Ultra-short wave threat signal sensing method based on support vector data description

Abstract

The invention discloses an ultra-short wave threat signal sensing method based on support vector data description. The method mainly comprises a signal acquisition module, a time-frequency analysis module, a data dimension reduction module, a support vector data description model training module, an online threat sensing module and the like. The method mainly comprises the following steps: acquiring our signals in a space by using a radio frequency board card; performing time-frequency analysis on the baseband complex signal according to a short-time Fourier transform theory to generate a time-frequency spectrogram, and constructing a our signal time-frequency spectrogram database; converting the time-frequency spectrogram into a low-dimensional feature vector according to an image preprocessing technology and a principal component analysis theory; and training a support vector data description model to construct a minimum hypersphere capable of including all target samples, determining our signal by the samples falling into the hypersphere, and determining the samples located outside the hypersphere as potential threat signals. According to the method, the support vector data description model of the non-threat signals is constructed only by utilizing a signal database of our party, so that the problem of opening sets of the threat signals is effectively solved.

Description

technical field [0001] The present invention relates to spectrum sensing technology in cognitive radio, more precisely, the present invention mainly relates to a method for sensing ultrashort wave threat signals based on support vector data description, which belongs to the cutting-edge technology of cognitive radio spectrum sensing. Background technique [0002] Ultrashort wave communication is a communication method that uses radio waves with a frequency range of 30MHz to 300MHz to transmit information, mainly relying on ground wave propagation and space line-of-sight propagation. Ultrashort wave communication equipment usually has a radiation distance of several kilometers to more than ten kilometers. It has the characteristics of wide communication frequency band and large communication capacity. It is widely used in television, FM radio, mobile communication, radar detection, military communication and other fields. [0003] With the increasingly complex electromagnetic...

AI Technical Summary

Problems solved by technology

[0004] At present, spectrum sensing technology in electronic warfare scenarios mainly focuses on interference detection and identification, non-cooperative communication signal modulation identification, and specific signal identification. The mainstream spectrum sensing methods include energy detection, matched filtering, and cyclostationary detection. The energy detection method has limited ability to adapt to noise, the matched filter method needs to know the prior information of the detected signal, the cyclostationary detection method has high computational complexity, and these conventional methods are difficult to detect a large number of threat signals of various forms; The spectrum sensing method of deep learning, the deep neural network needs a large amount of training data to optimize the model parameters, and the threat signal data is expensive to obtain, and the threat signal is changing with each passing day. Its data set belongs to a typical open set. Building a relatively complete threat signal database has the advantages It is very difficult, so deep learning methods are currently mostly aimed at spectrum sensing of signals with specific modulation methods

Images