摘要
针对随机海浪作用下船舶运动的非平稳、非线性特性,文章提出了基于卡尔曼(Kalman)滤波原理的非线性二阶Volterra级数自适应预报模型。通过把Volterra级数核向量作为状态向量,利用随机游动模型建立系统的状态方程,一步Volterra级数预报模型作为系统的观测方程,从而进一步提高了Volterra级数模型的核估计的收敛速度。同时验证了利用AIC准则对Volterra级数预报模型定阶的可行性,通过迭代法实现了自适应多步预报。仿真结果表明文中提出的基于Kalman滤波算法的自适应预报模型应用于船舶运动极短期预报是可行的,该方法在理论和工程应用方面具有重要的意义。
Aiming at non-stationary and non-linear nature of the ship motion, the application of Kalman filtering in non-linear second-order Volterra series model is proposed.The system＇s state space model is established by taking the Volterra series kernals vector as the state vector, the observation equation can be obtained by Volterra series model. Then the convergence rate is improved further and carried out adaptive ship motion prediction combining AIC criterion.The multi-step simulation results obtained show that non-linear second-order Volterra series model based Kalman filtering is feasible in the predicting skill.The approach has vital significance in the theory and the engineering application.
出处
《船舶力学》
EI
北大核心
2010年第7期732-740,共9页
Journal of Ship Mechanics
