Research On Related Parameters Estimation Techniques Of Frequency Hopping Signals

With the advantages of anti-jamming performance, low probability of interception and networking capability, Frequency Hopping(FH) communication has been widely applied in military and civilian communications, which leads to a severe challenge to carry on receiving technology in non-cooperative mode. As a major task of FH signals processing, parameters estimation has very important realistic significance. This paper is aiming to investigate the related parameters estimation techniques of FH signals, the main contents and innovations are detailed as follows:1. The parameter estimation methods based on time- frequency backbone line are investigated. Firstly, aiming at reducing the complexity of algorithm, the fractal theory is introduced to estimate the hop duration. Making use of relations between the fractal box dimension and the frequency, the time- frequency backbone line is replaced by box dimension curve, which converts the hop duration estimation to box dimension period estimation. Simulation shows that this method is lower at complexity than the method based on short time Fourier transform(STFT), while with low performance lost. Secondly, to improve the frequency estimating precision, a method based on slide-ESPRIT is proposed. The method extracts the frequency backbone line with the ESPRIT’s advantage of high-resolution, which improves the frequency estimation accuracy greatly.2. The parameter estimation methods based on self- mixing are investigated. First, aiming at estimating the hop duration, the self- mixing signal is constructed by delayed-conjugated multiplying, then the relation between the low- frequency components, time delay and hop duration is discussed, which indicates that the hop duration could be estimated by detecting the position of inflexion on the curve. To estimate the hop timing, a method based on self- mixing by subsection is proposed. The signal is mixed through section by section. With the relation of low component, offset time and hop timing, we prove that the offset time corresponding to the maximal low frequency components signifies the hop timing information. Simulation shows that both of the proposed methods can estimate hop duration or hop timing effectively and adapt to multiple FH signals.3. The problem of tracking FH signal is investigated. Firstly, the model with hop frequency as system state is studied, and a tracking method based on particle filtering is introduced. To improve the tracking ability, an auxiliary particle filtering method based on ESPRIT is introduced. The improved method improves the tracking performance by providing reference while updating particles. Due to the defection when tracking multiple FH s ignals by particle filtering, a method based on sparse recovery is developed to tracking multiple FH signals. Making use of the FH signals’ sparsity in frequency and spatial domain, the sparsity presentation model is established. The model is solved by sparse Bayesian learning(SBL) algorithm, the tracking of multiple FH signals is achieved by frequency estimation and hop timing detection as well as DOA estimation. Simulation shows the effectiveness of the method.