微机电陀螺仪Allan方差算法研究

Research on Allan variance algorithm for micro electro mechanical gyroscope

Allan方差是评估随机误差的有力工具。本文基于微机电系统(micro electro mecianicel system,MEMS)陀螺仪的静态数据,对比分析了非交叠式Allan方差、非完全交叠式Allan方差以及完全交叠式Allan方差的噪声性能。结果表明:非交叠式Allan方差计算效率较高,但曲线出现振荡,不利于噪声参数识别;完全交叠式Allan方差曲线最为平滑,但计算时间明显增加;而非完全交叠式Allan平衡了计算时间和精度。

Allan variance is a powerful tool for analyzing random errors.In this paper,the static data of micro electro me-chanical system(MEMS)gyroscope is used to compare the noise analysis performance of the non-overlapping,not fully overlapping and fully overlapping Allan variance.The results show that calculation efficiency of the non-overlapping Allan vari-ance is higher,but the curve has oscillation,which is not conducive to noise parameters identification.The fully overlapping Al-lan variance curve is the smoothest,but the calculation time increases obviously;while the not fully overlapping Allan variance keeps a balance between calculation time and precision.