An Augmented State Robust Unscented Kalman Filtering for Gravity Gradient Matching Navigation

As a completely autonomous and passive positioning method, gravity gradient matching can provide high-precision positioning services for submarines during long underwater voyages, holding great military application prospects. Aiming at the problem of measurement disturbances in the gravity gradiometer, an augmented state robust unscented Kalman filtering (ASRUKF) gravity gradient matching algorithm is proposed. This algorithm combines the Huber-based robust unscented Kalman filter (RUKF) and the augmented state unscented Kalman filter (ASUKF) to automatically address various interference scenarios, such as measurement biases, discrete outliers, and mixture Gaussian distributions. It compensates for the limitations of single algorithms and exhibits stronger practicality. Ship experiment results demonstrate that the proposed algorithm can effectively mitigate the impact of colored noises, including measurement biases, discrete outliers, and mixture Gaussian distributions, significantly improving the navigation performance of the system. Additionally, the effects of data-sampling periods and gradiometer noise levels on the system are also evaluated.