Identification of polynomial phase signals by extended Kalman filtering

The paper addresses the problem of estimating the parameters of polynomial phase signals embedded in Gaussian noise. We consider an estimation method based on an approximate linear state space representation of the polynomial phase signal. This approach offers the opportunity to use a nonlinear but exact measurement equation and guide the estimation of the states of these signals to an extended Kalman filtering algorithm. Procedure simulations were made on linear and quadratic phase modulation signals with time-varying amplitude and are consistent with the theoretical approach. The results given by this new algorithm are compared with the performances of a standard Kalman technique.