基于水下航行器的水平面线性水动力模型,分别使用欧拉法和四阶龙哥库塔法进行操纵运动仿真。然后分别使用欧拉型和积分型辨识样本结构进行水动力导数的辨识。通过辨识误差的比较,证明积分型辨识样本机构具有更高的精度和稳定性。通过分析指出,提高参数辨识精度的一个有效途径是减少数据样本同辨识样本结构的局部截断误差。
Based on the linear hydrodynamic model in horizontal plane of an underwater vehicle, 1-order Euler method and 4-order Runge-Kutta method are respectively used for simulating the ma-neuvering motion. Euler sample structure and integral sample structure for identification are respec-tively used for parametric identification of the hydrodynamic derivatives. By comparing the identifi-cation errors, integral sample structure for identification is proved to be more stable and accurate. By analysis, an effective way for improving the precision of parametric identification is pointed out, i.e., decreasing the local truncation error between the data sample and the sample structure for identifica-tion.