Another unexplained acceleration

[Mr2Mike]

98elise said:

Its very simple to design a fail safe. For example the pedal operates two potentiometers in opposite directions. This will give two pedal different position signals which can be checked against each other, yet don't behave the same in a failure

If you take a pedal sensor apart you'll see that is exactly how they are designed, there are multiple resistance tracks rather than just one. Also designed so that a break in a track or failure of the wiping contact results in a safe failure mode. The angle sensor at the throttle end uses a similar scheme.

[saaby93]

Roger Irrelevant said:

The Wookie said:

I think good old Occam's Razor applies here

Yep I'll go with that.

Oh but wait, the lady said that if she'd pressed the throttle she'd have admitted it. That throws a whole different complexion on it. So now it's not just:

'mystery throttle error for which there is no explanation that has failed to manifest itself in millions of other cars with the same system vs old lady gets driving wrong',

it's now:

'mystery throttle error for which there is no explanation that has failed to manifest itself in millions of other cars with the same system vs old lady gets driving wrong and is mistaken about it'.

...Occam would have had a tough time with that one.

My money's on Occam siding with the lady
and the car being the complex device with many failure modes
But it could be the other way around

Trouble is you cant convict on best guess - or can you

[framerateuk]

TheInternet said:

Safety critical custard. I'd also wager that there are fewer DBW incidents than cable throttle incidents.

I'd agree. I've had the throttle cable snap on my Caterham. In a light car like that it's almost as bad as slamming on the brake.

[98elise]

durbster said:

98elise said:

Its very simple to design a fail safe. For example the pedal operates two potentiometers in opposite directions. This will give two pedal different position signals which can be checked against each other, yet don't behave the same in a failure. It is sudden unintended acceleration, but the cause is very very likely to be the driver.

Why add complexity?

It doesn't matter if the accelerator is flat to the floor - if you press the brake the car will stop. The problem is how the human mind can trick you into thinking your body is doing something it's not.

The only reliable solution is autonomous cars.

Its not complicated at all to add a fail safe like that, and it would prevent it happening in the first place. I'd be very surprised if drive by wire relied on a single sensor for throttle position. I can't see a car company admitting that full throttle can be induced by a single potentiometer short.

[0000]

GroundEffect said:

Also I'm not going to take the word of a 82yo, sorry.

What about a 79 year old?

[Durzel]

I would presume that car software is written in such a way that if it fails then it's like a dead man's switch, in other words the software has to be operational (not crashed) for the car to (continue to) move at all.

If drive-by-wire was at risk of sticking open the throttle, or could actually crash in the traditional sense, we'd be hearing about it all of the time.

[saaby93]

Durzel said:

I would presume that car software is written in such a way that if it fails then it's like a dead man's switch, in other words the software has to be operational (not crashed) for the car to (continue to) move at all.

If drive-by-wire was at risk of sticking open the throttle, or could actually crash in the traditional sense, we'd be hearing about it all of the time.

Why? It doesnt happen very often

[LeoSayer]

mac96 said:

I am in the 'more likely to have pushed wrong pedal' camp.

But how can a foot slipping off the pedal, or pushing the wrong one by accident, deserve a prison sentence- it is ludicrous. and yet she seems to be looking at a minimum 2 year tariff ?(happy to be corrected on that).

She was driving a car that killed someone, blamed the car and the police don't believe her. I doubt she'd be facing the same charges if she had admitted it was her own fault at the scene or couldn't say.

I imagine some kind of compromise will be reached where she is banned/prevented from driving and doesn't end up in prison.

[98elise]

Mr2Mike said:

98elise said:

Its very simple to design a fail safe. For example the pedal operates two potentiometers in opposite directions. This will give two pedal different position signals which can be checked against each other, yet don't behave the same in a failure

If you take a pedal sensor apart you'll see that is exactly how they are designed, there are multiple resistance tracks rather than just one. Also designed so that a break in a track or failure of the wiping contact results in a safe failure mode. The angle sensor at the throttle end uses a similar scheme.

