Abstract:

The phase difference correcting spectrum method based on all phase fast Fourier transform （FFT） calculates the conventional FFT without effects of negative frequency. As a result, there is a decline in accuracy of lowfrequency signal parameters estimation. Aiming at this problem, an improved phase difference correcting spectrum method based on all phase FFT is proposed. The effects of negative frequency to FFT are considered in classic FFT. The all phase FFT spectrum analysis theory and the phase difference correcting method are introduced. The cause that the negative frequency has a negative impact on low frequency signal spectrum analysis is analyzed. The equations of frequency, amplitude and phase estimating for single frequency signals sampled in non integer period are derived with the effects of negative frequency to spectrum analysis being considered. The simulation results show that the improvement of considering negative frequency contribution is able to improve the accuracy in low frequency cosine signal parameter estimation in the conventional way, and the antinoise ability of the original algorithm is also improved.