注：国家自然科学基金资助项目（NO. 61571053）

作者：张伯源，高国伟

单位：北京信息科技大学传感器重点实验室，北京 100101

中图分类号：V241.5                 

文献标识码：A                    

文章编号：1006-883X(2019)03-0018-07

收稿日期：2019-02-27

 

摘要：针对MEMS陀螺仪的输出随机漂移误差影响测量精度的问题，提出一种改进的卡尔曼滤波方法进行MEMS陀螺仪误差补偿。传统的卡尔曼滤波方法是针对时域内的随机序列采用统计特性进行递推估计，从而得到测量所需要的信号。本文在传统卡尔曼滤波算法的基础上引入衰减因子和差分控制项，以此自适应地估计卡尔曼滤波量测噪声方差，并结合硬件系统将该算法进行静态性能试验和动态性能试验，使用Allan方差分析法对原始陀螺仪信号以及误差补偿后的陀螺仪信号进行对比分析。对比数据结果表明，陀螺仪静态随机误差得到了有效的抑制，从而验证了该算法在陀螺仪静态数据处理方面具有一定的应用价值。

关键词：MEMS陀螺仪；Allan方差；随机漂移；自适应卡尔曼滤波

 

Application of Improved Kalman Filter in Signal Processing of MEMS Gyroscopes

ZHANG Bo-yuan, GAO Guo-wei

Beijing Sensor Key Laboratory, Beijing Information Science & Technology University, Beijing 100101, China

Abstract: Aiming at the problem that the random output drift error of MEMS gyroscopes affects the measuring accuracy, an improved Kalman filter method is proposed to compensate the error of MEMS gyroscope. Traditional Kalman filter is a recursive estimation of random sequences in time domain using statistical characteristics, so as to obtain the signal needed for measurement. In this paper, a attenuation factor and a difference control item are introduced on the basis of the traditional Kalman filtering algorithm to estimate the variance of the noise measured by Kalman filter adaptively. Combining with the hardware system, the static performance test and dynamic performance test of the algorithm are carried out, and the results of the original gyroscope signal and the gyroscope signal after error compensation are compared and analyzed by Allan Variance analysis, and the comparison data show that the static random error of the gyroscope can be effectively suppressed. Thus, it is proved that the algorithm has certain application value in the static data processing of gyroscope.

Key words: MEMS gyroscope; Allan variance; Random drift; adaptive Kalman filter

 

 

备注：2019年 第25卷 第03期