Research On Detection And Parameter Estimation Of Low-Probability-of-Intercept Signals Based On Cyclostationarity

Electronic countermeasures are of great importance in the field of information warfare. Low-probability-of-intercept (LPI) techniques are widely utilized in electronic countermeasure systems. Therefore, researches on effective interception and parameter estimation methods of LPI signals have great practical significance and remarkable application value. This dissertation mainly deals with the detection and parameter estimation of phase-coded signals and linear frequency modulated (LFM) signals, which are typical waveforms applied in LPI radar and communication systems, under the framework of cyclostationarity. The main contents are as follows:Chapter 1 introduces the research background and significance of our work firstly. Then a brief review of the advances and the state-of-the-art of LPI signal detection and parameter estimation techniques is presented and the evolution of the cyclostationary theory is summarized.In Chapter 2, the fundamentals of the cyclostationary theory, including the basic concepts of cyclostationarity and the estimation of cyclic statistics, are briefly reviewed. The links as well as the distinctions between cyclostaitionary approaches and time-frequency representations are provided. Finally, the advantages of cyclostaitionary methods in noise and interference suppression are explicated.The cyclostationarity of phase-coded signals is analyzed in Chapter 3, then the explicit formula of the cyclic autocorrelation functions of binary phase-coded signals, as well as those of quaternary phase-coded signals are derived. Single cycle detectors are presented in the form of cyclic autocorrelation function, and their detection performance is analyzed employing the deflection theory. A cyclic feature based detection scheme for phase-coded signals is proposed for non-cooperative cases, which could perform blind detection under rather low signal-to-noise ratios.Based on the cyclic autocorrelation functions derived in the last chapter, the influence of lags on the cyclic autocorrelation is analyzed in Chapter 4. A new cyclic autocorrelation based phase-coded signal blind parameter estimation method is proposed, and the effect of stationary noise on the estimation performance is also provided. The method avoids multi-dimensional searching and has raised the computational efficiency.In Chapter 5, the cyclic autocorrelation envelope based detection method for LFM signals is studied in detail and its performance is evaluated. A cyclic autocorrelation based constant false alarm ratio detection scheme is developed by extending the asymptotically optimal chi-squared test for cyclostationarity, which is proposed by Dandawate et al, to LFM signal detection. Both methods could perform blind detection of LFM signals.The cyclostationary LFM signal phase parameter estimation method is presented in Chapter 6 and its performance is analyzed with the derived approximate mean-square-error analytical expressions of the parameters to be estimated. In order to overcome the drawbacks of the conventional cyclostationary method, an iterative estimation algorithm based on cyclostationarity is proposed. By choosing the lag values for iteration different from those for initialization, the proposed algorithm achieves increased estimation accuracy, reduced error propagation effect and operation over a wider range of phase parameter values. The new iterative algorithm has better performance as compared with the conventional cyclostationary estimation method and is very close to the Cramer-Rao lower Bounds.A summary of the whole dissertation is given in Chapter 7 and a prospect of the issues concerned in this field is also made from the author's research perspective.