Identification of a class of signals imbedded in high measurement noise

The Kalman filtering algorithm is used to identify a class of signals imbedded in high amplitude measurement noise. The considered class of signals is first modeled empirically as a nonlinear equation. The equation is then linearized and formulated as a Kalman filtering state estimation problem. Computer simulations yield excellent results for a variety of examples, a couple of which are presented in this paper.