基于神经网络及系统辨识的舵机带宽测试  

Testing system for rudder’s performance based on neural networks and linear sys-temidentification

作　　者：习赵军[1] 李敏[2] 李昌禧[1] 

机构地区：[1]华中科技大学控制科学与工程系 [2]武汉理工大学电信系,湖北武汉430074

出　　处：《自动化与仪器仪表》2008年第5期58-59,73,84,共4页Automation & Instrumentation

摘　　要：讲述了一种应用神经网络辨识算法测试舵机带宽的实用方法。文章简要介绍了舵机的工作特性和舵机模型的选取,概述了辨识算法的选取及实现过程,并在输入信号的选取和辨识数据的预处理等方面作了基本的探讨。实验仿真结果表明,基于线性神经网络的系统辨识具有很高的辨识速度和精度。An efficient method for testing the bandwidth of actuator, linear system identification based on neutral network, is presented in this paper.The paper briefly introduces the working characteristics of actuator,the method for choosing the model of actuator and the identification arithmetic. It also discusses how to chose the input and preproeessing the data for identification.The simulation results indicate that linear neural network based system identification do have very high speed and precision of identification.

关 键 词：舵机带宽 系统辨识 神经网络 

分 类 号：TP274.5[自动化与计算机技术—检测技术与自动化装置]