Wavelet denoising for chirp-like signal in time-frequency domain and application to aircraft flight flutter testing

A novel denoising method for chirp-like signal that relies on the time-frequency analysis is described. Because the region of chirp signal in time-frequency (TF) map is well-concentrated relative to the widely distributed noise, the wavelet coefficients corresponding to the true signal can be extracted by a masking operation. And a new procedure is presented to design the mask and reconstruct the true signal from the filtered wavelet coefficients. Our simulation result indicates that using the method proposed in this paper can increase SNR value of chirp signal significantly. Finally, the denoising method is applied to aircraft flight flutter testing in an attempt to better estimate the nonparametric transfer function from the flight test data.