Bootstrap analysis of polynomial amplitude and phase signals

In many areas where it is desirable to be able to estimate the form and parameters of an observed AM/FM signal existing procedures require that assumptions must be made of the distribution of the noise or the form of the signal. This paper presents procedures for estimating both the model order and confidence limits for the coefficients of signals whose amplitude and phase are polynomials. The procedures are formulated using minimal assumptions. Simulations run using the presented procedures show that they quite accurately estimate the parameters when the signal to noise ratio is high.