The fundamental problem of obtaining the parameters of the model (assumed) of a physical system from measured samples of its output response to a known input, is one that impacts on many different and important fields of interest. Naturally, it has a long and continuing history. The purpose of this work is to break fresh ground by utilizing a new simple deterministic theory founded squarely on well established passive network concepts. Specifically, the approach is used to achieve two main goals: 1) stable rational minimum-phase transfer functions can be identified without a priori knowledge of either numerator or denominator degrees, and 2) stable rational minimum-phase Pade-like approximations appear to be generated automatically in the nonrational case. Detailed theoretical analysis of the basic ideas and extensive numerical simulation will be carried out.
