Vibration signal analysis of automobiles based on nonstationary spectrum

This paper describes an analysis method and its application to prognosis and diagnosis of automobiles based on nonstationary spectral estimation. An instantaneous frequency spectrum is considered first at a single point of time based on the instantaneous representation of autocorrelation. A kinetic model is introduced into spectral time-variations. The instantaneous maximum entropy method (IMEM) is used for the nonstationary spectral estimation of vibration and acoustic sound signals of automobiles. A two-dimensional (2-D) space is composed based on the variables which are obtained by the IMEM. Position, velocity, and acceleration vectors are calculated for state points on the above 2-D space. A new space is composed by axes of position and acceleration vectors. The proposed method was confirmed useful to implement automatic supervising for automobile engine operations.<
>