http://www.audizine.com/forum/showth...=1#post8382275
" Just something to think about, as "adaptive" control system software could be affected in unexpected ways depending upon operating conditions and usage patterns."
Good job Tomasz and $Doc. It seemed intuitive that an electronically controlled system must have a parameter that would allow or prohibit intervention in the system's operation on the part of the computers / sensors, and it looks like you guys finally dug it out. What's puzzling to me is why some cars' sensor systems would learn to apply that corrective behavior with such alarming frequency. What about steering wheel position, wheel speed, slip sensor inputs, frammis tensioning, etc. etc. would cause some of them to not only learn differently, but to then over-use (abuse?) the "correction" feature?
South FL roads (where I live and drive 98% of the time) are mostly all flat, smooth, and largely devoid of grooves or ruts, except for maybe the far right lanes of major highways that see constant abuse from heavy vehicles - where I DON'T drive 99.9% of the time. Perhaps the adaptive steering feature is coded a bit too liberally, in that if it observes or samples > X instances of what it thinks is steering behavior that requires correction over a certain period of time - for example if a car is driven extensively on grooved or rutted highways - it thinks "hey, I see wandering or inattention, I'm needed here!" [yank the wheel], when the reality is it's just a crappy road. Oversimplification of what I'm sure is a complicated algorithm, but if your fix holds, it's gotta be the case, and could probably be coded to factor in such circumstances with future revisions.
Good luck, and I think I'll look up someone local w/a VAG COM and get that bit unset just for my own peace of mind.
Bookmarks