基于多新息最小二乘法的船舶操纵响应模型参数辨识

摘要

提出基于多新息最小二乘法的船舶二阶非线性响应型模型参数辨识方法.在实验室环境下对模型船开展Z形试验,分别应用最小二乘递推法和多新息最小二乘法对试验数据进行参数辨识;用得到的模型分别进行Z形试验预报,比较2种方法的准确性和辨识结果的收敛性.试验结果表明:基于多新息最小二乘法的船舶操纵响应模型参数辨识结果收敛更迅速且精度满足要求.%A parameter identification method for second order nonlinear ship response model is put forward based on the multi-innovation least squeres algorithm.Ordinary recursive least squares algorithm and the proposed MILS (Multi-Innovation Least Square) algorithm are applied respectively in parameter identification with the Z test data of the ship model.The identified parameters are used to do Z test prediction.Prediction accuracy and convergence of identified maneuvering response model parameters from two processes are compared.The comparison proves that identification results based on the MILS algorithm is more accurate and have a higher convergence speed.