right side of the brake pedal. If I then end up having to seriously apply braking then the right side of my foot depresses the accelerator along with the brake.
So I am "relearning", teaching, my right foot to move more agressively to the center of the brake pedal. Or I suppose I could saw an inch or so off the right side of the brake pedal.
Note that the above is really only possible with an automatic transmission.
My guess is that the CC code is getting locked into a "deadly embrace", executing ONLY a tight loop of instructions to the detriment of ALL other engine/transaxle ECU control functions, and thereby continuously executing ONLY the CC "accel" mode.
If I might bring up the dreaded Occam's Razor again: as proponents of the electronic glitch theory are forced to come up with ever more complex explanations for simultaneous systems failures, this suggests (to me) a solution in search of facts, rather than the other way 'round.
Has anyone, even Toyota, checked to see if the braking capability is enough to override the engine at WOT on a FWD vehicle..? The evidence to date, however meager, indicates NOT.
Toyota demostrated that a normally operating Prius could be brought to a quick stop "from" WOT but that was becuase the Prius goes automatically into regen mode (throttle "override..") when the brakes are used.
Complex..??, Forced...?, not at all.
Having been in the industry, specifically the REAL-TIME process control industry, since '72 I can tell you that while software "deadly embrace" scenarios are VERY rare, they do happen, definitely.
How many of you out that have, at one time or another, have had to cycle the power off and back on in order to recover it to operational status with your PC or Mac..??
Why....no one knows or really cares.
But make use of commercial off-the-shelf PC hardware in a REAL-TIME environment...?
You either "care" or you do not survive.
You can see how this goes:
"Well X caused the cruise control to go to WOT"
What about the brake failure
"Well, X + Y in conjunction with Z would explain that"
Well, what about the claim that putting the car in neutral didn't work?
"Well, X + Y + Z + A + B + C, if D=E, and if F, at certain times, defaulted to G---then that might explain that"
And on and on it goes.......
In the meantime we work with what we have.
For the moment, I expect all would agree, a CC "accel" deadly embrace firmware instruction loop theory covers all information currently available.
Nor does it really explain how, after the car is brought to a stop, no one, EVER, can replicate the problem.
So what you really need is the "CC-Deadly Embrace-Hydraulic Depressurizer-Ignition Cutoff-Gearshift-Locker--Disappearing Software Bug Theory"
More epicycles....
You can't put ANY modern day car into neutral unless the engine/transaxle controlling ECU is "paying attention" to the shifter gear ratio selector SWITCH group/set. Not locked in a deadly embrace, repetively executing only a narrow sequence of instructions, in this case most likely the CC "accel" instruction sequence.
The same applies to the PB start/stop function, and maybe even the ones with ignition keys/tumblers.
I have a senior, top level, programer working on such a problem at this very minute, and has been for about 4 weeks now. The problem "crops up" at customer sites around the world just barely often enough that it MUST be addressed.
We have been able to replicate the problem in our offices only once or twice in those past weeks and it seems that anything we change, instruction trace, say, causes the problem to abate.
You're welcome to contribute.