The Study on the Acceleration Signal Processing Based on the TVAR Algorithm

Penetration acceleration signal is a kind of non-stationary random signal, the occurrence of which lasting a very short period of time, and the conventional time-frequency analysis and non-parametric modeling methods can not effectively analyze and deal with such signals. In this paper, time-varying autoregressive (TVAR) algorithm is used to achieve the parametric modeling of the non-stationary random penetration signal, through the time-varying spectrum estimation the method is verified with high resolution. Based on the simulation analysis, this paper resolves the model parameters with the application of the base function expansion. Through the study on the pole parameters phase information, the characteristics of information of the instantaneous frequency is extracted, while the acceleration signal characteristics such as layer events, void event are effectively obtained, and these works provided the algorithm basis and reference for the penetration of precise burst control technology.