IMU calibration and validation in a factory, remote on land and at sea

This paper treats the IMU calibration and validation problem in three settings: Factory production line with the aid of a precision multi-axis turntable, in-the-field on land and at sea, both without specialist test equipment. The treatment is limited to the IMU calibration parameters of key relevance for gyro-compassing grade optical gyroscopes and force-rebalanced pendulous accelerometers: Scale factor, bias and sensor axes misalignments. Focus is on low-dynamic marine applications e.g., subsea construction and survey. Two different methods of calibration are investigated: Kalman smoothing using an Aided Inertial Navigation System (AINS) framework, augmenting the error state Kalman filter (ESKF) to include the full set of IMU calibration parameters and a least squares approach, where the calibration parameters are determined by minimizing the magnitude of the INS error differential equation output. A method of evaluating calibrations is introduced and discussed. The two calibration methods are evaluated for factory use and results compared to a legacy proprietary method as well as in-field calibration/verification on land and at sea. The calibration methods shows similar navigation performance as the proprietary method. This validates both methods for factory calibration. Furthermore it is shown that the AINS method can calibrate in-field on land and at sea without the use of a precision multi-axis turntable.