Based On The Cyclostationary Low Probability Of Intercept Signals Interception Technology

The techniques of Low Probability of Intercept (LPT) are developing rapidly, and are used widely in Radar and Communication. In order to counteract the LPI techniques, there is necessary to research the new methods of detection and estimation of LPI signals. In the dissertation, phase-coded signals and chirp signals are studied. Because the two signals have cyclic features, the cyclostationary theory has been applied in the detection and estimation of both signals. Some novel methods of detection and estimation are proposed after intensive investigation.Based on the single cycle detector, the finite cycle detector is proposed for the detection of phase-coded signals. The finite cycle detector has a good performance under unknown the spreading code sequence, initial phase, spreading code timing epoch. Based on the amplitude and phase of two-order cyclic spectrum of phase-code signals, a novel cyclic spectrum estimation approach is proposed to estimate the carrier frequency, chip width, amplitude, initial phase, timing epoch of spreading code under no parameter information. This approach is computationally efficient, and is easy to be implemented, and it shows a good estimation accuracy over analysis.Two detection methods of chirp signals are presented in the dissertation: the cyclic autocorrelation detection method and the correlation detection method of cyclic autocorrelation function. It is shown that the two methods both have an excellent performance of blind detection. The performance analyses of the two detection methods are both carried out for additive noise or multiplicative noise, and the output signal-to-noise ratio (SNR) expressions of the two methods are also derived. The cyclic estimation approaches of chirp signals in additive nose or multiplicative noise are proposed, it presents a way for the parameters estimation of zero mean amplitude chirp signals. The estimation performances in the additive noise or multiplicative noise are analyzed by the first order perturbation analysis method, and the error variance expressions of the parameters estimation under large samples are derived.Moreover, a large amount of computer simulations have been carried out to evaluate the performance of the proposed methods. Results of simulations confirm the analysis results in this dissertation, which proves the feasibilities of those methods further.