A dynamical identification method based on conditioned observations for a sound insulation system with the order determined by a new criterion index

This paper describes a new trial on the identification method for an arbitrary sound insulation system with arbitrary random excitations, in a practical case when the transmitted signal is strongly contaminated by an external additive noise. Firstly, the actual sound insulation system is expressed by using a time series model in a power scale rather than using the usual spectral density in a frequency domain. Here upon, based on a pre-experiment with a test input of white noise, the order of this time series model is determined. Secondly, its system parameters are successively estimated by a wide sense digital filter based on the Bayes´ theorem conditioned input level. Thirdly, by use of those estimated parameters, the probability distribution of output power fluctuation is predicted for the present system with arbitrary random excitations. Finally, the effectiveness of the proposed identification method is experimentally confirmed by applying it to many actual sound insulation systems.