Sadly, it's one of those things you won't really know until you need it, unless someone finds out more details about the inputs the IMU uses.
Totally agree. If someone does find more detailed specifications from Bosch regarding how specifically these systems work, please let us know. I have searched several times, but haven't found anything beyond the usual marketing materials.
If I were a wagering man, I'd bet that Kawasaki implemented the system with a significant safety margin relative to the predicted failure limits of the system. This is why first-rate riders describe the TC and ABS interventions as sometimes "too conservative" or "premature" when riding hard. Let's say the Kawasaki-engineered margin of error is maybe 10 - 20%. And then they set the dashboard fault threshold at something well within that margin -- maybe 5% as your excellent list of known data points suggests.
Given all that speculation, I would further speculate that if your changes don't cause a dashboard error, then you are safely within the ICU's operational margin of error (as OCL implied). The 190/55 tire is almost certainly an example of a "within the margin of error" modification. But I also agree with RC that the safer, more considered way to adjust the geometry on these ICU-equipped bikes is via the shock or linkage.
I really wish I knew more about how these systems work. The older ABS systems were cool in their own way, but the operational parameters were pretty easy to comprehend. There were only really two inputs:
1) has this wheel's brake been activated?
2) is the wheel in question rotating?
If the brake was activated and the wheel wasn't rotating, then the tire was sliding and the brake intervention would begin modulating brake pressure.
These new ICU systems appear to be making
predictions about when traction is likely to be lost. Because if the system were waiting for actual slippage, like the old ABS, then it wouldn't need to know tire circumference, would it? But, at the same time, it can't totally be prediction-oriented because it's impossible for the system to know, for instance, whether or not the road is wet. So it must be using
measured slippage somehow to set the real-time traction threshold.
Like I said, I really wish I knew more.