Elman神经网络及其在激光陀螺输出误差建模中的应用研究  

Neural Network and Its Application in RLG Output Error Modeling

作　　者：刘永红[1] 杨孟兴[1] 吴一[1] 

机构地区：[1]中国航天时代电子公司第十六研究所,西安 710100

出　　处：《导航与控制》2005年第3期16-20,共5页Navigation and Control

摘　　要：惯性器件误差补偿技术对提高捷联惯导系统的导航精度具有十分重要的意义，而误差补偿的关键在于误差模型的辨识。本文探讨将Elman神经网络应用于惯性器件误差建模中，详细介绍了Elman网络和其对应的学习算法。仿真结果证明了该方法的可行性，并且，该方法具有网络收敛速度快、跟踪性能好、稳定性好等优点。The technology of inertial sensors error compensation is important to improve the accuracy of SINS, and the identification of error model is the key to compensating errors . This paper tried to model errors of RLG by use of Elman neural network, Elman neural network and learning arithmetic were presented detailedly in this paper. From the results of simulation show the feasibility of this method and also the Elman networks has some advantages for modeling errors of laser gyro ： rapid convergence, good performance of tracking and stabilization.

关 键 词：捷联惯导系统 激光陀螺 误差模型 ELMAN神经网络 学习算法 误差建模 网络应用 误差补偿技术 ELMAN网络 输出 

分 类 号：TP183[自动化与计算机技术—控制理论与控制工程] TG659[自动化与计算机技术—控制科学与工程]