基于神经网络的MIMU动态误差标定与补偿方法  

MIMU Dynamic Error Calibration and Compensation Based on Neural Network

作　　者：马一鸣 陈帅[1] 王国栋 张琨[1] 程玉 MA Yiming;CHEN Shuai;WANG Guodong;ZHANG Kun;CHENG Yu(Nanjing University of Science and Technology,Nanjing 210000,China;Beijing Institute of Aerospace Control Devices,Beijing 100000,China)

机构地区：[1]南京理工大学,南京210000 [2]北京航天控制仪器研究所,北京100000

出　　处：《电光与控制》2022年第7期91-95,共5页Electronics Optics & Control

基　　金：国防基础科研计划项目(WDZC20190303)。

摘　　要：由于微惯性测量单元(MIMU)易受环境影响,且存在输出非线性以及剧烈角运动和线运动下精度低的缺点,而传统的多项式标定方法难以精确地补偿动态误差,因此,利用深度学习方法对MIMU整体动态误差进行建模与补偿,分别使用浅层神经网络与深度循环神经网络建立MIMU的整体动态误差模型。设计了基于三轴带温箱位置速率转台的标定流程,使三轴带温箱转台的内、中、外3轴同时施加角运动并且施加温度变化,建立MIMU的多因素影响误差训练集。实验结果表明,浅层神经网络模型相对于传统模型在误差补偿效果上略有提升,深度循环神经网络模型补偿后残差均值与均方差显著下降,其中,门控循环单元(GRU)神经网络模型补偿效果最好,并且需要训练的参数较少、计算负担小。The Miniature Inertial Measurement Unit(MIMU)is vulnerable to environmental influencesand has the shortcomings of non-linear output and low accuracy under severe angular and linear motion.Considering that the traditional polynomial calibration method is difficult to compensate for the dynamic error accuratelydeep learning method was used to model and compensate for the overall dynamic error of MIMU.The shallow neural network and the deep recurrent neural network were used to establish the overall dynamic error model of MIMU.The calibration process based on three-axis temperature control position and rate turntable was designedso that angular motion and temperature changes were applied to the three axes of the turntable of the innermiddle and outer axis simultaneously.An MIMU multi-factor influence error training set was also built.The experimental results showed that:1)The shallow neural network model has a slight improvement on the error compensation effect compared with the traditional modeland the deep recurrent neural network model can reduce the residual mean and mean square error significantly through compensation;2)The effect of Gated Recurrent Unit(GRU)neural network model is the bestand there are fewer parameters to be trained and low computational burden.

关 键 词：微惯性测量单元 神经网络 非线性误差 动态误差补偿 

分 类 号：V241[航空宇航科学与技术—飞行器设计]