基于无线射频技术的入侵设备识别方法研究

Research on intruder device detection based on RF technologies

大多数无线设备的识别研究都是利用小波变换、傅里叶变换(Fourier transform,FT)和机器学习等方法对已知协议的单一设备进行识别,这些研究中设备种类的多样性不足,并且没有一种方法可以在多个设备同时存在时对它们进行识别。为解决上述问题,提出了一种基于自相关检验的方法,可以在多个不同协议的设备同时存在时对它们进行识别,并且无论是否已知设备的协议,都可以采用该方法。首先通过通用软件无线电设备(universal software radio peripheral,USRP)监测入侵设备的信道来采集信道信号的基带(in-phase and quadrature,IQ)数据,然后利用自相关检验的方法来提取时域信号数据的周期特征。在室内环境下,对无人机、WiFi路由器、ZigBee传感器3种设备中的两两设备同时存在时进行了30次测试,识别的成功率为100%.

Most existing researches on wireless device recognition are based on wavelet transform,Fourier transform(FT)or machine learning to identify the single device with known protocol.However,the diversity of device types in these studies is insufficient,and there is no way to identify multiple devices when they exist simultaneously.In order to solve these problems,a detection approach based on autocorrelation test is proposed,which can identify multiple devices with different protocols when they exist at the same time,and it can be used whether the protocol of the device is known or not.First,universal software radio peripheral(USRP)defined platform is used to detect the channel of invasive device and collect the in-phase and quadrature(IQ)data of the channel signal.Second,autocorrelation test method is used to extract the periodic characteristics of the time-domain signal data.In the indoor environment,30 tests are carried out on the simultaneous existence of two devices of unmanned aerial vehicle(UAV),Wifi router and ZigBee sensors,and the success rate of detection is 100%.