An Iterative Parameters Estimator for Damped Exponential Signal in Noise

The paper proposed a hybrid-iteration algorithm to estimate the parameters (frequency and damping factor) of damped exponential signal. With regard to core idea of this method, the new algorithm utilizes the phase difference based on asymmetric window as the first implementation and interpolation Discrete Fourier Transform (IpDFT) as the second one. The estimator introduces the asymmetric phase difference method to the decaying case and avoids the error generated from wrong location of spectral line. The root mean squares error (RMSE) was investigated in the case of white Gauss noise to evaluate the capability of robustness against adaptive noise. Extensive numerical simulations show that the RMSEs of the proposed algorithm closely attach to Cramer-Rao lower bound (CRLB). In addition, simulation results also indicate that the new method has excellent performance against random noise compared with other interpolation DFT algorithms.