船舶动力定位系统的广义预测控制方法研究

Research on Generalized Predictive Control Algorithm in Ship Dynamic Positioning System

船舶在海上的航行是一种复杂的非线性运动,其水动力参数很难精确确定,并且会遭遇来自风、流和浪的随机干扰;因此,船舶须要具有鲁棒性和自适应能力的动力定位系统。鉴于广义预测控制算法在非线性控制方面的独特优势以及神经网络具有自学习和自适应的能力,作者研究了基于支持向量机的广义预测混合控制算法,并将其应用于船舶动力定位系统。仿真结果表明,该方法具有较好的鲁棒性和自适应性,提高了动力定位系统的精度和性能。

Ship motion is complicated and nonlinear during voyage course.It is fairly difficult to determine the hydrodynamic parameters related to ship motion,which may be disturbed randomly by wind,current and wave.Therefore,it is necessary for ship to be equipped with Dynamic Positioning System possessing robust and adaptive properties.In view of the superiority of Generalized Predictive Control in nonlinear control and neural networks in self-learning and adaptability,the hybrid algorithm of Generalized Predictive Control based on Support Vector Machine is studied in this paper.This algorithm has been applied to ship dynamic positioning system.It is shown that in the study case the algorithm has better robust and adaptive properties and the System precision and performance of dynamic positioning are improved.