Intrusion detection method for constructing visibility graph network based on channel state information

Abstract

The invention discloses an intrusion detection method for constructing a visibility graph network based on channel state information. The intrusion detection method is divided into two stages: offlinetraining stage and the online testing stage; at the offline training stage, an abnormal value of the CSI time sequence signal is removed through Pauta method, the high-frequency noise is eliminated through a low-pass filter, and finally the smooth filtering processing is performed by using the a sliding average filter, and then 30 sub-carriers are reduced to the one-dimension time sequence signalby using the PCA dimension-reduction method, and the time sequence signal is intercepted by selecting the window size, and the time sequence signal is constructed as the network by using the visibility graph network construction method, and the related network feature is extracted as the machine learning classification feature; at the online testing stage: the intrusion detection estimation classification is performed by using different machine learning algorithms, the amplitude and the phase information are combined at this time, and the accuracy and the stability of the detection are improved by synthesizing multiple pairs of antennas. The intrusion detection method disclosed by the invention is low in cost, free from detection dead zone, and the simpler and more feasible in deploymentrealization.

Description

technical field [0001] The invention relates to the field of intrusion detection, in particular to an intrusion detection method for constructing a visual network based on wireless network channel state information (ChannelState Information, CSI). Background technique [0002] Intrusion detection technology currently has a wide range of application scenarios. For example, building an intrusion detection system can be used to monitor empty nesters and babies in delivery rooms, and implement security monitoring for some museums and banks. A large number of researchers are attracted by the above application requirements. For example, some researchers have preliminarily realized the intrusion detection system by using technical means such as video, GPS, ultrasonic, infrared, sensor network or radio frequency identification. However, the above-mentioned technical means have more or less problems in practical application, which limit the effect of practical application. For exam...

AI Technical Summary

Problems solved by technology

[0005] In order to overcome the shortcomings of the existing intrusion detection methods, such as high cost, detection blind spots, and complex deployment implementation, the present invention provides a channel state information-based construction method with lower cost, no detection blind spots, and simpler and feasible deployment implementation. The intrusion detection method of the visual graph network, which can effectively realize the classification detection of whether an intrusion occurs

Images