Identification with ARMA model application to modeling of track geometry irregularity
Ying Li, W.-D. Wang, Sh.-B Wei, Shuai Yuan · 2010
Aiming at question that low identification precision of time series model system in noise, the ARMA parameters are estimated using a damped sinusoidal model representation of the autocorrelation function of the noise ARMA signal. The AR parameters are obtained directly form the estimates of the damped sinusoidal model parameters with guaranteed stability. The MA parameters are estimated using a correlation matching technique. The simulation results show that with this method only less calculation work is needed and good convergence and accuracy can be achieved in various signal-to-noise systems. This method can be successfully applied to signal modeling of track geometry irregularity. The experiment result shows model established can reflect track geometry irregularity tendency with reasonable accuracy.