Detection And Modulation Classification Of Low-probability-of-intercept Communication Signals In High Frequency Channel

In recent years, the fast development of high frequency(HF) communication technology andthe application of new communication system, brought opportunities and challenges to thenon-cooperative communication. For the third party in terms of communication, its main task isto intercept the signal, azimuth measurement, recognition and ultimately get the signalcommunication content. With HF channel electromagnetic environment becoming more denseand complex, the wide application of low probability of intercept signal, in a non-cooperativereceives fast and complete detection of low probability of intercept signal and identify themodulation scheme is the basis for its subsequent processing.Therefore this paper studies the HFlow probability of intercept signal blind detection and modulation recognition, the main contentand innovation as follows:1. According to the research background of this paper, aiming at the problem of channelmodulation recognition, to study the characteristics of HF channel and multipath fading model.Aiming at the receiving condition of frequency hopping signal, the paper study the technology ofwideband digital receiver, and discusses the broadband receiver under the conditions ofshort-wave complex electromagnetic environment, the adaptive threshold detecting broadbandsignal is given, laying the foundation for the follow-up research.2. Due to the characteristics of short burst signals of real-time detection, from the timedomain algorithm of signal detection, aming at the complex of the test statistic, the Kolmogorovdetection method is proposed which is simple, aming at the detection performance, theautocorrelation detection method is proposed which have good performance, aming at theconstant false alarm threshold, the fractal box dimension method is proposed without noiseestimation. And then the threshold with constant false alarm rate of the algorithm was deduced,and finally through the simulation and the complexity of the algorithm analyse the advantagesand disadvantages of each algorithm and the suitable occasion. Then burst signal extractedevaluation index is given, and the use of wavelet decomposition reconstruction algorithm fordetecting signal load-point optimization accuracy is improved, the simulation verifies theeffectiveness of the algorithm.3. The problem on the HF frequency hopping signal blind detection under complicatedelectromagnetic environment is researched. Mathematical model of the frequency-hopping signaland technical indicators is given, and the existing method of time-frequency analysis is improved,and the time-frequency diagram processing fixed frequency elimination, binarization andmorphological filtering and cluster analysis of the processing done blind detection of frequencyhopping signal, the actual signal test proved the feasibility of the algorithm. Introduced shown to information presented for the shortcomings of the frequency graph algorithms based false alarmrate and missed alarm rate, based shows to hopping signal blind detection algorithm. Completedthrough the use of related Interferometer and boundary projection the hopping signal blinddetection, the simulation verifies the feasibility of the algorithm.4. The problems of common modulation recognition classification of the burst signal andfrequency hopping signal in HF channels and non-ideal reception conditions are researched. Firstthe cyclic autocorrelation of the single carrier and multi-carrier signals are analyzed in multipathfading channel. Through improvement of search scope, and according to the discrete spectrumline on cyclic autocorrelation section location information which accomplish the detection ofmulti-carrier signal. Aiming at the shortcomings of the existing single carrier modulationrecognition algorithm, analysis in the multipath fading channel was deduced under various singlecarrier signal first-order, second-order and fourth-order cyclic statistics feature, and by usingCAFR detection algorithm of detection, through the five layer tree judgment on common analogand digital signal recognition and classification.Simulations verify the feasibility of thealgorithm, resistance to multipath fading and robust performance under non-ideal receptionconditions.