多分量雷达辐射源信号模型和检测估计算法研究
|
摘要
雷达辐射源信号检测和参数估计是电子情报侦察系统中的关键信号处理过程,也是电子对抗信号处理研究中的热点和难点。在现代电子战环境中,信号密度越来越大,雷达信号也越来越多地采用复杂调制方式。密集复杂的信号环境会导致脉冲同时或相继到达接收机并重叠或交叠在一起,形成多分量雷达辐射源信号。多分量雷达辐射源信号的存在使依赖单脉冲检测的瞬时频率估计方法难以有效执行,易导致瞬时频率估计错误或关键信息缺失,破坏同一辐射源脉冲列的各脉冲间的统计特性,降低辐射源识别率,甚至导致识别错误。在脉冲数极为有限的情况下,该处理方式会严重影响雷达辐射源情报信息的获取。随着电子战环境中的脉冲信号密度越来越大,脉冲发生交叠的概率也越来越大,对交叠脉冲信号形成的多分量雷达辐射源信号进行检测,估计各信号分量的参数已经成为现代电子侦察系统不可避免而又必须解决的问题,但目前尚缺乏对多分量雷达辐射源信号及其检测估计算法的系统研究。因此,本论文对多分量雷达辐射源信号及其检测估计算法中存在的问题和难点进行了探索性、系统性的深入研究,以实现密集信号环境中复杂雷达辐射源的情报获取。
本论文先在分析多分量雷达辐射源信号来源、特性、表现形式和交叠概率的基础上,界定多分量雷达辐射源信号的含义,建立多分量雷达辐射源信号模型；然后,从时频分析方法和参数估计两方面设计多分量雷达辐射源信号检测估计算法,以获得各信号分量的时域、频域、调制规律等有用信息,在算法研究中考虑了不同信号类型、时频交叠严重及强噪声干扰情况下的多分量雷达辐射源信号检测,在采用基于时频分析方法的多分量雷达辐射源信号检测估计算法中,分别讨论S方法(SM)检测估计算法、重排sM检测估计算法和SHE(SM、Hough变换和时频面消除法的结合算法)检测估计算法；随后,参数检测估计算法被用来处理由能量强弱相差较大的信号分量形成的多分量信号和含高阶次相位信号分量的多分量雷达辐射源信号。最后,从算法实现过程分析、计算复杂度和检测估计性能三个方面对不同检测估计算法进行对比分析,并给出结论。在对雷达辐射源信号及其检测估计算法进行系统研究中,获得的主要研究成果如下：
(1)界定多分量雷达辐射源信号含义,建立多分量雷达辐射源信号模型。目前关于多分量雷达辐射源信号的研究是电子对抗信号处理中的新问题和研究难点,多分量雷达辐射源信号的含义和模型是研究检测估计算法的基础。论文从多分量雷达辐射源信号的来源、特性、表现形式和检测必要性出发,对多分量雷达辐射源信号进行研究。在此基础上,针对雷达辐射源信号的特点,对多分量雷达辐射源信号的含义进行界定,建立多分量雷达辐射源信号数学模型。从信息获取和需要解决的问题两方面讨论多分量雷达辐射源信号检测的特点,为检测估计算法的研究提供重要参考。
(2)深入分析SM检测机理,提出多分量雷达辐射源信号SM检测估计算法。从交叉项去除、噪声性能评估和检测性能三方面研究了SM检测机理,得到SM在检测多分量信号时的优势及存在的问题。利用SM在检测多分量信号时的特点,设计多分量雷达辐射源信号SM检测估计算法。仿真实验表明,SM检测估计算法具有伪Wigner-Ville分布的高时频分辨力,又避免了交叉项,可以准确估计出各辐射源脉冲信号分量的瞬时频率和脉冲起止点,对噪声信号有很强的处理能力；估计值围绕理想值上下波动,但波动范围很小。在检测含噪声信号(信噪比为0dB)时,虽然估计值会在局部偏离理想值,但不影响估计信号与理想信号的逼近,仍能保持高的估计精度。
(3)提出多分量雷达辐射源信号重排SM(RSM)检测估计算法,并从算法过程分析讨论了RSM检测估计算法的可行性和执行效率。针对多分量雷达辐射源信号特点,将时频重排用来提高信号SM时频面的时频聚集性,提高SM在避免交叉项干扰过程中因采用频窗而降低的时频聚集性,使信号更容易被检测出来,并获得更高的估计精度。对算法实现流程和实现的分析表明,检测估计算法的主要计算过程是窗内信号的FFT,比较简单且容易通过硬件实现。仿真实验结果表明,RSM检测估计算法采用SM消除信号分量间的交叉项,通过重排提高时频分辨率,既能获得与Wigner-Ville分布接近的高时频分辨率,又能避免Wigner-Ville分布所固有的交叉项,可有效处理线性和非线性调频多分量雷达辐射源信号。
(4)提出多分量雷达辐射源信号SHE检测估计算法,给出SHE检测估计算法模型和引入以实现多个信号分量依次检测为目的的时频面消除法。SHE检测估计算法将SM、Hough变换和提出的时频面消除法有机结合,在计算信号SM时频分布的基础上,通过在Hough参数空间中对线性调频信号和相位编码信号的累积,提高处理低信噪比信号的能力；通过提出的时频面消除法实现多个信号分量的依次检测,避免不同信号分量间的干扰和信号定位中对Hough变换进行多值检测；通过脉冲检测阈值来估计信号分量的起止时间,通过相位判别阈值来估计相位信号的相位。仿真实验表明,SHE检测估计算法能准确估计多分量辐射源信号中各信号分量持续时间和瞬时频率；SHE方法在检测低信噪比信号(信噪比为-5dB)时仍能获得高的估计精度；通过对时频面提取的信号作相位判别,可以获得信号分量的相位编码信息。
(5)提出多分量雷达辐射源信号PHAF(Product High Ambiguity Function)检测估计算法,在检测中采用剥洋葱策略,使信号分量按照能量强弱逐次进行检测,并能有效检测多分量雷达辐射源信号中具有高阶相位的信号分量和能量较弱的信号分量。检测估计算法通过将多个时间延迟的HAF相乘的方法,强化信号分量形成峰值,削弱没有相关性的交叉项和噪声,避免HAF在处理多分量信号时易受噪声和交叉项形成的杂波干扰,同时避免采用最大似然估计方法时的高计算复杂度。在检测过程中采用剥洋葱策略,使能量较弱的信号分量因避免了强能量信号分量的影响及交叉项的干扰而被检测出来。仿真实验结果表明,PHAF检测估计算法能有效检测混合阶相位多分量雷达辐射源信号,对多个信号分量的检测顺序跟信号分量的能量(幅度)有关,能量大的信号分量先被检测出来,但并不影响其它信号分量的检测；不同相位阶次信号分量的检测顺序与阶次高低没有影响,参数估计精确程度与信号能量有关,能量越大的信号估计精度越高；在对含噪声信号(信噪比为OdB)进行检测时,估计误差只有小幅提高,估计精度没有显著降低。
(6)比较研究多分量雷达辐射源信号的多种时频检测估计算法和参数检测估计算法的性能,分析各检测估计算法的优缺点和适用范围,概括总结检测估计算法的性能指标,并在算法实现过程分析中提出了基于SM的快速时频处理实现方法。检测估计算法的性能分别从算法实现过程、计算复杂度和检测估计性能三方面进行比较分析。采用提出的SM快速时频方法降低时频检测估计算法的计算复杂度。检测估计算法的计算复杂度分析表明,时频检测估计算法的复杂度要高于参数检测估计算法的复杂度；在时频检测估计算法中,SHE检测估计算法的复杂度最高,RSM检测估计算法的复杂度略高于SM检测估计算法。检测估计性能是衡量检测算法有效性的指标,其研究主要考虑估计精度、噪声抑制能力和算法的适用范围。时频检测估计算法的检测性能分析以时频分布的性能为基础,参数检测估计算法的检测性能通过计算估计值的方差和克拉美-罗界获得。检测估计性能分析表明,RSM具有比SM更高的时频分辨率,继承了SM在交叉项和噪声抑制方面的优势,是比较理想的时频分布；SHE检测估计算法具有最强的检测低信噪比信号的能力；PHAF检测估计算法具有最好的高阶相位调制信号检测能力,可实现混合相位阶次多分量雷达辐射源信号的检测和弱能量信号分量检测。在总结检测估计算法性能的基础上,给出各检测估计算法性能指标比较表,以作为不同情况下检测估计算法选择的参考,并可作为多分量雷达辐射源信号检测估计算法的进一步讨论和应用研究的基础。
本论文研究工作得到国家自然科学基金项目(60971103、60702026、60572143)、四川省杰出青年基金项目(09ZQ026-040)和西南交通大学博士生创新基金项目资助。
