Source Number Enumeration Approach Based on CEEMD

Empirical mode decomposition (EMD) can be used to decompose complex signals into a limited number of intrinsic mode functions (IMFs), and each decomposed IMF component contains local characteristic signals of different time scales of the original signal. Thus, EMD can be used to transform the problem of source number enumeration into a pattern recognition problem, thereby improving the accuracy of source number enumeration. However, EMD is prone to mode mixing, which also affects the accuracy of source number enumeration. To overcome this issue, this research work proposes the replacement of EMD decomposition with the Complementary Ensemble Empirical Mode Decomposition (CEEMD) to suppress the mode mixing phenomenon. The result of the experimental simulation in this proved that CEEMD achieved a better accuracy of source number enumerations compared to three other approaches namely based on EMD, SORTE and GED.