基于UKF联邦滤波的动力定位船舶运动状态估计

摘要

针对动力定位船舶，采用基于无迹卡尔曼滤波(UKF)的联邦滤波技术，利用多种位置参考系统和电罗经的测量值来实现对船舶运动状态的估计。从传感器测量原理出发，采用非线性的测量方程，实现对多种位置参考系统传感器特征的描述，利用 UKF 滤波方法和联邦滤波结构将动力定位船舶配备的多个冗余位置参考系统的测量数据进行有效融合，从而提高对船舶运动状态估计的精度和可靠性。仿真结果表明，采用基于UKF的联邦滤波方法可以有效实现对动力定位船舶相关运动状态的估计。%For the dynamic positioning ship, based on the Unscented Kalman Filter (UKF), a federated filter is designed to estimate the ship motion states by using the measurements from different position reference systems (PRS) on the ship. From the point of measurement principle, the nonlinear measurement equations are used to describe the characteristic of different kinds of PRS. With the combination of UKF filter method and federated filter structure, this method can effectively achieve a good data fusion effect for the redundancy PRS to accomplish the goal of estimating the ship motion states with good accuracy. The simulation results show that the UKF based federated filter can meet the need for the state estimation for the dynamic positioning ship.