A STATE ESTIMATION USING A DIGITAL-FILTER BASED ON A SYSTEM MODEL OF THE TRANSITION-PROBABILITY TYPE AND ITS APPLICATION TO THE SOUND ENVIRONMENT

The phenomena observed in the actual sound environment are inevitably contaminated by additional external noise of arbitrary distribution. F urthermore, the actual stochastic process shows various types of linea r and nonlinear time transition forms and cannot be exactly expressed only by simple models which smooth the fluctuation pattern of the time series data. In this paper, a digital filter for estimating a specifi c signal under the existence of an external noise with various forms o f probability distribution is proposed by introducing a system model o f the transition probability type. The effectiveness of the proposed t heoretical method is experimentally confirmed by applying it to the ac tual problem of estimation of the sound environment.