Radar emitter signal detection and parameter estimation is a key process of signal processing in electronic intelligence systems and also a focus and difficulty in the signal processing of electronic countermeasure. In modern electronic warfare, the density of signals becomes denser and denser, and radar signals use more and more complex modulations. The dense and complex signal environment produces the situation that radar signal pulses reach receivers coincidently or successively and overlap each other, which forms multi-component radar emitter signals. The existence of multi-component radar emitter signals deteriorates the instantaneous frequency estimation method that uses a single pulse to perform detection, which easily results in estimation errors of instantaneous frequencies or the absence of important information, and therefore destroys statistic characteristics of pulses in a pulse train from one emitter and lowers correct recognition rates, even wrong recognition, of radar emitters. In the case of very limited number of pulses the method greatly affects the intelligence information gain of radar emitters. As the density of pulses grows denser and denser in the electronic warfare environment, the probability that pulses overlap is bigger and bigger. So the detection and estimation of multi-component radar emitter signals caused by overlapped pulses becomes an unavoidable and critical problem in modern electronic reconnaissance systems. Until now there is not a systematic study on detection-estimation algorithms for multi-component radar emitter signals. Therefore, this dissertation carries out a systematic, innovative and deep investigation on the problems of multi-component radar emitter signals and their detection-estimation algorithms to obtain the intelligence information of advanced radar emitters in a dense and complicated signal environment.
This dissertation first discusses the meaning and model of multi-component radar emitter signals, on the basis of the analysis of source, characteristics and representation ways of multi-component radar emitter signals, and the present probability of overlapped pulses. Then, time-frequency analysis approaches and parameter estimation methods are used to design several detection-estimation algorithms for multi-component radar emitter signals to gain various useful information of signal components in time and frequency domains and modulation types. The detection-estimation algorithms consider various types of signals and the detection-estimation of multi-component radar emitter signals in the case that there are many overlaps and noise. The detection-estimation algorithms based on time-frequency analysis consist of S-method (SM) detection-estimation algorithm, reassigned SM detection-estimation algorithm and SHE (a hybrid approach of SM, Hough transformation and elimination method in time-frequency plane) detection-estimation algorithm. The detection-estimation algorithms based on parameter estimation is devised to detect the multi-component radar emitter signals composed of signal components with greatly different energies or containing high-order phase modulation signal components. Finally, these detection-estimation algorithms are comparatively analyzed from the three aspects of the algorithm implementation processes, computational complexity and detection-estimation performances, and conclusions are drawn. Main research fruits achieved in the systematic study of multi-component radar emitter signals and their detection-estimation algorithms are as follows:
(1) The definition of multi-component radar emitter signals is given and the model of multi-component radar emitter signals is constructed. The investigation of multi-component radar emitter signals is a new and difficult problem in the signal processing of modern electronic countermeasure. The meaning and model of multi-component radar emitter signals is the basis of detection-estimation algorithms. Multi-component radar emitter signals are analyzed from their source, characteristics, representation ways and detection necessity. On the basis of analyzing radar emitter signals, the definition of a multi-component radar emitter signal is provided and the mathematical model of a multi-component radar emitter signal is given. Then, the features of multi-component radar emitter signal detection are discussed from the aspect of information collection and problems to solve, which can provide an important reference for detection-estimation algorithm design.
