MEMS惯性传感器随机误差分析与去噪研究

Analysis and Denoising MEMS Inertial Sensors Stochastic Error

针对MEMS传感器中存在的误差,采用Allan方差法分析法分析其存在的误差类型,并通过改进小波阈值函数、调整分解尺度观察存在的误差项在去噪前后的变化,探究各误差项与阈值函数、分解尺度之间存在的关系,从而有针对性地对MEMS惯性传感器中存在的特定随机误差进行降噪;结果表明:几类改进阈值函数对角(速)度随机游走的抑制效果与软、硬阈值函数无明显差异,效果并不理想;不同的尺度分解可以去除不同的误差项,从而提高MEMS传感器精度。

To address the errors in MEMS inertial sensor,the Allan variance was used to analyze the error types.By means of improving the wavelet thresholding function and changing the decomposition level to observe how the error types change before and after denoised and explore weather there is a qualitative relationship between them.The result shows that the improved wavelet thresholding functions have no advantage in restraining angle（velocity）random walk than hard and soft thresholding function,but different decomposition levels can remove different error term,so it can be used to improve the MEMS sensor precision.