Improving Empirical Mode Decomposition based on up-sampling

Abstract

The paper proposes an improved Empirical Mode Decomposition method based on up-sampling, due to the energy leakage in traditional Empirical Mode Decomposition for insufficient sampling rate(digital domain frequency greater than 0.2). The method uses the signal interpolation to improve sampling rate before EMD, and then recovers the original scale by corresponding down-sampling and low pass filtering. The numerical results show that it can partly recover the accurate position of extreme points and effectively reduce the energy leakage. Three typical interpolations are also employed and the result shows that the effect of using cubic spline interpolation with 4 times is the best relatively.