I have a question about IMUs. Nowadays, many IMUs come with built-in sensor fusion using Kalman filters, such as the Mbientlab-MetamotionR IMU. If I'm obtaining angular rate data from the gyroscope, how does the built-in sensor fusion work in this case? Typically, when I integrate angular rate to get the angle of rotation, there's a risk of drift in the measurement.
You're asking if the roll rate is derived directly from the Gyro, from a filter applied to the Gyro, or some calculation based on the absolute orientation determined by the Kalman filter? I can't imagine why you wouldn't derive it from the Kalman filter if you're already running one. In the case of the sensor you mention, the onboard Kalman filter produces a CorrectedAngularVelocity value.
It's worth mentioning, in the Openplotter ecosystem, sensor fusion is performed on the Raspberry pi by an instance of Pypilot, so if you're interested you can always have a look through the Pypilot repo.