| With the contradiction between the tight supply of land resources and the rapid population growth becoming increasingly prominent,the development and utilization of marine resources have become an inevitable trend all over the world.In recent years,with the rapid development of marine resources exploration and exploitation at continuously increasing water depths,the traditional anchor positioning system has been unable to meet the positioning requirements.In order to get improved positioning accuracy,safety and efficiency,dynamic positioning system is therefore an alternative and,sometimes,a supplement to the more traditional solution of anchoring a ship to the seabed.Therefore,dynamic positioning vessels are usually required to be equipped with position reference systems.How to effectively utilize measurement information of position reference systems to improve the accuracy and reliability of the dynamic positioning system is an urgent problem to be solved.Information fusion is an effective method to optimize the combination of multi-source information as it can take full advantage of the complementarities between different messages and improves the accuracy and credibility of measurement information,thereby improving the reliability of the system.Therefore,it is of great significance to get involved in the research of information fusion for position reference systems of dynamic positioning system of ships.The contributions of this thesis are focused on the following work:firstly,a mathematical model of three degree of freedom movement of surface ship is established,at the same time the measurement principles of taut wires system,hydroacoustic positioning system and differential global positioning system are introduced respectively,and the corresponding measurement models are established.Secondly,taking into account the nonlinear characteristics of the dynamic positioning system of ships,two kinds of nonlinear filtering methods,such as particle filter and unscented particle filter,are used to deal with the position measurements that are from the position reference systems of dynamic positioning system.In order to verify the two filtering methods,simulations of the two filtering methods are carried out in Matlab/Simulink environment.And then the two filtering methods mentioned above are compared based on their filtering effects.The simulation results show that the filtering effect of unscented particle filter method is better than particle filter.Finally,the information fusion of measurements of the position reference systems is carried out by using the federated filter with the structure of multi-level fusion.The local filters in multi-level federated filter are used to filter the position measurement values of ship using the unscented particle filter method and the master filter uses the filtered values from local filter and the error-estimation covariance matrix to obtain the optimal position estimation value by estimating the position information of position reference system layer by layer.The designed multi-level federated filter was simulated in the same environment.The simulation results show that the designed multi-level federated filter can achieve good fusion effect and improve the accuracy of ship position measurement information. |
