The imu is essentially a stream of data. The timing of this stream can be roughly calibrated by comparing with radio and motor data. (Motor changes, then either radio or imu changes, then the other.)

With the idea of sliding comparison, you might transform this data to look at absolute magnitude of change over time, and normalise it. Reviewing and comparing the streams as a visual graph would help inform whether or not they are then automatically correlatable.

So, you'd make a calibration test function that produces short bursts of graphable data around instigating motion.

Then you'd make the calibration function itself, but you'd keep the test function so that people could use it to troubleshoot issues they run it.