基于AR模型的MEMS惯导随机误差分析方法  

AR Method for MEMS Strap-Down Inertial Navigation Random Error Analysis

作　　者：吕春红 李洋[1] 赵坤[1] 董纯[1] Lv Chunhong;Li Yang;Zhao Kun;Dong Chun(Beijing Aerospace Automatic Control Institute,Beijing 100854,China)

机构地区：[1]北京航天自动控制研究所,北京100854

出　　处：《航天控制》2018年第6期31-35,共5页Aerospace Control

摘　　要：由于微机械惯性器件(MEMS)捷联惯导系统的惯组误差和漂移较大,在较短的时间内也会由于器件误差积累和模型算法误差引起很大的导航偏差。采用Allan方差法分析了MEMS陀螺的随机误差,并对陀螺中主要的5项随机误差系数进行辨识。辨识结果显示,角度随机游走在随机误差中占主要部分。针对MEMS陀螺的实测数据,运用时间序列方法对其随机误差进行分析,并建立了AR模型,试验结果表明,根据所建立的AR(1)模型,采用Kalman滤波方法可以减小其随机误差,滤波后的陀螺输出精度可提高3倍。The deviation and drift of MEMS strap-down inertial measurement unit are large.Meanwhile, under the high dynamic motion of missile,the deviation of navigation system can increase sharply in a short time because of the error of inertial measurement unit and the imprecision of navigation algorithm.Allan variance method is used to analyze the MEMS gyro random errors,and the top main five random error coefficients are identified.Identification results show that the angle random walk is main part of random errors For the measurements of the MEMS gyroscope and using time series method to analyze the random error,the AR model is established,which uses Kalman filtering method to reduce random errors,and filtering precision of the gyro output can be increased by three times.

关 键 词：MEMS惯导 ALLAN方差 时间序列分析 KALMAN滤波 

分 类 号：V241.5[航空宇航科学与技术—飞行器设计]