Undergraduate Graduation Project: Time Delay Estimation Using Cross-Power Spectrum

Cross-power spectrum time delay estimation is a classical signal processing technique primarily used to measure the time difference between two signals. This method finds wide applications in acoustic localization, radar ranging, communication synchronization, and related fields.

The fundamental principle involves computing the cross-power spectrum of two signals, then performing an inverse Fourier transform to obtain the cross-correlation function. The peak position of this function corresponds to the time delay estimate. Compared to directly calculating the cross-correlation function, the frequency-domain approach significantly reduces computational complexity and allows for optimization techniques like frequency-domain weighting.

When implementing in MATLAB, special attention should be paid to sampling rate and frequency resolution settings, as these directly impact time delay estimation accuracy. The R2009b version, though older, contains sufficient signal processing toolbox functions to support algorithm implementation. For undergraduate projects, further considerations may include performance comparison under different signal-to-noise ratio conditions or combining window functions to improve estimation results. Key MATLAB functions would include cpsd for cross-power spectral density calculation and ifft for inverse Fourier transform operations.