(2) The detection mechanism of SM is deeply analyzed and a SM-based algorithm for detecting multi-component radar emitter signals is proposed. The detection mechanism of SM is studied by analyzing cross-term elimination, noise suppression and detection performance to gain the advantages and disadvantages of SM in the process of detecting multi-component signals. Making use of the characteristics of SM in detecting multi-component signals, a multi-component radar emitter signal detection-estimation algorithm is designed. Simulation experiments show that the SM detection-estimation algorithm has high time-frequency resolution, like pseudo Wigner-Ville distribution, and avoids the interference of cross terms, and that the algorithm is able to accurately detect the instantaneous frequency, starting and ending instants of each signal component and has strong detection ability to process noised signals. The detected values fluctuate around their ideal values, but the fluctuation ranges are small. When the algorithm detects noised signals (signal-to-noise ratio (SNR) is 0 dB), the detected values may violate their ideal values, but this does not affect the approximation of the detected values to their ideal values, thus the algorithm still obtains high estimation precision.
(3) A multi-component radar emitter signal detection-estimation algorithm based on reassigned SM (RSM) is presented and its feasibility and implementation efficiency are discussed through algorithm process analysis and hardware realization. Aiming at the characteristics of multi-component radar emitter signals, time-frequency reassign is applied to improve the time-frequency concentration which is decreased by adopting frequency-domain window in the process of avoiding cross-terms of SM. Using the time-frequency reassign, the signal components in the SM distribution plane are detected more easily, and better estimation precision can be obtained. The analysis of algorithm implementation process and hardware realization shows that the main computing process of this detection-estimation algorithm is Fourier transform of the signal segment within a time-domain window and the detection-estimation algorithm is relatively simple and easy to realize for hardware. Results of simulation experiments show that RSM detection-estimation algorithm uses SM to remove cross-terms between signal components and applies time-frequency reassign to enhance time-frequency resolution. Thus, RSM detection-estimation algorithm can obtain high time-frequency resolution close to Wigner-Ville distribution and avoid the disturbance of cross-terms that Wigner-Ville owns. Also RSM detection-estimation algorithm can effectively detect both linear and non-linear frequency-modulated multi-component radar emitter signals.
(4) A detection-estimation algorithm, SHE, is proposed to detect multi-component radar emitter signals. The model of SHE detection-estimation algorithm is given. An elimination method performed in the time-frequency plane is introduced to detect signal components one by one. SHE appropriately combines SM, Hough transform with time-frequency elimination method. On the basis of the computation of SM time-frequency distribution of a signal, the noise suppression ability is improved by accumulating the linear frequency modulation signals or phase encoding signals in the parameter space of Hough transform. The introduced time-frequency elimination method is employed to detect signal components one by one, which can avoid the disturbance between signal components and multi-valued detection in the signal location of Hough transform. The starting and ending time of signal components is obtained by using a pulse detection threshold. The phase codes can be achieved by using a phase detection threshold. Simulation experiments show that SHE is able to accurately detect the instantaneous frequency and the duration time of each signal component, and that SHE can still achieve high estimation precision for the signals with low SNR (SNR is-5dB). Also the experiments show that the phase coding information of signal components can be gained by using phase elimination approach in the time-frequency plane.
(5) A multi-component radar emitter signal detection-estimation algorithm based on Product High Ambiguity Function (PHAF) is proposed. In this algorithm, a novel strategy like peeling onion is presented to detect signal components sequentially according to their energy magnitudes. The algorithm can effectively detect the multi-component radar emitter signals containing high-order phase modulation signal components or weak energy signal components. Through multiplying several high ambiguity functions with different time-delays, the proposed detection-estimation algorithm can strengthen the useful signal components and weaken noise and irrelative cross-terms to avoid the case that High Ambiguity function (HAF) is easily affected by noise and clutter waves formed by cross-terms when HAF is employed to process multi-component signals. Also, the detection-estimation algorithm reduces high computing complexity due to the use of a maximum likelihood estimation method. The onion-peeling strategy in the detection-estimation algorithm is helpful to detect the signal components with weaker energy because it eliminates the side-effects of signal components with stronger energy and cross-terms. Results of simulation experiments show that the presented algorithm can detect the multi-component radar emitter signals with various order phase modulations effectively. The detection order of signal components depends on the energy (amplitude) of signal component, so the signal components with stronger energy are first detected and therefore they do not affect the detection of other signal components. The detection order of the signal components with various orders of phase is not relative to their orders of phase. The estimation precision of parameters relates the energy of signal components. The stronger the signal components are, the higher the estimation precision is. When noised signals (SNR is OdB) are detected, the estimation error increases only in a small scale and there is not significant decrease for estimation precision.
