【摘要】 为获得横摇运动在不同时间尺度下的演变规律,提出基于小波变换(WT)理论进行船舶横摇运动非线性时间序列分析与预测的方法.通过小波变换对横摇运动时间序列进行多分辨率分析(MRA),将原序列分解为多个相对简单的准周期信号,对信号的趋势项、周期项和随机项进行分离,并采用人工神经网络(ANN)模型对上述准周期信号进行预报和集成.仿真结果表明:该方法有效提高了预报长度,并可获得较高建模及预报精度.更多还原

【Abstract】 The analysis and prediction approach based on wavelet transform(WT) was presented and applied to the prediction of ship roll motion nonlinear time series to obtain the evolvement rule of ship roll motion under different time scale.Multi-resolution analysis(MRA) using WT was applied to the ship roll motion time series decomposed into some relative simple and regular period signal series according to the scale.The trend term,periodic terms and stochastic terms were separated from original series,and the artificial neural network(ANN) prediction mode were employed to predict these approximate period signals.Simulation results show this method can improve the prediction length and has better prediction precision.更多还原