An electromagnetic signal spectrum data classification method based on esmd sample entropy combined with fcm

Abstract

The invention discloses a method for classifying electromagnetic signal spectrum data using ESMD sample entropy combined with FCM, which belongs to the field of electromagnetic signal processing; the details are as follows: step 1, use a spectrum analyzer to measure different electromagnetic equipment respectively, and obtain spectrum data of electromagnetic signals; step 2 1. For each kind of spectrum data of every kind of electromagnetic equipment, decompose with ESMD method respectively, obtain the intrinsic mode function component of each spectrum data; Step 3, for each kind of spectrum data, utilize the sample entropy of intrinsic mode function component as feature , constitute the eigenvector of the spectral data; step 4, input the sample entropy vectors of all spectral data into the FCM algorithm, and output the optimal membership matrix and the optimal clustering center; step 5, according to the optimal membership matrix and the optimal The clustering center analyzes the classification results of the electromagnetic signal spectrum data. The method has the advantages of: performing ESMD decomposition on the extracted electromagnetic signal spectrum data to obtain each intrinsic mode function; and having better adaptability.

Description

technical field [0001] The invention belongs to the field of electromagnetic signal processing, and relates to an electromagnetic signal spectrum data classification method combining ESMD sample entropy with FCM. Background technique [0002] The development of electronic information technology has greatly promoted the progress of science and technology, but the increasing number of electronic equipment has also caused serious electromagnetic interference problems, resulting in the decline of the normal working performance of electronic equipment in the environment. In order to reduce electromagnetic interference, electromagnetic compatibility is generally designed to find out various data in the electromagnetic signal spectrum. By classifying the electromagnetic signal spectrum data and analyzing the source of electromagnetic interference, it has a good auxiliary effect. [0003] At present, the classification of electromagnetic signals is mainly aimed at time-domain data,...

AI Technical Summary

Problems solved by technology

[0005] Among the existing classification algorithms, the K-means algorithm is simple and easy to implement, but if its initial value is not selected properly, it will be difficult to obtain ideal clustering results

The artificial neural network method can determine the optimal classification result in a short time, but it requires a sufficient number of samples, which is often difficult to meet in practice

Images