基于小波变换的车轮力传感器信号的去噪研究

Research on De-noising of Signals of Wheel Force Transducer Based on Wavelet Transform

在汽车道路试验中,通过多维车轮力传感器(WFT)可以测量每个轮所受的各维力和力矩。在测量过程中,信号会不可避免地受到各种噪声的干扰,而且,在将测量数据从车轮坐标系转换到车辆坐标系时,车轮转角的误差使测量结果产生了更严重的噪声。这些宽带随机噪声严重影响了车辆性能的分析。小波分析是一种信号的时间-尺度分析方法,特别适合于非平稳信号的分析,具有多分辨率分析特性,而且在时频两域都具有表征信号局部特征的能力。针对车轮力信号的特点,在MATLAB环境下编程进行车轮力信号小波变换去噪研究,试验结果表明,在选择了适当的小波基本函数和阈值的情况下,采用小波变换的阈值去噪方法对多维车轮力信号进行去噪处理,可以取得良好的效果。

Multi-dimensional forces and torques can be measured by multi-axes wheel force transdueer（WVF） in automobile mad experiment. The signals were inevitably influenced by kinds of noises, and awful noises were generated when test data were transformed from wheel coordinates to vehicle coordinates because of the measurement error of wheel rotating angle. These wide band stochastic noises badly influenced the analysis of vehicle performance. Wavelet transform is an analytical method of time-freqneney, which is especially suitable for the analysis of non-stable signal. The method has a eharaetex of multi-resolution analysis, and has ability of showing part character of signals in time-frequency domain. In view of the characteristics of wheelforce signals, de-noising methods based on wavelet transform were researched under MATLAB. The results suggest that effective outcome of wheel force signals could be got by the method of threshold value de-nuising if the suitable wavelet base and threshold value were chosen.