6 ms·

Is the software actually designed to "dupe" regulators? Might there be many legitimate reasons why the emissions vs performance tradeoff might be tailored to d

Is the software actually designed to "dupe" regulators? Might there be many legitimate reasons why the emissions vs performance tradeoff might be tailored to driving conditions? Of course when the car is idling it would make sense to minimize emissions as much as possible, just as an idle CPU can utilize a lowered clock speed.

aaronbrethorst 11y ago
Do you think that VW's CEO would have resigned if this was not the case?
- jfoutz 11y ago
  I think the VW software actually checks to see if it was in test mode, and altered emissions. Varying emissions based on load is perfectly reasonable.
  - grandalf 11y ago
    > software actually checks to see if it was in test mode, and altered emissions I saw this reported, but also saw that it actually just looks for acceleration and steering conditions often found in testing scenarios. There was no mention of whether these conditions are also found in routine use of the vehicle. Not sure which is true.
- dustinchilson 11y ago
  I'm thinking the 25% drop in stock prices might've had more to do with it.
  - grandalf 11y ago
    Agreed. Once the accusations were made public and the stock price dipped, the optimal solution may involve the CEO stepping down... and likely goes a long way to appease regulators and deflect further scrutiny.
  - aaronbrethorst 11y ago
    Not according to him: http://www.volkswagenag.com/content/vwcorp/info_center/en/news/2015/09/Statement.html http://www.volkswagenag.com/content/vwcorp/info_center/en/ne...
downandout 11y ago
Apparently in the VW case, this was their specific intent. Also, limits are limits. You can't have the software allow violations of emissions limits because the driver wants to go faster at a particular point. That isn't to say that the CEO is at fault, nor would it be expected that the CEO of an enormous company with thousands of engineers would know about or direct specific engineering decisions (and in fact he denies that he knew). More likely than not, he issued unrealistic directives to the engineering team, who met the demands in the only way they could at the time. That doesn't make it right by any stretch, but that appears to be what happened.
- grandalf 11y ago
  > You can't have the software allow violations of emissions limits because the driver wants to go faster at a particular point. It depends on how the emissions are measured. They could be measured per RPM, per mile, per gallon, per pound transported, or by any of those over time, or some basket of utilization patterns modeled after real-world use. If you slam the accelerator to the floor, the vehicle produces more emissions than if you'd continued along at a steady pace. Emissions control is a combination of active and passive systems, along with careful moderation of inputs to the engine. So it seems plausible that a complex, computer-controlled vehicle would optimize emissions vs performance by modifying many of the available variables in response to driving conditions.
  - downandout 11y ago
    >It depends on how the emissions are measured. That is certainly a good point. However in this case, the specifics of what they were doing obviously crossed the line. Normally I roll my eyes when I hear about EPA issues, given that agency's penchant for making mountains out of molehills and ravenous appetite for horrendous fines. But in this case, many rational, scientifically credible people all seem to agree that they were intentionally and egregiously violating the law.
    - grandalf 11y ago
      > scientifically credible people all seem to agree that they were intentionally doing this. That does seem to be the case. I'm curious when the dust settles if the spirit or the letter (or neither, both) of the EPA requirement was violated.
rconti 11y ago
From this article and every other article I've read, yes. It specifically detects when it is being tested versus normal road driving.
- grandalf 11y ago
  I haven't seen specific proof of this, other than describing the computer as detecting a lack of steering, etc. Conditions that could also describe the car idling or being driven in real world use cases. It would seem that this point is the crux of the accusation, so I find it odd that more detail hasn't been offered.
  - wiredfool 11y ago
    They have admitted that the software detects tests.
    - grandalf 11y ago
      Was the software designed only to detect tests? Or is detecting tests (and other cases where limiting performance has no adverse consequences) a side-effect of the system's intended behavior. For instance, suppose the vehicle was programmed to shut off its engine when stopped at a stoplight. Suppose the test scenario measures emissions while stopped at a stoplight...
      - wiredfool 11y ago
        All of the reports indicate that the ECUs were programmed to detect tests and enable the changes that reduce NOx emissions. The intended behaviour was to have one performance profile for emissions tests, and one for power/efficiency elsewhere. The engine shutoff feature is irrelevant, as it happens in the wild when not running emissions tests.

