基于循环平稳特性的低截获概率信号的截获技术研究

【作者】 黄春琳；

【导师】 周一宇；

【作者基本信息】 国防科学技术大学 ， 信息与通信工程， 2001， 博士

【摘要】 目前低截获概率(LPI)技术发展迅速，并在雷达和通信中获得广泛应用，为对抗这些技术带来的检测和估计问题，需要研究低截获概率信号检测和估计的新方法。本文主要对两种主要低截获概率信号(相位编码信号和线性调频信号)进行了研究和探讨，基于两种信号都具有循环平稳性，将循环平稳理论应用于信号的检测和参数估计，通过深入研究提出了一些有效的检测和估计方法。 针对相位编码信号的检测，提出了基于单循环检测器的有限多循环检测器，可在没有编码序列、初相、编码序列初始时间等信息的情况下，取得理想的检测效果。基于相位编码信号循环谱的幅度和相位特性，提出了二阶循环平稳参数估计方法，可庄没有任何参数信息情况下，实现载频、码片时宽、幅度、初相、编码序列初始时间等参数的估计，估计方法计算量少，易于实现，且有良好的估计精度。 通过研究提出了两种线性调频信号的检测方法—基于循环自相关函数包络的检测方法和基于循环自相关函数的相关函数包络的检测方法，两种方法均可实现信号的盲检测，有良好的检测性能，文中还分析了加性噪声及存在乘性和加性噪声情况下两种检测方法的性能，推导出了输出信噪比公式。针对加性噪声情况及存在乘性噪声的情况，提出了线性调频信号参数的循环平稳估计方法，解决了零均值乘性噪声的信号参数估计问题，并采用一阶扰动分析方法，对两种噪声情况下的估计性能进行了分析，推导出了各参数的估计误差方差公式。 此外，通过仿真实验对各种循环平稳检测和估计方法的性能进行了评估，验证了理论分析结果的正确。更多还原

【Abstract】 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.更多还原