Reconnaissance Of Frequency-Hopping Communication Based On Time-Frequency Analysis

Frequency-hopping (FH) signals are widely used in both military and commercial communication for their good properties in term of:anti-jamming, low probability of interception, good capability against fading, code division multiple access. The reconnaissance and monitoring of FH signals is thereby the research emphasis of communication countermeasures and radio management.FH signals are typical non-stationary signals for their frequency changing with time. A single time domain analysis or frequency domain analysis is helpless to them. The rise of time-frequency analysis provides an effective solution for this kind of signals. The idea of time-frequency analysis is to design a joint function of time and frequency, and use it to describe the power density on each time and frequency simultaneously. These years, the researches on time-frequency analysis are hot in signal processing.In this dissertation, firstly, according to the time variant property of FH signals, a deep investigation is made on time-frequency distribution (TFD) suitable to FH signals in complicated electromagnetic environment. Then based on these TFD, a series of studies on blind separation, blind detection and blind parameter estimation of FH signals is made. Now list the main achievements as follows:●The limitation of WVD and the existing time-frequency distributions of Cohen’ class to FH signals is firstly be analyzed. To reduce the cross-terms in WVD, a new kernel function is put forward. Compared with other kernel functions, the new kernel function is rather matched to the ambiguity function of a FH signal. On the condition that kernel parameters are selected properly, the new kernel function can suppress the cross-terms of FH signals, and preserves a higher time-frequency resolution simultaneously. To optimize the kernel parameters, a method based on the second order Renyi entropy normalized by its volume is proposed. With this method, the kernel can obtain optimal match to the ambiguity function of a FH signal, and a TFD with high time-frequency concentration and without cross-terms is available.●To reduce the cross-terms in WVD, according to the spectrum property of FH signals, a new time-frequency analysis method based on frequency decomposition is proposed. In this method, a multi-component FH signal is firstly decomposed into multiple single-component signals with a Gauss filter bank. Then the WVD of each component is sum up to form a new TFD. The theoretical analysis and simulation result shows that the new TFD can reduce the cross-term interference of FH signals, and preserves a higher time-frequency resolution simultaneously. So the new TFD can describe the time variant property of FH signals more accurately.●To resist the noise in TFDs, the TFDs of noisy signals are deduced. A threshold filtering method is proposed to reduce the noise in TFDs. This method can reduce the power of noise and improve the readability of TFDs.●To separate different FH signals, a new blind source separation (BSS) method based on TFD is proposed. This method exploits the difference in the time-frequency properties of the source signals to solve the blind source separation problem based on the joint diagonalization of a combined set of time-frequency distribution matrices. On the condition of source signals are non-correlative and have different time-frequency properties, the new method is effective on blind separation of FH signals. Due to the effect of spreading the noise power while localizing the source energy in the t-f domain, compared with those methods based on signal independence, the proposed BSS method has an increasing robustness with respect to noise.●For the single signal achieved after blind source separation, via a comparison on instantaneous frequency (IF) of different signals, a blind detection method based on IF is proposed. Since the method exploits the particular property of FH signals different from other signals, it is robust for its high detection probability and low false alarm probability in complicated electromagnetic environment.●A blind parameters estimation method for FH signals based on IF is proposed. According to the transient frequencies of FH signals, via the wavelet transform of IF, the hop period can be estimated accurately, and the switching time instants and the hopping frequencies can be estimated thereby. The method can accurately estimate the parameters of FH signals under a low SNR. It has a superior performance compared with the accepted method these days.The studies in this dissertation provide some new ideas and approaches for the reconnaissance of FH signals. Undoubtedly, it has important theoretical significance and practical value to the development of FH communication countermeasures and expanding the application of the time-frequency analysis.