优化NARX神经网络重力仪稳定平台模型辨识方法  

Model Identification Method of Gravimeter Stable Platform Based on Optimized NARX Neural Network

作　　者：李昂 鲁鑫 林恩凡 LI Ang;LU Xin;LIN Enfan(No.3 Military Representative Office Stationed in Beijing Area,Naval Equipment Department,Beijing 100071;No.1 Military Representative Office Stationed in Tianjin Area,Naval Equipment Department,Tianjin 300131;College of Electrical Engineering,Naval University of Engineering,Wuhan 430033)

机构地区：[1]海军装备部驻北京地区第三军事代表室,北京100071 [2]海军装备部驻天津地区第一军事代表室,天津300131 [3]海军工程大学电气工程学院,武汉430033

出　　处：《舰船电子工程》2021年第8期151-154,共4页Ship Electronic Engineering

摘　　要：针对重力仪稳定平台非线性因素多,且实验条件限制,通过传统的模型辨识方法难以建立平台的精确模型的问题,提出了一种基于优化带外输入的非线性自回归模型(Nonlinear Auto Regressive Models with Exogenous Inputs,NARX)的神经网络模型辨识的方法对重力仪稳定平台进行精确建模,从而提高模型与实际系统的一致性,为PID参数整定方法应用于工程奠定基础。通过实验建模及验证的结果表明:该方法相比较于传统方法建模精度提高7倍~23倍左右,模型精度提高一个数量级,且方法简单,应用范围更广。In view of the many non-linear factors of the gravimeter stable platform and the limited experimental conditions,it is difficult to establish an accurate model of the platform through traditional model identification methods,a method based on optimizing the nonlinear autoregressive models with exogenous inputs(NARX)neural network model identification method is proposed to accurately build the gravimeter stable platform.It can improve the consistency between the model and the actual system and facilitate the tuning of PID parameters.The results of experimental modeling and verification show that compared with the traditional method,the modeling accuracy of this method is increased by about 7 to 23 times,the model accuracy is improved by an order of magnitude,and the method is simple and has a wider application range.

关 键 词：重力稳定平台 模型识别 神经网络 双闭环PID控制 参数整定 

分 类 号：U666.121[交通运输工程—船舶及航道工程]