I would introduce an ideal set of data, and introduce an error, and see how fast it approaches the ideal gain. After that, its just a matter of making sure the gyros are spitting out the correct data.

Update:

http://rp181.fortscribe.com/?p=520

10MP picture: http://api.ning.com/files/9RiMwTT9l0T-v ... 0_5450.JPG
I soldered together the helicopter IMU after the long wait for the board (It was worth it, batchPCB doubled my order). The only things not connected is the camera, GPS, Wireless transmitter, MAG, and pressure sensor. All of these go off with cables to various parts of the helicopter. I still have two expansion ports available on seaperate cores available, the only thing planned for the first core is the wireless connection (all other mentioned sensors connect to the shield), and the second will be a memory controller for external RAM, in order to proccess larger images.
You can see the ADC's poking out, one samples the gyros and the temperature outputs on the gyros, and the other does the accelerometer. I have some spare analog inputs, as well as around 12 digital spare (6 broken out on board).
Got all the helicopter parts, but on first spin up, the tail blade got stuck in some blankets and destroyed the belt transmission gear.
I did some surgery to remove a bad cell from the 6s 5000mah battery I got from hobbyking, ended cutting open a cell to get it off as its glued together very well. Very nerve-racking.

Thanks for the update! How long do you see the battery lasting for flight time?

It's hard to tell, but about 10 minutes.