Federated filter for multi-sensor data fusion of dynamic positioning ship

Abstract

High precision and a good capability of fault tolerant are more and more required for today's dynamic positioning system. In this paper, a new kind of data fusion method is presented for the DP system by using federated filters and Unscented Kalman Filter technique. This method could overcome the single sensor's weakness, and the UKF technique bring the algorithm with a good precision for nonlinear problem, while the federated structure significantly enhances the filters' capability of multi-rate information fusion, fault tolerance, and system modularity.