A method of RCS Sequence extraction based on CAC + EMD

Abstract

When the spatial object possesses the periodic motion, an obvious periodic phenomenon can be produced by its RCS sequence. Some methods such as Fast Fourier Transform Algorithm (FFT) and Circular Autocorrelation Function (CACF) can be used to judge the existence as well as the extraction of the sequence’s periodicity. Autocorrelation Function, Average Magnitude Difference Function (AMDF) and their combined transformation are essentially estimated by the correlation of their periodic signal. More than two periods of observation time is needed for accurate estimate. Furthermore, the requirement of the date rate is high and. Furthermore, such methods always accompanied by the errors of frequency multiplication and frequency division. In this paper, a periodic discriminant and estimation method combining cyclic autocorrelation and empirical mode decomposition is proposed. The periodicity of the original signal can be enhanced by the cyclic autocorrelation (CAC), and the signal component with the relatively complete periodicity can be obtained by the empirical mode decomposition (EMD). The periodic estimation of the measured data showed that the period estimated by CAC +EMD is more stable. In addition, this method is also effective in extracting the signal period with the weaker periodic phenomenon. In this paper, a periodic discriminant and estimation method combining cyclic autocorrelation (CAC) and empirical mode decomposition (EMD) is proposed (CAC+EMD), and a comprehensive comparison with the Single method is made, both CAC and CAC+EMD can extract the stable signal’s period effectively. With the gradual increase of noise discriminant effect of CAC+EMD on signal period is much better than CAC. Because CAC does not have the ability to process the non-stationary signal, so this method has a poor discriminant effect on signal period. As for the discriminant effect of CAC+EMD, there is no difference between stable signal and non-stationary signal.