(6) The performances of several multi-component radar emitter signal detection-estimation algorithms based on time-frequency distributions and parameter estimation are comparatively investigated. The application scopes, advantages and disadvantages of the detection-estimation algorithms are analyzed. Performance indexes for the detection-estimation algorithms are summarized. In the process of analyzing the implementation procedure of algorithms, a fast time-frequency analysis method based on SM is proposed to deal with multi-component radar emitter signals. The performances of detection-estimation algorithms are analyzed through algorithm implementation process, computing complexity and detection capability. The proposed fast time-frequency analysis method is used to decrease the computing complexity of time-frequency detection-estimation algorithms. The analysis of computing complexity of detection-estimation algorithms shows that the parameter estimation detection algorithm is better than the time-frequency detection-estimation algorithms. Among the time-frequency detection-estimation algorithms, the computing complexity of SHE is the highest and RSM detection-estimation algorithm is slightly worse than SM detection-estimation algorithm. The detection-estimation performance, mainly considering detection precision, noise suppression and application scope, is an index to evaluate the effectiveness of detection-estimation algorithms. The performance analysis of time-frequency detection-estimation algorithms is based on time-frequency distributions. The detection-estimation performance of parameter estimation detection algorithm is analyzed through computing the variance of estimation values and Cramer-Rao Bound. The detection-estimation performance analysis shows that RSM has higher time-frequency resolution than SM because RSM inherits the good points of SM about noise suppression and cross-terms elimination, so RSM is an ideal time-frequency distribution. The analysis also shows that SHE has better detection ability than other algorithms in processing the signals with low SNR, and that PHAF detection-estimation algorithm has the best detection ability in processing the signals with high order phase modulations and can detect multi-component radar emitter signals with mixed order phase modulations or weaker energy signal components. After summarizing the performance of detection-estimation algorithms, a table containing the performance indexes for comparing detection-estimation algorithms is provided to choose the detection-estimation algorithm for different situations and also to be a basis of further discussion and application research of multi-component radar emitter signals.
This work is supported by the National Natural Science Foundation of China (60971103、60702026、60572143), Scientific and Technological Funds for Young Scientists of Sichuan (09ZQ026-040), and the Doctorial Innovation Foundation of Southwest Jiaotong University.
本论文先在分析多分量雷达辐射源信号来源、特性、表现形式和交叠概率的基础上,界定多分量雷达辐射源信号的含义,建立多分量雷达辐射源信号模型；然后,从时频分析方法和参数估计两方面设计多分量雷达辐射源信号检测估计算法,以获得各信号分量的时域、频域、调制规律等有用信息,在算法研究中考虑了不同信号类型、时频交叠严重及强噪声干扰情况下的多分量雷达辐射源信号检测,在采用基于时频分析方法的多分量雷达辐射源信号检测估计算法中,分别讨论S方法(SM)检测估计算法、重排sM检测估计算法和SHE(SM、Hough变换和时频面消除法的结合算法)检测估计算法；随后,参数检测估计算法被用来处理由能量强弱相差较大的信号分量形成的多分量信号和含高阶次相位信号分量的多分量雷达辐射源信号。最后,从算法实现过程分析、计算复杂度和检测估计性能三个方面对不同检测估计算法进行对比分析,并给出结论。在对雷达辐射源信号及其检测估计算法进行系统研究中,获得的主要研究成果如下：
(1)界定多分量雷达辐射源信号含义,建立多分量雷达辐射源信号模型。目前关于多分量雷达辐射源信号的研究是电子对抗信号处理中的新问题和研究难点,多分量雷达辐射源信号的含义和模型是研究检测估计算法的基础。论文从多分量雷达辐射源信号的来源、特性、表现形式和检测必要性出发,对多分量雷达辐射源信号进行研究。在此基础上,针对雷达辐射源信号的特点,对多分量雷达辐射源信号的含义进行界定,建立多分量雷达辐射源信号数学模型。从信息获取和需要解决的问题两方面讨论多分量雷达辐射源信号检测的特点,为检测估计算法的研究提供重要参考。
(2)深入分析SM检测机理,提出多分量雷达辐射源信号SM检测估计算法。从交叉项去除、噪声性能评估和检测性能三方面研究了SM检测机理,得到SM在检测多分量信号时的优势及存在的问题。利用SM在检测多分量信号时的特点,设计多分量雷达辐射源信号SM检测估计算法。仿真实验表明,SM检测估计算法具有伪Wigner-Ville分布的高时频分辨力,又避免了交叉项,可以准确估计出各辐射源脉冲信号分量的瞬时频率和脉冲起止点,对噪声信号有很强的处理能力；估计值围绕理想值上下波动,但波动范围很小。在检测含噪声信号(信噪比为0dB)时,虽然估计值会在局部偏离理想值,但不影响估计信号与理想信号的逼近,仍能保持高的估计精度。
