基于MPU6050加速度信号的匹配小波设计及去噪

Design of matching wavelet and denoising based on MPU6050 acceleration signal

MPU6050传感器的加速度信号在实际测量中往往会存在一些干扰噪声。目前针对加速度信号去噪,人们往往会选择前人所构造的小波基,诸如Harr小波、Daubechies小波等。这些小波基没有从实际的信号出发去最大化地匹配滤波器的输出,导致加速度波形会出现较为明显的失真情况。本文采用了能量匹配的方法,利用构造出的9/7小波结合能量匹配原则找出最优小波基。在此基础上再用相邻的高频系数波形图比较自适应确定最优分解层算法确定分解层次,然后结合构造出的最优小波基进行去噪,得到不失真的加速度信号波形。

During the experiment, the signal of the MPU6050 often exists interference noise. Currently people often choose wavelet bases constructed by predecessors for the acceleration signal de-noising, such as Harr wavelet, Daubechies wavelets, these wavelets have not maximized the matched filter output signal for the actual acceleration signal ,which resulted in obtaining the signal that can not match the original acceleration signal greatly. What＇s more ,it resulted in an acceleration waveform appearing more obvious distortion. In this paper, we use the matching method to construct 9/7 wavelet, and use the principles of matching energy to find the optimal wavelet basis. On this basis, we compare the adjacent frequency coefficients waveform diagram to determine the optimal decomposition level adaptively, and then use the constructed optimal wavelets bases to denoise for matching the original acceleration signal to the maximum extent. Furthermore ,we can get the waveform of acceleration signal which is not distorted in the process.