一种基于参数估计的自适应舰船运动预报方法

摘要

Motion prediction is important to some on board systems. In order to predict the ship motion effectively, an adaptive method was developed based on estimation of the peak frequency of sea weave. Ship motion and wave excitation were formulated, and a self-reset estimator for the peak frequency of sea weave was established based on least squares approximations. The ship motion was predicted by Autoregressive Integrated Moving Average (ARIMA) model fitting, and the sampling frequency of the ARIMA model was adap-tively adjusted by the estimated value of the peak frequency. Therefore, the prediction method could adapt to the fluky sea state with high precision. A comparison between the proposed method, ordinary ARIMA method, and Back-Propagation (BP) neural network was made, and the result shows that the proposed method is good in both performance and robustness.%运动预报是部分舰船系统的重要组成部分.为了有效地解决这一问题,文章提出了一种基于海浪峰值频率估计的自适应舰船运动预报方法.在舰船运动与海浪激励的建模基础上,建立了基于最小二乘估计的自复位海浪峰值频率估计器.采用自回归移动平均(ARIMA)模型拟合方法预报舰船运动,并通过海浪峰值频率估计值自适应调节ARIMA模型的采样周期,提高了复杂海况下对舰船运动的预报能力.该方法与常规ARIMA模型方法、反向传播神经网络方法的仿真对比结果表明了该方法在解决舰船动态预报问题上的良好精度和鲁棒性.