(3)提出多分量雷达辐射源信号重排SM(RSM)检测估计算法,并从算法过程分析讨论了RSM检测估计算法的可行性和执行效率。针对多分量雷达辐射源信号特点,将时频重排用来提高信号SM时频面的时频聚集性,提高SM在避免交叉项干扰过程中因采用频窗而降低的时频聚集性,使信号更容易被检测出来,并获得更高的估计精度。对算法实现流程和实现的分析表明,检测估计算法的主要计算过程是窗内信号的FFT,比较简单且容易通过硬件实现。仿真实验结果表明,RSM检测估计算法采用SM消除信号分量间的交叉项,通过重排提高时频分辨率,既能获得与Wigner-Ville分布接近的高时频分辨率,又能避免Wigner-Ville分布所固有的交叉项,可有效处理线性和非线性调频多分量雷达辐射源信号。
(4)提出多分量雷达辐射源信号SHE检测估计算法,给出SHE检测估计算法模型和引入以实现多个信号分量依次检测为目的的时频面消除法。SHE检测估计算法将SM、Hough变换和提出的时频面消除法有机结合,在计算信号SM时频分布的基础上,通过在Hough参数空间中对线性调频信号和相位编码信号的累积,提高处理低信噪比信号的能力；通过提出的时频面消除法实现多个信号分量的依次检测,避免不同信号分量间的干扰和信号定位中对Hough变换进行多值检测；通过脉冲检测阈值来估计信号分量的起止时间,通过相位判别阈值来估计相位信号的相位。仿真实验表明,SHE检测估计算法能准确估计多分量辐射源信号中各信号分量持续时间和瞬时频率；SHE方法在检测低信噪比信号(信噪比为-5dB)时仍能获得高的估计精度；通过对时频面提取的信号作相位判别,可以获得信号分量的相位编码信息。
(5)提出多分量雷达辐射源信号PHAF(Product High Ambiguity Function)检测估计算法,在检测中采用剥洋葱策略,使信号分量按照能量强弱逐次进行检测,并能有效检测多分量雷达辐射源信号中具有高阶相位的信号分量和能量较弱的信号分量。检测估计算法通过将多个时间延迟的HAF相乘的方法,强化信号分量形成峰值,削弱没有相关性的交叉项和噪声,避免HAF在处理多分量信号时易受噪声和交叉项形成的杂波干扰,同时避免采用最大似然估计方法时的高计算复杂度。在检测过程中采用剥洋葱策略,使能量较弱的信号分量因避免了强能量信号分量的影响及交叉项的干扰而被检测出来。仿真实验结果表明,PHAF检测估计算法能有效检测混合阶相位多分量雷达辐射源信号,对多个信号分量的检测顺序跟信号分量的能量(幅度)有关,能量大的信号分量先被检测出来,但并不影响其它信号分量的检测；不同相位阶次信号分量的检测顺序与阶次高低没有影响,参数估计精确程度与信号能量有关,能量越大的信号估计精度越高；在对含噪声信号(信噪比为OdB)进行检测时,估计误差只有小幅提高,估计精度没有显著降低。
(6)比较研究多分量雷达辐射源信号的多种时频检测估计算法和参数检测估计算法的性能,分析各检测估计算法的优缺点和适用范围,概括总结检测估计算法的性能指标,并在算法实现过程分析中提出了基于SM的快速时频处理实现方法。检测估计算法的性能分别从算法实现过程、计算复杂度和检测估计性能三方面进行比较分析。采用提出的SM快速时频方法降低时频检测估计算法的计算复杂度。检测估计算法的计算复杂度分析表明,时频检测估计算法的复杂度要高于参数检测估计算法的复杂度；在时频检测估计算法中,SHE检测估计算法的复杂度最高,RSM检测估计算法的复杂度略高于SM检测估计算法。检测估计性能是衡量检测算法有效性的指标,其研究主要考虑估计精度、噪声抑制能力和算法的适用范围。时频检测估计算法的检测性能分析以时频分布的性能为基础,参数检测估计算法的检测性能通过计算估计值的方差和克拉美-罗界获得。检测估计性能分析表明,RSM具有比SM更高的时频分辨率,继承了SM在交叉项和噪声抑制方面的优势,是比较理想的时频分布；SHE检测估计算法具有最强的检测低信噪比信号的能力；PHAF检测估计算法具有最好的高阶相位调制信号检测能力,可实现混合相位阶次多分量雷达辐射源信号的检测和弱能量信号分量检测。在总结检测估计算法性能的基础上,给出各检测估计算法性能指标比较表,以作为不同情况下检测估计算法选择的参考,并可作为多分量雷达辐射源信号检测估计算法的进一步讨论和应用研究的基础。
本论文研究工作得到国家自然科学基金项目(60971103、60702026、60572143)、四川省杰出青年基金项目(09ZQ026-040)和西南交通大学博士生创新基金项目资助。
Radar emitter signal detection and parameter estimation is a key process of signal processing in electronic intelligence systems and also a focus and difficulty in the signal processing of electronic countermeasure. In modern electronic warfare, the density of signals becomes denser and denser, and radar signals use more and more complex modulations. The dense and complex signal environment produces the situation that radar signal pulses reach receivers coincidently or successively and overlap each other, which forms multi-component radar emitter signals. The existence of multi-component radar emitter signals deteriorates the instantaneous frequency estimation method that uses a single pulse to perform detection, which easily results in estimation errors of instantaneous frequencies or the absence of important information, and therefore destroys statistic characteristics of pulses in a pulse train from one emitter and lowers correct recognition rates, even wrong recognition, of radar emitters. In the case of very limited number of pulses the method greatly affects the intelligence information gain of radar emitters. As the density of pulses grows denser and denser in the electronic warfare environment, the probability that pulses overlap is bigger and bigger. So the detection and estimation of multi-component radar emitter signals caused by overlapped pulses becomes an unavoidable and critical problem in modern electronic reconnaissance systems. Until now there is not a systematic study on detection-estimation algorithms for multi-component radar emitter signals. Therefore, this dissertation carries out a systematic, innovative and deep investigation on the problems of multi-component radar emitter signals and their detection-estimation algorithms to obtain the intelligence information of advanced radar emitters in a dense and complicated signal environment.
