A One-Stop Method for EMI Analysis Based on Wavelet Packet and SOM

The analysis of electromagnetic interferences (EMI) has been a heated problem in the field of Electromagnetic Compatibility (EMC). As the demand of efficiency and effectiveness is getting higher, the traditional methods have become the short board in analysis process. These methods havent provided a solution to analyze the relation among multiple EMI signals, and the data clustering and mining are currently done manually. To address this problem, in this paper we propose a one-stop method based on the wavelet packet decomposition (WPD) and self-organized feature map (SOM), aiming to provide a systematical and solution to extract and analyze multiple EMI signals. Experimental results are also provided to demonstrate the validity and efficiency of the proposed method.