A Filtered-Marine Map-Based Matching Method for Gravity-Aided Navigation of Underwater Vehicles

Abstract

Gravity-aided inertial navigation system (INS) has been widely applied in the application of autonomous underwater vehicle. Due to the uneven distribution of the gravitational field, the matching performance varies unsteadily when underwater vehicles pass through real marine environment. In this article, a filtered marine map-matching method based on sliding window iterated closest contour point (ICCP) algorithm is proposed. The filtered marine-map records the information of the contour line, which is extracted from the gravity anomaly value. The trajectory can be represented by different vectors between the contour lines corresponding to the sampled gravity points. In addition, a model-free matching area selection method is proposed. A multilayer support vector classifier is established by hierarchical classification method. The genetic algorithm is adopted to optimize the penalty factor and kernel function parameters. The marine experimental results indicate that the proposed matching area selection method can fully excavate the gravity information and provide directional guidance. The proposed filtered marine map-matching algorithm effectively improved the matching accuracy for underwater vehicles.