A New Loosely Coupled DCM Based GPS/INS Integration Method.

ABSTRACT In this paper, we derive a GPS/INS integration algorithm based on the direction cosine matrix (DCM) attitude representation model and analyze its performance. The motivation for this work is to find an algorithm that is capable of tracking the variation of the gyro bias vector during operations and that benefits from the established DCM based attitude estimation algorithms. Filter performance is evaluated using a simulation based on an air vehicle trajectory profile and is compared with an Euler angles-based unscented Kalman filter (UKF). We investigate the effects of improper initialization of the state vector on the filter performance. These results are supplemented using experimental data collected along a car route with two different grades of inertial measurement units (IMUs). The DCM based model has some advantages over other attitude representations due to its relatively moderate computational load and applicability with different categories of inertial sensors. Copyright © 2012 Institute of Navigation [ABSTRACT FROM AUTHOR]