K. Joo et T. Bose, IDENTIFICATION OF A CLASS OF SIGNALS IMBEDDED IN HIGH MEASUREMENT NOISE, Journal of the Franklin Institute, 330(6), 1993, pp. 995-1003
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 equa
tion is then linearized and formulated as a Kalman filtering state est
imation problem. Computer simulations yield excellent results for a va
riety of examples, a couple of which are presented in this paper.