This is Brad's suggestion (WPIlib code here) so you can drive around independent of heading. This has been tried in the past on the Cortex and it didn't work well enough to be used for a match. The gyro drifts too much. V5 would have to perform substantially better for this to work.

Alternatively, instead of taking a gyro object, we could do this the WPIlib way and just take the value of the heading so we aren't locked in to using a gyro. Then the heading measurement could come from a combination of quads & a gyro.

Gyro is waiting on Vex fixing the ADI port analog in sampling rate.