Study On Detection, Parameters Estimation And Seperation Algorithm For Frequency-hopping Signals

Because of superior anti-jamming performance, low interception probability and strong capability of multiple access networking of frequency-hopping(FH) signals, FH communication has been widely used in the military and civilian field in recent years.For these reasons, FH signals reconnaissance encounters a great challenge. So, it is urgent to study FH signals interception, parameters estimation and signals separation.Taking the FH signals in the complex electromagnetic environment as the research object, the key issues in the process of FH signals processing, including blind detection of FH signals, single and multiple FH signals parameters estimation and multiple FH signals separation are investigated in the thesis. Aiming to the existing problems of the current FH communication reconnaissance, the main contents and innovations are summarized as follows:1. A wideband digital channelized receiving frame based on polyphase filter banks is adopted to realize fast and real-time processing of FH signals. In the strong noise environment, a joint algorithm based on spectrogram and noncoherent integration is proposed to improve detection and recognition performance. According to the difference of time-frequency characteristics between FH signals and other modulated signals, a recognition algorithm for FH signals based on frequency difference sequence is investigated. The overall computational complexity of the proposed FH signals recognition scheme is small, and simulations show that this scheme is still effective in more wicked noise environment.2. For single FH signal, the spectrogram and multiple differences are jointly used to estimate hop duration, time-hopping and carrier frequency. Simulation results show that the estimation variance of hop duration and time-hopping are both better than910-when signal-to-noise ratio(SNR) is above 2dB. For multiple FH signals, a new algorithm of hop duration estimation based on FH center time transform is proposed.Compared with the sequential difference histogram, the proposed one improves the estimation accuracy of FH signal parameters, especially solves the problem of degraded performance when the FH signal is lost. Simulation experiments illustrate that the hop duration estimation variance of the proposed algorithm is an order of magnitude lower than the sequential difference one.3. Regarding multiple FH signals separation, an efficient and real-time separation algorithm for asynchronous networking is studied. The FH data loss is sufficiently considered in this algorithm, and the correct rate of FH signal separation is given under different data loss rate. Simulations validate that this algorithm is suitable for fast separation of FH signals. Furthermore, to address the issues of low separation correct rate and high computational complexity of multiple FH stations, a novel method of multiple FH signals separation based on sparse Bayesian learning(SBL) is proposed.Simulation experiments verify the effectiveness and superiority of the proposed method.On the basis of these works, the complete simulations of multiple FH signals parameters estimation and separation are given, which provides practical guidance in the engineering application.