Information Fusion For Position Reference Systems Of Dynamically Positioned Ship

The exploration and exploitation of ocean resource such as oil and gas accelerate the development of dynamic positioning system (DPS). The reliability and accuracy of DPS have received more and more attention. According to the relevant rules of International Maritime Organization (IMO) and other classification societies, all the dynamically positioned ships should be equipped with multiple position reference system (PRS) so that the stability and accuracy of DPS are guaranteed. This paper studies the information fusion of multiple PRS to improve the reliability and accuracy of DPS.Firstly, the motion of dynamically positioned ship is analyzed to achieve the mathematical model of ship. The mathematical models of three kinds of PRS including Taut Wire system, Hydro-acoustic system and DGPS, are discussed according to the principles of position reference. Secondly, both of unscented Kalman filter (UKF) and adaptive unscented Kalman filter (AUKF) are employed to filter the data from each PRS to obtain the ship position. Finally, the paper adopts federated filter to fuse the information from three kinds of PRS. The federated filter is a two-stage filter composed of several local filters and a master filter. The local filters are designed based on AUKF for one of the PRSs. The master filter fuses the results of the local filters to obtain the optimal ship position according to their error variances. Fault detection is introduced in front of the master filter based on fusion-reset federated filter, so as to design a fault tolerant federated filter with fault detection and isolation.Simulation study involving a supply ship is included to illustrate the effectiveness of filters and federated filters. As the results showing, both UKF and AUKF have the ability to filter the data from PRSs and estimate the ship position. However, the performance of AUKF is better than UKF. In addition, the accuracy of information fusion based on the federated filter is better than that of single PRS. Thus, using the information fusion method can not only improve the accuracy of position measurement of DP vessel, but also make the measurement system has a certain degree of fault tolerance.