基于Sage-Husa自适应滤波器的MEMS陀螺随机误差建模补偿

Stochastic Error Modeling Compensation of MEMS Gyroscope Based on UKF Filter

MEMS陀螺随机漂移误差是制约惯性导航精度的关键因素,该文所设计的方法主要为了能够提升陀螺漂移精度。首先针对标准Kalman滤波器陀螺漂移处理方法中,随机动态系统的结构参数和噪声统计特性参数不准确的问题,采用自适应SHAKF(Sage-Husa Adaptive Kalman Filter)滤波器进行参数实时估计,然后通过建立了ARMA随机误差模型,搭建了MEMS陀螺组件实验系统,通过高精度三轴转台静态测试采集陀螺数据。Aallan方差分析表明,零偏不稳定性经线性KF滤波后提升17.4%,经自适应SHAKF滤波后提升26.2%。SHAKF滤波较标准Kalman滤波方式有明显优势。

The random drift error of MEMS gyroscope is the key factor to the precision of inertial navigation.According to the standard of gyro drift of Kalman filter processing method,statistical characteristics of structural parameters and noise parameters of random dynamic system precision problem,using adaptive SHAKF(Sage-Husa Adaptive Kalman Filter)filter for real-time estimation of parameters,improve the accuracy of gyro drift.Based on this idea,the ARMA random error model is established,and the experimental system of MEMS gyro component is built,and the gyro data are collected by the high precision three axis turntable.Aallan variance analysis shows that the zero bias instability is increased by 17.4%after the linear KF filter,and the adaptive SHAKF filter is improved by up to 26.2%.