Thanks, I didn't know that, but I have some experience in control systems that employ multiple sensors. They are designed that way so that a if both fail the same way, you don't end up with two corresponding (but wrong) signals sent. As an example if you flooded the sensor casing both would probably give the same output, but using an opposing output this would be seen as an error.

[phil4]

So many flaws in her story:

"I went into drive and I put my foot very gently on the accelerator to cross the road and then the car just took off. It surged forward." So perhaps she wasn't as gentle as she thought? Perhaps she thought the engine had "auto stopped", she gave it a prod to start the engine, but it was already started.

"I can remember taking my foot off the accelerator quickly and I'm sure it went on to the brake and the brake didn't appear to be working.". Right... gave it a really good shove did she? I'm not saying I try this often, but I'm pretty sure a full on brake press, would do a fair bit to slow or stop the car even if temporarily. Last time I did it by mistake it was a bit like hitting a brick wall.

[Markbarry1977]

My gut instinct is that the dear old lady has probably by the law of averages got the wrong pedal but firmly believes she hasn't.

Having said that, how many times have I had to restart a pic or iPhone.

In my previous job I wrote software for a specific air traffic control system. The software had been in use worldwide for some 20 years. One day in a simulator running regression testing on a new modification there was a major safety error in the software found, nothing to do with the new change but in the original baseline software. Admittedly the scenario and user/operator switch and key actions requires to trigger the error were very specific and the chances of it happening millions to one but a student controller during a free play session happened to trigger it. He argued that it had happened we all said no you must have imagined it. In the end we replayed a recording of the test session. He was right we were wrong and all resources were dropped on everything to fix this problem.

My long winded point is that even when someone says it's impossible, there's always that one in a million sequence of things that can cause unexpected errors.

[mac96]

LeoSayer said:

mac96 said:

I am in the 'more likely to have pushed wrong pedal' camp.

But how can a foot slipping off the pedal, or pushing the wrong one by accident, deserve a prison sentence- it is ludicrous. and yet she seems to be looking at a minimum 2 year tariff ?(happy to be corrected on that).

She was driving a car that killed someone, blamed the car and the police don't believe her. I doubt she'd be facing the same charges if she had admitted it was her own fault at the scene or couldn't say.

I imagine some kind of compromise will be reached where she is banned/prevented from driving and doesn't end up in prison.

I hope so. Trouble is, with the threat of prison, what is she supposed to say? In any case, as someone said above, even if it was her error, she probably genuinely believed it wasn't. She is therefore in the unpleasant position of possibly being further penalised for pleading Not Guilty, when she genuinely believes she was not guilty.

Even assuming it was her error and not the car, I find it hard to see how this can be deserving of penalties on the scale of the charged offence. Even if the victim had been a friend or relative of mine I cannot imagine wanting to see an old lady sent to prison following an accident like this.

I would of course feel differently if she had a history of near misses, but we don't know that.

[SonicShadow]

phil4 said:

So many flaws in her story:

"I went into drive and I put my foot very gently on the accelerator to cross the road and then the car just took off. It surged forward." So perhaps she wasn't as gentle as she thought? Perhaps she thought the engine had "auto stopped", she gave it a prod to start the engine, but it was already started.

"I can remember taking my foot off the accelerator quickly and I'm sure it went on to the brake and the brake didn't appear to be working.". Right... gave it a really good shove did she? I'm not saying I try this often, but I'm pretty sure a full on brake press, would do a fair bit to slow or stop the car even if temporarily. Last time I did it by mistake it was a bit like hitting a brick wall.

Yup, blatantly obvious what happened. Throttle could be pinned wide open but if you fully on the brakes they will win every time.

[J4CKO]

Markbarry1977 said:

My gut instinct is that the dear old lady has probably by the law of averages got the wrong pedal but firmly believes she hasn't.

Having said that, how many times have I had to restart a pic or iPhone.

I get that it is all electronics but a DBW throttle has just one job to do, doesnt run an operating system, it cant really be user altered (apart from drive modes) as it has no connection to the outside world, PC's are millions of lines of codes to suit a multitude of different hardware and configurations, you can connect anything to it and to be fair, nowadays they work pretty well, not had to reboot my new one, built last year at at all.

