Deal with boundary effect of EMD method based on improved Characteristic Wave Extending

Empirical Mode Decomposition (EMD) is a new way to process the non-stationary and nonlinear data. But the edge effect or boundary effect appears when spline interpolation is used to get two envelopes of the data. A novel method based on improved characteristic wave algorithm is proposed to prolong the data series and get extra-extremes to deal with the edge effect. Two simulated signals are decomposed by three methods for comparison: `Extending Method with Characteristic Wave´, `Extending Method with Mirror Imaging´, and `Extending Method with Improved Characteristic Wave´. Experimental results show that the proposed method restrains the edge effects effectively.