This paper deals with parameter estimation of product signals consisting of hyperbolic FM and chirp factors. Two computationally simple algorithms which decouple estimation of the chirp parameters from those of the hyperbolic FM part are presented. Both methods rely on transformations that remove either the hyperbolic or the chirp component. Therefore, at each step we are faced with the problem of estimating either a hyperbolic FM signal or a chirp signal. For the former problem, a Nonlinear Least Squares (NLS) approach is advocated. The latter is solved by the multi-lag High-Order Ambiguity Function (ml-HAF). Numerical examples show that these two methods have performance close to the Cramer-Rao Bounds.
|Original language||English (US)|
|Number of pages||4|
|Journal||International Conference on Signal Processing Proceedings, ICSP|
|State||Published - Dec 1 1998|