A modulation based phase difference estimator for real sinusoids to compensate for incoherent sampling

Phase difference estimation is a fundamental problem in numerous applications. However, incoherent sampling (IS) is an inevitable factor which degrades the precision of many correlations or Fourier transform-based approaches. In this paper, IS and the spectral superposition of real signals are both considered. A novel estimator is developed based on modulation and the discrete Fourier transform (DFT). With the estimated frequency, the phase difference can be obtained by calculating four DFT samples of the modulated signals. Simulations and the experimental results have proved their validity as well as their superiority over five other methods designed for IS, particularly at high signal-to-noise ratios. Furthermore, the proposed method can maintain high accuracy even when a significant bias occurs with the frequency estimation.