A New Cubature Kalman Filtering with M-method for Attitude Determination

The inertial navigation system (INS) aided by a magnetic compass (MCP) is designed for the underwater glider to obtain underwater navigation with higher accuracy, higher stability and faster tracking even though sudden change of attitudes. Even if the measurement becomes unobservable, the CKF still has the better determination precision. However, the computation of CKF is much larger than that of EKF. The paper proposes a novel algorithm integrating cubature Kalman filtering with M-method (CKF-M), which takes full advantage of efficient computation and anti-interference to acquire the navigation information with higher accuracy during long-term gliding. In the real glider experiments, the CKF-M has a good effect in the aspects of accuracy, stability and tracking even if the sudden change of attitude.