【作者】 王大勇；

【导师】 张广莹；

【作者基本信息】 哈尔滨工业大学 ， 控制科学与工程， 2009， 硕士

【摘要】 光纤陀螺自问世以来,在短短30多年时间里取得了日新月异的发展。由于其自身独特的优点,目前光纤陀螺已广泛应用于航空、航天、航海及陆地的军用和民用的各个领域。本文研究了光纤陀螺的工作原理;分析了影响其测量精度的各种随机误差源,这些误差源主要包括量化噪声、角度随机游走、零偏不稳定性、速度随机游走和速度斜坡等;介绍了用于识别这些误差源的Allan方差分析法,并用这种方法对光纤陀螺的实测数据进行了分析。时间序列分析是数理统计的一个重要分支,它是采用参数模型对所观测到的有序随机数据进行分析与处理的一种数据处理方法。本文借助时间序列分析知识,对光纤陀螺的实测数据进行了分析处理,得到了满足平稳时间序列建模要求的随机漂移信号。在此基础上建立光纤陀螺随机误差的时间序列模型,为采用滤波方法对随机误差进行补偿做好了准备。卡尔曼滤波是一种递推算法,在模型精确、噪声统计特性准确的情况下,可以得到系统状态的最优估计。鲁棒滤波是针对系统模型、噪声统计特性不精确性的一种滤波算法。本文通过仿真实验,比较了这两种滤波算法的特点,并用两种滤波算法分别设计了光纤陀螺随机误差的补偿方案,通过对实测数据的滤波处理,借助Allan方差分析法,验证了鲁棒滤波比卡尔曼滤波更加适用于光纤陀螺随机误差的补偿。更多还原

【Abstract】 Fiber-optic gyro(FOG) develops rapidly in 30 more years after being invented. Due to its advantages, fiber-optic gyros have been widely used in military and civilian fields of aviation, aerospace, navigation and land currently.This paper researches working principle of FOG; analyses all kinds of random error sources affecting its measurement precision, these error sources mainly contain Quantization noise, Angle random walk, Bias instability, Rate random walk, Rate ramp; introduces Allan variance law which is used to identify these error sources and analyses measurement data of FOG with this method.Time series analysis is an important branch of Mathematical statistics, which is a data processing method using parametric model to analysis and process random data measured which has order. With the help of time series analysis, this paper analysis and processes measurement data of FOG, obtains the random drift signal which is be able to modeled in the law of steady time series, laying a good foundation for compensating random errors using filtering method.Kalman filtering is a recursive algorithm, it can obtain optimal estimation of system states when there are precise model of system and precise statistical properties of noises. Robust filtering is a filtering algorithm against uncertainty of system model and noises statistical properties. This paper compares features of these two filtering algorithms through simulation and designs the compensation program of random errors of FOG. By means of processing the measurement data and with the help of Allan variance analysis, this paper verifies that robust filtering does better in compensating random errors of FOG than Kalman filtering.更多还原