aaronbrethorst 11y ago

Do you think that VW's CEO would have resigned if this was not the case?

jfoutz 11y ago
I think the VW software actually checks to see if it was in test mode, and altered emissions. Varying emissions based on load is perfectly reasonable.
- grandalf 11y ago
  > software actually checks to see if it was in test mode, and altered emissions I saw this reported, but also saw that it actually just looks for acceleration and steering conditions often found in testing scenarios. There was no mention of whether these conditions are also found in routine use of the vehicle. Not sure which is true.
dustinchilson 11y ago
I'm thinking the 25% drop in stock prices might've had more to do with it.
- grandalf 11y ago
  Agreed. Once the accusations were made public and the stock price dipped, the optimal solution may involve the CEO stepping down... and likely goes a long way to appease regulators and deflect further scrutiny.
- aaronbrethorst 11y ago
  Not according to him: http://www.volkswagenag.com/content/vwcorp/info_center/en/news/2015/09/Statement.html http://www.volkswagenag.com/content/vwcorp/info_center/en/ne...

downandout 11y ago

Apparently in the VW case, this was their specific intent. Also, limits are limits. You can't have the software allow violations of emissions limits because the driver wants to go faster at a particular point. That isn't to say that the CEO is at fault, nor would it be expected that the CEO of an enormous company with thousands of engineers would know about or direct specific engineering decisions (and in fact he denies that he knew). More likely than not, he issued unrealistic directives to the engineering team, who met the demands in the only way they could at the time. That doesn't make it right by any stretch, but that appears to be what happened.

grandalf 11y ago
> You can't have the software allow violations of emissions limits because the driver wants to go faster at a particular point. It depends on how the emissions are measured. They could be measured per RPM, per mile, per gallon, per pound transported, or by any of those over time, or some basket of utilization patterns modeled after real-world use. If you slam the accelerator to the floor, the vehicle produces more emissions than if you'd continued along at a steady pace. Emissions control is a combination of active and passive systems, along with careful moderation of inputs to the engine. So it seems plausible that a complex, computer-controlled vehicle would optimize emissions vs performance by modifying many of the available variables in response to driving conditions.
- downandout 11y ago
  >It depends on how the emissions are measured. That is certainly a good point. However in this case, the specifics of what they were doing obviously crossed the line. Normally I roll my eyes when I hear about EPA issues, given that agency's penchant for making mountains out of molehills and ravenous appetite for horrendous fines. But in this case, many rational, scientifically credible people all seem to agree that they were intentionally and egregiously violating the law.
  - grandalf 11y ago
    > scientifically credible people all seem to agree that they were intentionally doing this. That does seem to be the case. I'm curious when the dust settles if the spirit or the letter (or neither, both) of the EPA requirement was violated.

rconti 11y ago

From this article and every other article I've read, yes. It specifically detects when it is being tested versus normal road driving.

grandalf 11y ago
I haven't seen specific proof of this, other than describing the computer as detecting a lack of steering, etc. Conditions that could also describe the car idling or being driven in real world use cases. It would seem that this point is the crux of the accusation, so I find it odd that more detail hasn't been offered.
- wiredfool 11y ago
  They have admitted that the software detects tests.
  - grandalf 11y ago
    Was the software designed only to detect tests? Or is detecting tests (and other cases where limiting performance has no adverse consequences) a side-effect of the system's intended behavior. For instance, suppose the vehicle was programmed to shut off its engine when stopped at a stoplight. Suppose the test scenario measures emissions while stopped at a stoplight...
    - wiredfool 11y ago
      All of the reports indicate that the ECUs were programmed to detect tests and enable the changes that reduce NOx emissions. The intended behaviour was to have one performance profile for emissions tests, and one for power/efficiency elsewhere. The engine shutoff feature is irrelevant, as it happens in the wild when not running emissions tests.