High resolution time-delay estimation based on wavelet denoising

Wavelet denoising and TLS-ESPRIT (total least squares version of estimation of signal parameters via rotational invariance technique) are applied to time delay estimation. At first, the correlation method and a frequency domain deconvolution are used to convert the time delay estimation problem to an estimation of the frequencies of complex sinusoids in white nonstationary noise. After that, wavelet singularity detection is applied to reduce the impact of the noise. TLS-ESPRIT is then applied to estimate the unknown time delays. Computer simulation results show that the derived method can work at about 5 dB lower SNR than the ordinary TLS-ESPRIT.