The accurate synthesis of realistic waveforms conforming to certain specica-

tions is a fundamental step in random vibration testing. Since real-time imple-

mentation of digital signal processing systems for random vibration and noise

synthesis necessarily operates frame by frame, the overlap-add (OLA) method,

by which frames are windowed and overlapped, is widely used in practice to

avoid artifacts at frame boundaries. When a wide-sense stationary random

signal is desired, however, the OLA method presents a shortcoming, because

the inherent periodicity of the frame-by-frame process unavoidably produces a

cyclostationary signal, i.e., its statistics present an undesired periodic behav-

ior. We analyze the impact of the window coecients in the cyclostationarity

properties of the synthetic process, and then present algorithms for window de-

sign with the goal of maximizing a measure of its stationarity, considering both

second- and fourth-order statistical properties. The proposed designs are shown

to signicantly improve the stationarity properties when compared to commonly

used windows.

%B Mechanical Systems and Signal Processing
%V 122
%P 642-657
%8 05/2019
%G eng
%R 10.1016/j.ymssp.2018.12.038
%0 Conference Paper
%B Int. Conf. on Noise and Vibration Engineering, ISMA-USD 2018
%D 2018
%T Spectral and statistical evaluation of the properties of the vibration measured at the base of an automotive seat for non-Gaussian random noise synthesis
%A Damián González
%A R. López-Valcarce
%K random vibration
%X The first step in the definition of the specifications for the vibration tests of any automotive component

is the recording and evaluation of the properties of the input vibration during field measurements. In this

paper, we explore the properties of the vibration measured at the base of the driver seat when driving

through different surfaces, with the purpose of defining a model for non-Gaussian random vibration

testing. The recorded acceleration signals are first transformed to a 6 Degree of Freedom (DOF) space,

and then segmented into stationary sections that are later analyzed. The non-Gaussian nature of the

recorded vibration is demonstrated through different Gaussianity hypothesis tests, and the distribution of

the third and fourth order moments, as well as the crest factor, are computed and analyzed. The second-order

spectral content reveals significant correlation at specific frequencies, while the significance of

third order polyspectra is checked through hypothesis testing. Based on our observations, a practical

definition of the random process is proposed based on second-order spectral content, univariate kurtosis

and crest factor levels.

%B Int. Conf. on Noise and Vibration Engineering, ISMA-USD 2018
%C Leuven, Belgium
%8 09/2018
%G eng