Scientific paper ID 493 : 2011/2

The output signal of the MEMS inertial sensors, mounted on vehicles, contains noise components which introduced errors in the signal processing. Lots of methods are proposed to reduce these noisy components, but at the recent years wavelet transformation methods are studied due to their special characteristics. The paper briefly describes three of the methods used for the signal denoising. The properties of the orthogonal and biorthogonal wavelet families are explained. The advantages and disadvantages of each of them are analyzed. A transformation algorithm is proposed according to the described methods, which is distinguished with the choice of a decomposition level, used wavelet method and calculation of the threshold corresponded to the transformation coefficients are treated as zeros. Also an adopted criterion is proposed for the selection of the most appropriate wavelet, based on the filtration effectiveness. It is determined according to the mean square error or the noise reduction coefficient.