A Support Vector Machine Recognition Optimization Method Based on Multi-resolution Analysis

Abstract

The invention discloses a support vector machine recognition and optimization method based on multi-resolution analysis. It includes a step S6 of performing multi-resolution analysis on the waveform in the radio frequency fingerprint identification technology. The step S6 is to perform multi-resolution analysis on the coherent accumulation denoising signal obtained in step A. The present invention can improve the accuracy of radio frequency fingerprint identification by using support vector machine, reduce the computational complexity of support vector machine learning with a large number of sample signals, adopt the method of multi-resolution analysis to extract the feature points in signal samples, and ensure the accuracy of radio frequency fingerprint On the premise of features, the number of sample points for support vector machine learning is reduced, thereby reducing the computational complexity.

Description

technical field [0001] The invention relates to the field of physical layer access authentication of wireless equipment, in particular to a support vector machine identification and optimization method based on multi-resolution analysis. Background technique [0002] Feature extraction comes from image processing. In signal detection and processing, it is a method of analyzing and transforming the preprocessed signal to highlight the representative (with obvious physical or statistical significance) features of the signal and Extract saved method. The purpose of feature extraction is to reduce data storage and reduce input data redundancy, so as to facilitate the discovery of more meaningful latent variables, and help in-depth understanding and analysis of a large number of similar signals. RF fingerprint feature extraction is a key step in RF fingerprint identification. For the received wireless signal, the receiver must preprocess the signal, such as some necessary detect...

AI Technical Summary

Problems solved by technology

[0003] In the radio frequency fingerprint feature extraction technology, although the use of support vector machines can improve the accuracy of radio frequency fingerprint recognition, but the demand for the number of sample points for support vector machine learning is very large, resulting in high computational complexity and high computational resource overhead.

Images