Parametric modeling of hypersonic ballistic data based on time varying auto-regressive model

For describing target motion in hypersonic vehicle defense, a parametric analyzing and modeling method on ballistic data is proposed based on time varying auto-regressive method. Ballistic data are regarded as non-stationary random signal, where the hidden internal law is studied. Firstly, ballistic data are decomposed into smooth linear trend signal and non-stationary periodic skip signal with ensemble empirical mode decomposition method to avoid mutual interference between different modal data.Secondly, the linear trend signal and the periodic skip signal are modeled separately. The linear trend signal is approximated by power function regressive estimator and the periodic skip signal is modeled based on time varying auto-regressive method. In order to determine optimal model orders, a novel method is presented based on information theoretic criteria and the criteria of minimizing the mean absolute error. Finally, the consistency test is conducted by investigating the time-frequency spectrum characteristics and statistical properties of outputs of the parametric model established above and dynamics model under the same initial condition. Simulation results demonstrate that the parametric model established by the proposed method shares a high consistency with the original dynamics model.