Research On Parameters Blind Estimation Of Frequency Hopping Signals

Frequency hopping communication has been widely used in military frequency hopping radio, because of its LPI/LPD (low probability of interception/detection) and good anti-jamming features, which increase the difficulty of FH signal detection and interception for the third communication in non-cooperative communication. Therefore, researching effective interception and blind parameter estimation algorithms for FH signals have significant military meaning.Frequency hopping signal is a typical kind of non-stationary signal, while time-frequency analysis technique is a very powerful tool for analyzing non-stationary signal, and the research of time-frequency methods for FH signal analysis has been a hot spot in communication countermeasures area. The existing methods mostly solved the problem by discovering or improving new time-frequency analysis algorithms without further study of discovering how to estimate FH signal parameters efficiently by time-frequency analysis algorithms. For another word, the influence of time-frequency algorithms parameters on the parameter estimation performance is rarely mentioned in the publications. According to different types of FH signals, the thesis analyzed the influence of the time-frequency distribution parameters on parameter estimation performance. At the same time, studied the appropriate parameter estimation algorithms for different types of FH signal.Firstly, on the basis of learning various different time-frequency analysis algorithms (STFT, WVD, SPWVD, RSPWVD, WT etc.), their respective characteristics are simulated. Among them, focused on researching the impact of different time-frequency window function length on the time-frequency aggregation performance of SPWVD, and the optimal window length based on entropy measure for SPWVD are proposed, the simulation results showed that the optimal window length has a better estimation performance than the general window length. Moreover, parameter estimation algorithms based on time-frequency analysis technique for high and low speed hop rates of the FH signal are studied in-depth respectively, and the existed parameter estimation algorithms are reproduced by MATLAB simulation.Secondly, as hop duration estimation is a key part for parameter estimation algorithms, the thesis is aimed at two kinds of different FH sequences(i.e. adjacent non-repetition FH sequence and adjacent repetition FH sequence), and focus on the study of two major categories of hop duration estimation algorithms based on the peak curve and time-frequency backbone curve respectively. By combining different time-frequency algorithms and the hop duration estimation methods, simulation results found that the combining hop duration estimation method based on STFT and time-frequency backbone curve has a good estimation performance both for the two kinds of FH sequences. In addition, according to the different complexity of the FH sequences, the influence of time-frequency analysis algorithms on the parameters estimation performance are simulated, the simulation results showed that FH signals parameter estimation algorithms based on time-frequency analysis is not affected by FH sequence complexity Finally, in order to make the parameter estimation more close to the actual FH radio signal parameters, parameter estimation algorithms under low-pass and band-pass sampling principal for narrow band FH signal are studied and proper parameter estimation strategy are given. At the same time, the thesis analyzed the impact of the sampling frequency on frequency estimation precision. The simulation results showed that not the higher sampling frequency makes better frequency estimation precision.