结合舵减摇控制中通用线性模型与模糊线性模型的优缺点,采用了一种根据遭遇角的不同自动切换控制器设计模型的并行双模切换舵减摇控制方案。为在线自动辨识遭遇角,提出基于横摇功率谱密度柱状图的神经网络辨识方法。仿真结果表明,该方法辨识精度较高,且比传统重心法节省了在线计算量与存储数据库所需的存储空间,简化了辨识流程。
This paper dealt with the advantages and disadvantages of the general linear model and fuzzy linear model on rudder roll stabilization. A parallel two mode switched control scheme was adopted based on the encounter angle. In order to identify the encounter angle on line automatically, a neural network identification method was presented on the roll power spectral density (PSD) histogram. The simulation result shows that this method is not only high in precision, but also saves the computer quantity and memory space as compared with the traditional one, simplifying the identification process.