This dissertation first discusses the meaning and model of multi-component radar emitter signals, on the basis of the analysis of source, characteristics and representation ways of multi-component radar emitter signals, and the present probability of overlapped pulses. Then, time-frequency analysis approaches and parameter estimation methods are used to design several detection-estimation algorithms for multi-component radar emitter signals to gain various useful information of signal components in time and frequency domains and modulation types. The detection-estimation algorithms consider various types of signals and the detection-estimation of multi-component radar emitter signals in the case that there are many overlaps and noise. The detection-estimation algorithms based on time-frequency analysis consist of S-method (SM) detection-estimation algorithm, reassigned SM detection-estimation algorithm and SHE (a hybrid approach of SM, Hough transformation and elimination method in time-frequency plane) detection-estimation algorithm. The detection-estimation algorithms based on parameter estimation is devised to detect the multi-component radar emitter signals composed of signal components with greatly different energies or containing high-order phase modulation signal components. Finally, these detection-estimation algorithms are comparatively analyzed from the three aspects of the algorithm implementation processes, computational complexity and detection-estimation performances, and conclusions are drawn. Main research fruits achieved in the systematic study of multi-component radar emitter signals and their detection-estimation algorithms are as follows:
(1) The definition of multi-component radar emitter signals is given and the model of multi-component radar emitter signals is constructed. The investigation of multi-component radar emitter signals is a new and difficult problem in the signal processing of modern electronic countermeasure. The meaning and model of multi-component radar emitter signals is the basis of detection-estimation algorithms. Multi-component radar emitter signals are analyzed from their source, characteristics, representation ways and detection necessity. On the basis of analyzing radar emitter signals, the definition of a multi-component radar emitter signal is provided and the mathematical model of a multi-component radar emitter signal is given. Then, the features of multi-component radar emitter signal detection are discussed from the aspect of information collection and problems to solve, which can provide an important reference for detection-estimation algorithm design.
(2) The detection mechanism of SM is deeply analyzed and a SM-based algorithm for detecting multi-component radar emitter signals is proposed. The detection mechanism of SM is studied by analyzing cross-term elimination, noise suppression and detection performance to gain the advantages and disadvantages of SM in the process of detecting multi-component signals. Making use of the characteristics of SM in detecting multi-component signals, a multi-component radar emitter signal detection-estimation algorithm is designed. Simulation experiments show that the SM detection-estimation algorithm has high time-frequency resolution, like pseudo Wigner-Ville distribution, and avoids the interference of cross terms, and that the algorithm is able to accurately detect the instantaneous frequency, starting and ending instants of each signal component and has strong detection ability to process noised signals. The detected values fluctuate around their ideal values, but the fluctuation ranges are small. When the algorithm detects noised signals (signal-to-noise ratio (SNR) is 0 dB), the detected values may violate their ideal values, but this does not affect the approximation of the detected values to their ideal values, thus the algorithm still obtains high estimation precision.
(3) A multi-component radar emitter signal detection-estimation algorithm based on reassigned SM (RSM) is presented and its feasibility and implementation efficiency are discussed through algorithm process analysis and hardware realization. Aiming at the characteristics of multi-component radar emitter signals, time-frequency reassign is applied to improve the time-frequency concentration which is decreased by adopting frequency-domain window in the process of avoiding cross-terms of SM. Using the time-frequency reassign, the signal components in the SM distribution plane are detected more easily, and better estimation precision can be obtained. The analysis of algorithm implementation process and hardware realization shows that the main computing process of this detection-estimation algorithm is Fourier transform of the signal segment within a time-domain window and the detection-estimation algorithm is relatively simple and easy to realize for hardware. Results of simulation experiments show that RSM detection-estimation algorithm uses SM to remove cross-terms between signal components and applies time-frequency reassign to enhance time-frequency resolution. Thus, RSM detection-estimation algorithm can obtain high time-frequency resolution close to Wigner-Ville distribution and avoid the disturbance of cross-terms that Wigner-Ville owns. Also RSM detection-estimation algorithm can effectively detect both linear and non-linear frequency-modulated multi-component radar emitter signals.
(4) A detection-estimation algorithm, SHE, is proposed to detect multi-component radar emitter signals. The model of SHE detection-estimation algorithm is given. An elimination method performed in the time-frequency plane is introduced to detect signal components one by one. SHE appropriately combines SM, Hough transform with time-frequency elimination method. On the basis of the computation of SM time-frequency distribution of a signal, the noise suppression ability is improved by accumulating the linear frequency modulation signals or phase encoding signals in the parameter space of Hough transform. The introduced time-frequency elimination method is employed to detect signal components one by one, which can avoid the disturbance between signal components and multi-valued detection in the signal location of Hough transform. The starting and ending time of signal components is obtained by using a pulse detection threshold. The phase codes can be achieved by using a phase detection threshold. Simulation experiments show that SHE is able to accurately detect the instantaneous frequency and the duration time of each signal component, and that SHE can still achieve high estimation precision for the signals with low SNR (SNR is-5dB). Also the experiments show that the phase coding information of signal components can be gained by using phase elimination approach in the time-frequency plane.
(5) A multi-component radar emitter signal detection-estimation algorithm based on Product High Ambiguity Function (PHAF) is proposed. In this algorithm, a novel strategy like peeling onion is presented to detect signal components sequentially according to their energy magnitudes. The algorithm can effectively detect the multi-component radar emitter signals containing high-order phase modulation signal components or weak energy signal components. Through multiplying several high ambiguity functions with different time-delays, the proposed detection-estimation algorithm can strengthen the useful signal components and weaken noise and irrelative cross-terms to avoid the case that High Ambiguity function (HAF) is easily affected by noise and clutter waves formed by cross-terms when HAF is employed to process multi-component signals. Also, the detection-estimation algorithm reduces high computing complexity due to the use of a maximum likelihood estimation method. The onion-peeling strategy in the detection-estimation algorithm is helpful to detect the signal components with weaker energy because it eliminates the side-effects of signal components with stronger energy and cross-terms. Results of simulation experiments show that the presented algorithm can detect the multi-component radar emitter signals with various order phase modulations effectively. The detection order of signal components depends on the energy (amplitude) of signal component, so the signal components with stronger energy are first detected and therefore they do not affect the detection of other signal components. The detection order of the signal components with various orders of phase is not relative to their orders of phase. The estimation precision of parameters relates the energy of signal components. The stronger the signal components are, the higher the estimation precision is. When noised signals (SNR is OdB) are detected, the estimation error increases only in a small scale and there is not significant decrease for estimation precision.
