Research On Key Techniques Of Wideband Digital Reconnaissance Receivers And Signal Analyzing And Processing

In modern electronic warfare, the electromagnetic environment that the reconnaissance receivers confront is more complex than ever before. Wideband and digitization have been the basic requirement for reconnaissance receivers. How to design wideband digital reconnaissance receiver using reasonable and efficient methods, and how to analyze and process the intercepting signals have been the important issues in modern electronic reconnaissance domain. In this dissertation, a series of key techniques of wideband digital receivers in complex electromagnetic environment, including digital channelization technique and parameter estimation algorithms for intercepting signals, are mainly researched in detail. The main contributions of this dissertation can be summarized as follows:In Chapter 2, the dynamic channelization receiver is studied to solve the receiving problem of non-cooperative signals in electronic reconnaissance. The drawbacks of existing channelization methods are analyzed firstly, and then a new dynamic channelization receiver is proposed and its efficient implementation structure is deduced by combining parallel processing and polyphase decomposition. This structure can realize dynamic changing and non-uniformly channel division. All the channels can be processed in parallel and real time. This structure is more flexible: Even though the signal is dynamically changing, it can be synthesized according to the energy detection. This structure is computationally efficient and can be implemented in large-scale field-programmable gate array chips.In Chapter 3, the channelized receiver in the fractional Fourier domain(FRFD) is studied to solve the problem that the channelized receiver in the traditional Fourier domain can not deal with multiple wideband signals whose frequency spectra are aliasing. The theory of fractional Fourier transform(FRFT) is analyzed and the idea is pointed out that the spectrum shifting of the signal in the FRFD is equivalent to its spectrum shifting in the traditional Fourier domain. The channelized model in the FRFD with low-pass filter is proposed and in order to reduce the computational load, its efficient implementation structure is deduced based on the polyphase decomposition. By comparing with the efficient implementation structure in the traditional Fourier domain, the conclusion is drawn that the efficient architecture in the FRFD can be considered as a generalization of that in the traditional Fourier domain and the efficient architecture in the traditional Fourier domain can be regarded as a special case of that in the FRFD. Theoretical analysis and simulation results verify that by choosing an appropriate fractional Fourier transform(FRFT) order, the presented architecture can separate the broadband signals whose frequency spectra are aliasing. Thus the performance of signal detection and parameter estimation after channel output will be improved significantly.In Chapter 4, the parameter estimation algorithms for sinusoid signals and linear frequency modulation(LFM) signals are studied. We focus on solving the problems of frequency estimation for sinusoid signals and parameter estimation for multi-component LFM signals.(1) The characteristics of Rife and Jacobsen algorithms are analyzed and it is shown that the advantages and drawbacks of the two algorithms are complementary in different frequency domains. By combining the two algorithms, the synthetic approach is proposed. Simulation results indicate this approach has fine and steady frequency estimation capacity in the whole frequency range. Therefore, it is suitable for high precision, real-time and low SNR applications.(2) To solve the computational complexity problem caused by using the FRFT to estimate the parameters of LFM signals, the theory of integrated cubic phase function(ICPF) is analyzed and a new method based on ICPF and FRFT is presented. The new method only needs two one-dimension, instead of two-dimension, searching in the whole processing procedure. Thus the computational load is reduced enormously. When multi-component LFM signals with different amplitudes are present, a novel scheme is put forward. It utilizes the technique of signal separation in the FRFD to estimate the parameters of strong LFM signals and then eliminate them one by one until all of the signal components have been estimated. The scheme realizes parameter estimation of the weak LFM signals in cases involving in the strong LFM signals.In Chapter 5, the parameter estimation algorithm for polynomial-phase signals(PPSs) is studied. We concentrate on solving the problems of parameters estimation for multi-component cubic phase signals and high-order PPSs.(1) The theories and disadvantages of cubic phase function(CPF) and generalized cubic phase function(GCPF) are analyzed firstly, and then a new approach is proposed to estimate the parameters of multi-component cubic phase signals. This approach is based on the product generalized cubic phase function(PGCPF) and the product cubic phase function(PCPF), which are used to compute the cubic phase coefficient and chirp rate of the cubic phase signal, respectively. The parameters of the component with the strongest amplitude are estimated firstly. Then filtering out the strongest component, the parameters of other components in the residual signal are estimated one by one. This approach realizes parameter estimation of multi-component cubic phase signals with different amplitudes.(2) To deal with the problem of CPF-HAF method for analyzing multi-component PPSs, two improved approaches are proposed. One is based on GCPF and high-order ambiguity function(HAF) and has advantages of less calculation load and higher precision. However, it introduces higher SNR threshold due to involving in higher order nonlinearity. The other is based on Radon transform and CPF-HAF method. Compared with the former approach, it has lower SNR threshold while heavier calculation load and lower precision. In order to solving the two-dimension searching problem after Radon transform of the latter approach, a method is proposed which uses CPF-HAF method to estimate the next-highest-order phase parameter firstly, and then dechirps the intercepting signal to estimate the highest-order phase parameter. The two-dimension searching is transformed into two one-dimension searching and the computational complexity is reduced significantly.