I've driven two cars with rain-sensing wiper systems. Two is a poor sample rate, but since I drive/drove them all the time, and one of them in a place where rain is not seasonal, my sample rate of their performance is very high. And I have zombie Richard Feynman on my side.
Our Toyota Corolla Verso has four levels of "rain" sensitivity, as well as two, fixed speed settings. To me, this implies a speed spectrum like this:
0 - 0.2 - 0.4 - 0.6 - 0.8 - 1 - 2
This was certainly the case in all the cars I've driven that had variable speed wipers that were directly controlled by the driver.
However the actual spectrum I get on the Verso is:
0 - WTF! - 1 - 2
Changing the variable setting doesn't seem to affect the sensitivity of the system. What's more, the lowest sensitivity setting will cycle the wipers as fast as the highest setting. What??!!
This is not rocket science, people. Regardless of how much the "rain sensor" is a technical black box, the data it outputs should be some normalized value which reflects how "wet" the windshield is, which is compared against a trigger value that's associated with what the driver has set.
And this is probably one of the easiest features in a car to test. You don't need a track, you don't even need most of the car. A correctly configured windshield and wipers and a couple of spray bottles will iron out 80% of the bugs in the system. The other 20% will require misters and a shower head.
Get on this, people. And fix my system while you're at it.
 The Verso is something like a Sienna squished into the wheelbase of a Rav4. I wanted an actual mini-van, but neither Honda nor Toyota sold theirs in Europe.