(6) The performances of several multi-component radar emitter signal detection-estimation algorithms based on time-frequency distributions and parameter estimation are comparatively investigated. The application scopes, advantages and disadvantages of the detection-estimation algorithms are analyzed. Performance indexes for the detection-estimation algorithms are summarized. In the process of analyzing the implementation procedure of algorithms, a fast time-frequency analysis method based on SM is proposed to deal with multi-component radar emitter signals. The performances of detection-estimation algorithms are analyzed through algorithm implementation process, computing complexity and detection capability. The proposed fast time-frequency analysis method is used to decrease the computing complexity of time-frequency detection-estimation algorithms. The analysis of computing complexity of detection-estimation algorithms shows that the parameter estimation detection algorithm is better than the time-frequency detection-estimation algorithms. Among the time-frequency detection-estimation algorithms, the computing complexity of SHE is the highest and RSM detection-estimation algorithm is slightly worse than SM detection-estimation algorithm. The detection-estimation performance, mainly considering detection precision, noise suppression and application scope, is an index to evaluate the effectiveness of detection-estimation algorithms. The performance analysis of time-frequency detection-estimation algorithms is based on time-frequency distributions. The detection-estimation performance of parameter estimation detection algorithm is analyzed through computing the variance of estimation values and Cramer-Rao Bound. The detection-estimation performance analysis shows that RSM has higher time-frequency resolution than SM because RSM inherits the good points of SM about noise suppression and cross-terms elimination, so RSM is an ideal time-frequency distribution. The analysis also shows that SHE has better detection ability than other algorithms in processing the signals with low SNR, and that PHAF detection-estimation algorithm has the best detection ability in processing the signals with high order phase modulations and can detect multi-component radar emitter signals with mixed order phase modulations or weaker energy signal components. After summarizing the performance of detection-estimation algorithms, a table containing the performance indexes for comparing detection-estimation algorithms is provided to choose the detection-estimation algorithm for different situations and also to be a basis of further discussion and application research of multi-component radar emitter signals.
This work is supported by the National Natural Science Foundation of China (60971103、60702026、60572143), Scientific and Technological Funds for Young Scientists of Sichuan (09ZQ026-040), and the Doctorial Innovation Foundation of Southwest Jiaotong University.
引文
[1]R. G. Wiley. Elint the interception and analysis of radar signal [M]. Artech House, Norwood, MA,2006
[2]M. I. Skolnik.雷达系统导论[M].左群声,徐国良,马林and王德纯.第三版.电子工业出版社,2006
[3]G. W. Stimson.机载雷达导论[M].吴汉平.第二版.电子工业出版社,2005
[4]赵国庆.雷达对抗原理[M].西安电了科技大学出版社,1999
[5]H. K. Mardia. New techniques for the deinterleaving of repetitive sequences [J]. IEE Proceedings on Radar and Signal Processing,.1989,136(4):149-154
[6]D. J. Milojevic, B. M. Popovic. Improved algorithm for the deinterleaving of radar pulses [J]. IEE Proceedings on Radar and Signal Processing.1992,139(1):98-104
[7]P. S. Ray. A novel pulse TOA analysis technique for radar identification [J]. IEEE Transactions on Aerospace and Electronic Systems,.1998,34(3):716-721
[8]H. E. Hassan. Joint deinterleaving/recognition of radar pulses [A].2003. Proceedings of the International Radar Conference [C],2003:177-181
[9]孟建,胡来招.用于信号处理的重复周期变换[J].哈尔滨工业大学学报.1998,13(1)：1-7
[10]H. E. A. B. Hassan, F. Chan, Y. T. Chan. Queueing analysis of the deinterleaving of radar pulses in a dense emitter environment [A]. Canadian Conference on Electrical and Computer Engineering,2003. IEEE CCECE [C],2003:2015-2020 vol.2013
[11]N. Levanon, E. Mozeson. Rdar Signals [M]. John Wiley & Sons, Hoboken, New Jersey, 2004
[12]P. E. Pace. Detecting and classifying low probability of intercept radar [M]. Artech House, Norwood, MA,2004
[13]R. M. E. M. van Heijster. Universal precision ESM receiver based on software defined radio technology [A]. The European Conference on Wireless Technology,2005 [C], 2005:419-422
[14]林象平.雷达对抗原理[M].西北电讯工程学院出版社,西安,1985
[15]J. B. Moore, V. Krishnamurthy. Deinterleaving pulse trains using discrete-time stochastic dynamic-linear models [J]. IEEE Transactions on Signal Processing.1994, 42(11):3092-3103
[16]王乃和.关于ESM中雷达信号分选识别问题的探讨[J].电子对抗.1991.25(3)：44-51
[17]D. R. Wilkinson, A. W. Watson. Use of metric techniques in ESM data processing [J]. IEE Proceedings on Communications, Radar and Signal Processing.1985,132(4): 229-232
[18]J. Roe. A review of applications of artificial intelligence techniques to naval ESM signal processing [A].IEE Colloquium on Application of Artificial Intelligence Techniques to Signal Processing [C],1989:5/1-5/5
