针对船舶动力定位问题提出了基于线性核函数在线支持向量回归的模型预测控制方案。在线支持向量回归算法的引入可以通过在线调整,确保预测模型的精确性。基于线性核函数的模型预测控制能够方便、迅速地求取控制律的解析表达式,保证控制律的最优性以及算法的快速性。仿真结果证明了该控制方案的有效性。
Model predictive control strategy based on linear kernel online SVR is proposed to solve ship dynamic positioning problem. The introduction of online SVR algorithm can ensure the accuracy of predictive model by online adjustment. Model predictive control based on linear kernel is able to obtain the analytical solution of control law conveniently and fleetly, which guarantee the optimal control law and expeditious algorithm. The simulation results show that the proposed control strategy is valid.