Abstract: The cyclic autocorrelation function for PSK signals is derived, a cyclic autocorrelation based statistic is established, and by employing such a statistic a new blind parameter estimation method for PSK signals is proposed. This cyclic autocorrelation based method is computationally efficient and easy to realize due to the utilization of FFT (Fast Fourier Transform) in calculating the one-dimensional slices of cyclic autocorrelation function. Moreover, this estimation approach avoids multi-dimensional search and it applies to low signal-to-noise cases. Simulation experiments are provided for evaluating the parameter estimation performance, and experimental results have proved the validity of the method.

Key words: PSK signal, cyclic autocorrelation, cyclostationary, parameter estimation