The DBW throttle will be less code than a Nokia 3310 even, its a very simple system despite being electronic, not saying ti cant go wrong, EMF or the sensor breaks but like someone mentioned it is what it does in that scenario, suspect they default to a failsafe but if it thinks its got full throttle it will apply full throttle.

Grannies are more complex again than PC's and cant be patched

[LeoSayer]

mac96 said:

In any case, as someone said above, even if it was her error, she probably genuinely believed it wasn't. She is therefore in the unpleasant position of possibly being further penalised for pleading Not Guilty, when she genuinely believes she was not guilty.

I'm sure her lawyers would have explained the risks to her long before it went to trial.

Hopefully her lawyers have some backup for their claim about a malfunction due to undercharged battery.

I'm half her age but if I was in her circumstances I doubt I'd be able to say for certain what pedal I hit.

[Roger Irrelevant]

saaby93 said:

My money's on Occam siding with the lady
and the car being the complex device with many failure modes
But it could be the other way around

Trouble is you cant convict on best guess - or can you

No of course you can't convict on a best guess, but then neither can you get off a dangerous driving charge by simply saying that cars are complex and have 'many failure modes'.

The people who make up the jury will know that, in the vast, vast majority of cases, cars don't just do their own thing. In fact the car in this case had been driven without this supposed fault manifesting itself for at least seven years previously. So the prosecution will have put a straightforward case based on the elderly defendant making a mistake - the sort of thing that happens all the time. Now jury trials can be a bit of a lottery, especially when the defendant is of a type that will naturally garner sympathy (e.g. a little old lady), but in order to create the reasonable doubt that is required to get the defendant acquitted of a criminal charge I would have thought the defence would have to do more than trot out the 'many failure modes' line. Have they any evidence at all as to what the specific failure was? Why don't we see Qashqais (or other similarly complex machinery) careering out of control for unknown reasons on a regular basis?

Maybe she's got OJ Simpson's legal team on the case and will be found not guilty, but I don't think she'd stand a chance defending a civil claim.

[mac96]

LeoSayer said:

mac96 said:

In any case, as someone said above, even if it was her error, she probably genuinely believed it wasn't. She is therefore in the unpleasant position of possibly being further penalised for pleading Not Guilty, when she genuinely believes she was not guilty.

I'm sure her lawyers would have explained the risks to her long before it went to trial.

Hopefully her lawyers have some backup for their claim about a malfunction due to undercharged battery.

I'm half her age but if I was in her circumstances I doubt I'd be able to say for certain what pedal I hit.

I am with you there. What I really don't understand is the charge- others seem more appropriate.
On the face of it, she is guilty of 'Careless or inconsiderate driving arising from momentary inattention with no aggravating features' which if causing death leads to a community order I believe. Which seems much more sensible.

[MajorMantra]

WokingWedger said:

Don't these cases of 'pushing the wrong pedal' (which it almost certainly is) only happen in Autos ?

Sounds obvious I know, but think about it.

You are creeping forward (or back) without touching any pedals, but 'covering' one of them, possibly the throttle because it often doesn't quite creep faster enough.

An emergency happens and you get it wrong (you had been 'covering' the throttle)but stamp on it thinking you were covering the brake.

Doesn't happen with manual cars as you have to know where your feet are to make it move.

As there are a lot more autos out there in the hands of older people (and young)we are bound to see more of it.

I've been thinking this too. My impression is that there are a lot of this type of accident (cars mounting pavements, smashing through shopfronts etc.) in the US and I don't think it's a coincidence that this is the case in a country where the vast majority drive autos. It's MUCH harder to have a 'sudden unplanned acceleration' in a car with a manual clutch.

[saaby93]

Roger Irrelevant said:

Why don't we see Qashqais (or other similarly complex machinery) careering out of control for unknown reasons on a regular basis?

This doesnt have to be a regular basis failure mode (they'd have fixed that) - isn't more likely a very rare occurrence failure mode?

[Wills2]

The absence of any explanation as to how the car could leap forward with no driver input means it's going to be difficult to create a reasonable doubt in the mind of the jurors as to what actually happened.