Constellation diagrams' information extraction with carrier frequency offset and phase differences

The frequency offset and phase unsynchronization between the received signal and local carrier will cause rotation and dispersion of constellation points. However, these circular or discrete ‘problem’ constellation diagram still retain some residual characteristic information for original signal. In terms of this issue, we propose a method based on density clustering algorithm to evaluate the carrier frequency offset of ‘problem’ constellation diagram. Through clustering constellation points twice and combining data fitting, the method obtains the characteristic information under different SNR and frequency offset. When SNR is more than 25dB, the ‘problem’ constellation diagram information will be well extracted. By analyzing the constellation points’ clustering results and various parameters, the signal's possible modulation type can be determined according to the number of parameter vectors and that whether the signal has carrier offset can be judged through parameter vectors. The simulation results show that the proposed method is robust and achieve good results under high SNR.