实时小波滤波方法在硅微陀螺仪中的应用研究

Application of wavelet’s real-time filter in silicon micromachined gyroscope

采用实时小波滤波技术构建硅微陀螺仪数字化平台,对快速的Mallat算法选用何种小波基,是否采用软硬阈值处理以及选用何种尺度和采样点数进行深入分析。考虑到滤波效果即陀螺仪输出信号滤波前后Allan方差改进情况和硬件处理的实时性,通过比较最终选择3尺度16点软阈值db2小波Mallat快速算法。仿真计算和实验结果表明该滤波方法对改善微机械陀螺仪的零偏稳定性有一定的实用价值。

The paper presented a real-time wavelet filter technology for constructing digital platform, and analyzed which wavelet basis to select, whether to apply soft or hard threshold processing, and what scale and sampling sample numbers to select. Considering the filter effect （i.e. the Allan variance improvement extent before and after the filter processing, including the real time and high effective algorithm capability of hardware processing）, the db2 wavelet mallat fast algorithm with 3 level scale and 16 points was selected by comparison. The computation simulation and experiment results show that this filter processing method is much practical and valuable to improve the zero bias stability of micromachined gyroscope.