A new parameter estimation method for damped real-value sinusoidal signals in noise

To reduce the spectrum leakage influence of the negative frequency component of damped real-value sinusoidal signals, a high precision parameter estimation method is proposed in this paper. The fast Fourier transform algorithm is adopted to improve the real-time performance, and the negative frequency influence is suppressed by a spectrum subtract strategy based on an interpolation method. Then, the accurate frequency, damping factor, initial amplitude and phase of the sampled signals are obtained by an iterative procedure. The theoretical bias and variance are analyzed, and the analytic results indicate that the proposed method is an asymptotic unbiased estimation method. The simulation testing is performed, and the numerical results demonstrate that the proposed method suppresses the spectrum leakage influence of the negative frequency component and improves the estimation performance. Moreover, the efficacy of the proposed method is validated by experimental testing on a Coriolis mass flowmeter.