Using a test bench would be the real best way to test the whole sensor + filter + processing thing. Would need both linear motion and tilt. A true test bench would be best, for my part I "simply":
outputed the accelerations in NED frame (after quaternion rotation) and yaw pitch roll on serial
tested by moving the IMU around: "looking up" and oscillate up down / right left, "looking on the side" (ie 90 degrees from up) and oscillate up down / right left, tilt, etc, and check that the output is consistent
of course a nice test bench would be much better :) .
Using a test bench would be the real best way to test the whole sensor + filter + processing thing. Would need both linear motion and tilt. A true test bench would be best, for my part I "simply":
of course a nice test bench would be much better :) .