thanks again for your report. Just to give you some feedback:

I could successfully track down the problem to a bug in the 
UserClient part of Apple's IOUSBInterface service object. While I 
could reproduce your kernel panic within a 100 x launch/quit cycle, 
the release yesterday didn't show any problems at all even after 
running a 2500 x launch/quit cycle over the course of a day.

The release published yesterday works around the bug in a way so that 
the kernel panic can no longer occur.

I'll file a bug with Apple as soon as I find some time to put 
together some sample code for them to contribute to the overall 
stability of Mac OS X.

Best regards, 
Felix