摘  要：針對經驗模態分解（EMD）在分解信號時存在的端點效應問題，為抑制端點效應對信號分析帶來的影響，進一步提高EMD分解準確度，提出基于基因表達式編程（GEP）的EMD改進算法。通過仿真實驗與鏡像延拓等其他3種常用端點效應抑制方法作對比，并計算評價端點效應的兩個指標，最后，通過這4種方法分別測量出原始信號的瞬時頻率以作驗證。仿真結果表明：基于GEP的EMD改進算法在分解信號后各分量兩端發散程度最小，評價指標也均優于其他3種傳統的改進方法，且更加準確地測量出原始信號的瞬時頻率。證明該改進算法能更有效地抑制EMD端點效應，具有更高的應用價值。

關鍵詞：經驗模態分解；基因表達式編程；端點效應；瞬時頻率

文獻標志碼：A       文章編號：1674-5124（2015）12-0032-04

GEP-based suppression of end effect in EMD 

YANG Bo， LI Shiping

（The Second Artillery Engineering College，Xi’an 710025，China）

Abstract: An improved algorithm of empirical mode decomposition （EMD） based on gene expression programming（GEP） was proposed in this paper to solve the end effect existing in the signal decomposition process of EMD and to restrain the strong impact brought of end effect on signal analysis to improve further the decomposition precision of EMD. After comparing it with three other commonly-used methods namely simulation and experiment and mirror extension， the two indicators of end effect were calculated and evaluated and the instantaneous frequency of original signals was measured and verified ultimately. The simulation results indicate that the divergence degree at both ends of each component is the lowest， the evaluation indicators are more superior， and the instantaneous frequency of the original signal measured is more precise compared with the three traditional improved methods. It has proven that this improved algorithm is more precise and practical in restraining the end effect.

Keywords: EMD； GEP； end effect； instantaneous frequency