[ahecht]

The hard part with an INS is the actual programming. Building 190's INS didn't take very long, but getting it to communicate with the RC, getting it calibrated, filtering out noise, and general troubleshooting has been a major challenge. Our programming team spent the entire build season working on it, and has been working with a second INS and the EduRobot since the build season ended, and as I type this, they are probably busy in the pits at Seattle still perfecting it.

The other thing you need to know is that the gyros in the kit are not fast enough to work in a real world situation (I think they only track up to 110 degrees/sec), so we had to buy some 300 degree/sec gyros out of our additional electronics budget. Also, noise control is a big issue. With the noise we are getting, in a worst case scenerio, our INS can think it has reached escape velocity in 15 seconds while the robot is standing still (although it usually averages out to provide reasonable results during our autonoous mode, when we really need it). Don't forget that you need to do a double integral to determine position, so any noise in acceleration or yaw rate would be cubed.

I'd encourage you to go ahead and try it. It is always to to see FIRST teams pushing the envelope.