基于LS-SVM的半潜式海洋平台动力定位系统广义预测控制

摘要

The semi-submersible platform is a nonlinear object with large inertia and large time delay , and will be subject to random interference from wind , ocean currents and waves. To the features of platforms, the least square support vector machine (LS-SVM) with kernel RBF is used to identify the platform model in this paper. And it is combined with generalized predictive control to design a controller. Simulation results show that LS-SVM reduces calculating complexity with high calculating speed. And it also has good generalization ability androbustness.%半潜式海洋平台是一个具有大惯性、大时滞特性的非线性对象,会受到来自海风、海流和海浪等环境因素的随机干扰.针对平台的这一特性,本文采用径向基(RBF)核函数最小二乘支持向量机(LS-SVM)来辨识平台的模型,并且结合广义预测算法(GPC),设计出控制器.仿真结果表明,最小二乘支持向量机能够降低计算的复杂度,并且具有较快的计算速度和良好的泛化能力,该方法具有较好的鲁棒性和稳定性.