Research On Methods Of Detection And Parameter Estimation Of Frequency Hopping Signals And Their Performance Testing

Frequency hopping communication has strong anti-interference ability,low interception rate and multiple access networking capabilities,and is widely used in civil and military communications.At the same time,its rise has also brought severe challenges to frequency hopping communication reconnaissance.Based on different electromagnetic environments,in this paper,the blind detection method and parameter estimation methods of single and multiple frequency hopping will be studied,and the performance of the proposed algorithm will be tested.Firstly,in order to solve the problem that the detection accuracy of frequency hopping signals under low signal-to-noise ratio is greatly affected by interference and noise,a blind detection algorithm(TFC)for multi frequency hopping signals based on time-frequency analysis is proposed.Using the different time-frequency distribution characteristics of frequency hopping signal,fixedfrequency signals,and Gaussian white noise,the time-frequency diagram which is obtained by short-time Fourier transform is used to construct the time-frequency cancellation ratio.And then the time-frequency cancellation ratio of each signal is different through the theoretical analysis,so it is used as the detection statistics,and the detection thresholds are obtained through experiments.The simulation results show that the proposed algorithm has good robustness to noise power uncertainty,it has higher detection probability of frequency hopping signals and fixed frequency signals under the background of Gaussian white noise with the low signal-to-noise ratio and better frequency hopping signals detection performance under the complex electromagnetic interference environment of fixed-frequency signals,sweep signals and burst signals than the improved power spectrum cancellation method.Secondly,two parameter estimation algorithms of frequency hopping signal are proposed in response to different scene requirements.In the light of the problems of low estimation accuracy of frequency hopping signal parameters and the instability of algorithm performance under low signalto-noise ratio,a parameter estimation algorithm based on local contrast measure of time-frequency matrix with multi-scale windows(TFMLCM)is proposed.Using the mean value of the local energy comparison value of the time-frequency matrix under different scale sliding windows,the multi-scale local energy comparison feature matrix is obtained.The time-frequency matrix P that only retains the time-frequency information of the frequency hopping signal is obtained by the separated feature matrix and the adaptive threshold.And then with the time-frequency hopping information extracted from matrix P,the period,hopping time and frequency of the frequency hopping signal are accurately estimated.In accordance with the complexity of the actual electromagnetic environment,a parameters blind estimation algorithm of multi-frequency hopping signals based on multi-scale morphological filtering and time-spectrogram cancellation(MF-TFC)is proposed.Multi-scale morphological filtering is used to eliminate noise,burst signals and frequency sweep signals,and the spectrogram cancellation method is used to remove the fixed-frequency signals.Then the position information of the frequency hopping signals is obtained through the eight-connected domain mark,and the frequency hopping signals at different speeds are separated by the improved K-means clustering algorithm.The period,hopping time and frequency of the multi-frequency hopping signals are estimated according to the various cluster parameters.The simulation results show that the two algorithms proposed in this paper have better frequency hopping signal extraction effect and higher parameter estimation accuracy than similar algorithms.Finally,in order to verify the practical feasibility of the proposed algorithm,the performance test system of frequency hopping signals detection and parameter estimation algorithm is built on the DSP+FPGA hardware platform.It is programmed to realize the TFC algorithm and TFMLCM algorithm proposed.And the detection performance of the TFC algorithm and the parameter estimation performance of the TFMLCM algorithm are tested by using the frequency hopping signal sent by the signal generator SMBV100A to verify the correctness and practicability of the algorithms.