Automatic calibration and in-motion alignment of an odometer-aided INS

In-motion alignment problem has been regarded as a challenging problem for years in the extensive study on inertial navigation systems (INS's). In this contribution, the problem of odometer-aided in-motion alignment is investigated, where the nonholonomic constraints are efficiently employed in the designed Kalman filtering process with the measured data provided by a calibrated odometer. As the application of odometer under nonholonomic constraints requires the INS body axes to be well aligned with the vehicle body frame (VBF), INS-to-VBF alignment and calibration of odometer's scale factor are implemented in the INS/Odometer integration. To the end of in-motion alignment, the system and measurement equations for INS/Odometer integration are derived to construct a Kalman filter, which is used to process the integrated velocity information and then to obtain estimates of both odometer error states and INS error states. With these parameter estimates, the odometer outputs can be calibrated properly, which makes the in-motion alignment implementable and practical. The effectiveness of the proposed method is tested through ground based navigation experiments.