加速度计信号的小波形态滤波与样本熵分析

Noise Reduction for the Accelerometer Signals by Using Wavelet + Morphological Filtering Method and Sample Entropy Analysis

针对石英挠性加速度计的测量精度会受到噪声的影响而降低的问题,采用融合小波分析和数学形态学的方法对测量信号进行降噪处理。在采用统计参数的同时,提出了将样本熵复杂度指标用于降噪效果的评价参数,并分别在重力场静态条件和离心机动态条件下进行试验测试。试验结果表明,所采用的方法能有效去除加速度信号中的噪声,同时较好地保留了原始信号的主要特征,且样本熵能够有效描述加速度信号的复杂度和稳定性。

To deal with the problem of the measurement accuracy of quartz flexible accelerometer may decrease under influence of noises, by combining wavelet analysis and mathematical morphological filtering method, the noise reduction process for the measuring signals is conducted. Besides using statistical parameters, it is also proposed that the complexity index of sample entropy can be used to evaluate the effect of noise deduction. The experimental tests are carried out under static condition of gravitational field and dynamic condition of centrifuge respectively. The result of experiments shows that the noise in acceleration signal can be effectively reduced with this method, while the major features of original signal can be reserved, and the complexity and stability of the acceleration signal can be effectively described using sample entropy.