Parameter Estimation in Distributed Systems Using Transform Factoring

Abstract

This paper describes the application of Transform Factoring to parameter estimation in nonlinear and linear distributed systems. Transform Factoring is based on computing the Laplace transforms of experimental data for both linear and nonlinear terms in the differential equations, searching in the nonlinear parameter space for those nonlinear parameter values which yield zero phase angle of a complex-valued expression, and using these values in a functional evaluation to compute the linear parameter values. Search space dimensionality is determined only by the number of nonlinear parameters. The method is insensitive to measurement noise and distinguishes between true and spurious parameter sets.