Blind Carrier Detection for Signals with Unknown Modulation

This paper proposes a blind carrier identification method for phase and/or amplitude modulated signals. The carrier frequency (fc) of the received signal with unknown phase and/or amplitude modulation is identified from the ratio of the second derivative of the signal with itself. It is shown that this ratio is equal to –(2πfc)2. The second derivative is estimated by successive differences of the highly sampled received signal. It is shown that average of the ratio of estimated second derivative to the highly sampled received signal gives an accurate estimate of the carrier frequency. Since sinusoidal chips in the received signal are symmetrical around certain points, their convolution results in correlation up to a delay. While convolution of the noise, which is not symmetric by definition, with itself does not result in a peak as in the case of correlation. Thus to reduce the additive noise effect, different segments of the received signal are convolved with themselves and these convolution outputs are cross-correlated before computing the second derivative. This results in higher signal-to-noise ratio. This technique along with a bandpass filter bank improves the performance of the second derivative based carrier detection method in the presence of noise. The proposed carrier detection method is applied to M-PSK, QAM and CDMA signals and its performance in noise for each case is presented using root mean square error simulations.