Radar Signal Carrier Frequency Measurement Based on BP Neural Network
In modern electronic warfare,various information of enemy radars is often obtained through electronic reconnaissance.The carrier frequency of radar signals are important parameters for subsequent radiation source identification,interference and anti-interference.In view of the ex-cellent multi-dimensional function representation ability of neural networks to data,a simple neural network-based fast measurement method of radar signal carrier frequency is designed.Firstly,the radar signal containing noise is sampled,the frequency of each sampled signal is annotated,and the signal sample datasets are obtained by preprocessing.Then the datasets are divided into training set and testing set,and they are input into the BP neural network for frequency fitting.Finally,the time-domain signals obtained by radio frequency sampling are input to the trained network model,and the the instantaneous frequency value of the signals are output though network.Under the con-ditions that the input signal frequency range is 0.2～2.6 GHz and the signal noise ratio is 30 dB,random tests are imposed on the network repeatedly for several times.The experimental results show that the root-mean-square error of the input signal frequency estimation is better than 5 MHz.
BP neural networkcarrier frequency measurementradar signal
NETL
NSTL
万方数据