[19]J. Roe, S. Cussons, A. Feltham. Knowledge-based signal processing for radar ESM systems [J]. IEE Proceedings on Radar and Signal Processing 1990,137(5):293-301
[20]J. A. Anderson, M. T. Gately, P. A. Penz, et al. Radar signal categorization using a neural network [J]. Proceedings of the IEEE.1990,78(10):1646-1657
[21]A. L. Roe. Artificial neural networks for ESM emitter identification-an initial study [A]. IEE Colloquium on Neural Networks for Systems:Principles and Applications [C], 1991:4/1-4/3
[22]T. Jinsong, Z. Zhaoda. A kind of neural network similar to ART network with application to radar signal sorting [A]. Proceedings of the IEEE 1994 National Aerospace and Electronics Conference,1994. NAECON [C],1994:398-401 vol.391
[23]J. Matuszewski, L. Paradowski. The knowledge based approach for emitter identification [A].12th International Conference on Microwaves and Radar,1998. MIKON'98. [C],1998:810-814 vol.813
[24]E. Granger, Y. Savaria, P. Lavoie, et al. A comparison of self-organizing neural networks for fast clustering of radar pulses [J]. Signal Processing.1998,64(3): 249-269
[25]S. Ching-Sung, L. Chin-Teng. A vector neural network for emitter identification [J]. IEEE Transactions on Antennas and Propagation.2002,50(8):1120-1127
[26]E. Granger, M. A. Rubin, S. Grossberg, et al. A What-and-Where fusion neural network for recognition and tracking of multiple radar emitters [J]. Neural Networks.2001, 14(3):325-344
[27]E. Granger, M. A. Rubin, S. Grossberg, et al. Classification of incomplete data using the fuzzy ARTMAP neural network [A]. Proceedings of the IEEE-INNS-ENNS International Joint Conference on Neural Networks,2000. IJCNN 2000 [C],2000: 35-40 vol.36
[28]李广彪,张剑云.独立分量分析在雷达信号分选中的应用[J].航天电子对抗.2005,21(4)：18-21
[29]胡建伟,扬绍全.基于小波互信息的辐射源脉冲分类[J].系统工程与电子技术.2005,27(11)：1895-1898
[30]G. X. Zhang, L. Z. Hu, W. D. Jin. Complexity feature extraction of radar emitter signals [A].3rd Asia-Pacific Conference on Environmental Electromagnetics [C]. Hangzhou, PEOPLES R CHINA,2003:495-498
[31]G. X. Zhang, W. D. Jin, L. Z. Hu. Fractal feature extraction of radar emitter signals [A]. 3rd Asia-Pacific Conference on Environmental Electromagnetics [C]. Hangzhou, PEOPLES R CHINA,2003:161-164
[32]张葛祥,胡来招,金炜东.基于熵特征的雷达辐射源信号识别[J].电波科学学报.2005(04)
[33]G. X. Zhang, W. D. Jin, L. Z. Hu. Resemblance coefficient based intrapulse feature extraction approach for radar emitter signals [J]. Chinese Journal of Electronics.2005, 14(2):337-341
[34]G. X. Zhang, H. N. Rong. Improved quantum-inspired genetic algorithm based time-frequency analysis of radar emitter signals [A].2nd International Conference on Rough Sets and Knowledge Technology [C]. Toronto, CANADA,2007:484-491
[35]朱明,金炜东,普运伟,et al.基于Chirplet原子的雷达辐射源信号特征提取[J].红外与毫米波学报.2007(04)
[36]普运伟,金炜东,朱明,et al.雷达辐射源信号模糊函数主脊切而特征提取方法[J].红外与毫米波学报.2008(02)
[37]程柏林,马晓岩,陈蓓.基于时频重排多分量辐射源信号分析研究[J].系统工程与电子技术.2006,28(5)：684-686,778
[38]邹兴文,张葛祥.基于图像处理技术的雷达辐射源信号时频特性分析[J].电路与系统学报.2009,14(3)：135-140
[39]余志斌,金炜东.多分量LFM雷达辐射源信号的经验模式分解[J].西南交通大学学报.2009,44(1)：49-54
[40]Q. Shie, C. Dapang. Joint time-frequency analysis [J]. Signal Processing Magazine, IEEE.1999,16(2):52-67
[41]L. Cohen. Time-frequency distributions-a review [J]. Proceedings of the IEEE.1989, 77(7):941-981
[42]D. Gabor. Theory of communication [J]. Proceedings of the IEE.1946,93(111): 429～457
[43]C. H. Page. Instantaneous Power Spectra [J]. Journal of Applied Physics.1952,23(1): 103-106
[44]E. Wigner. On the Quantum Correction For Thermodynamic Equilibrium [J]. Physical Review.1932,40(5):749
[45]D. Devedeux, J. Duchene, C. Marque. Use of synthetic uterine signals for an optimum choice of time/frequency representation [A]. Engineering in Medicine and Biology Society,1994. Engineering Advances:New Opportunities for Biomedical Engineers. Proceedings of the 16th Annual International Conference of the IEEE [C],1994: 1246-1247 vol.1242
[46]E. Gil, M. Mendez, J. M. Vergara, et al. Detection of obstructive sleep apnea in children using decreases in the amplitude fluctuations of PPG signal and HRV [A]. Engineering in Medicine and Biology Society,2008. EMBS 2008.30th Annual International Conference of the IEEE [C],2008:3479-3482
[47]V. X. Afonso, W. J. Tompkins. Detecting ventricular fibrillation [J]. Engineering in Medicine and Biology Magazine, IEEE.1995,14(2):152-159
[48]S. Pola, M. Varanini, A. Macerata, et al. Bivariate spectral analysis of nonstationary cardiovascular variability series through cross Wigner distribution [A]. Computers in Cardiology 1996 [C],1996:81-84
[49]R. Bailon, P. Laguna, L. Mainardi, et al. Analysis of Heart Rate Variability Using Time-Varying Frequency Bands Based on Respiratory Frequency [A].29th Annual International Conference of the IEEEEngineering in Medicine and Biology Society [C], 2007:6674-6677
[50]R. Burnett, J. F. Watson, S. Elder. The application of modern signal processing techniques to rotor fault detection and location within three phase induction motors [A]. Instrumentation and Measurement Technology Conference,1995. IMTC/95. Proceedings.'Integrating Intelligent Instrumentation and Control'., IEEE [C],1995:
[51]J. Rosero, J. Cusido, A. Garcia, et al. Fault detection of eccentricity by means of joint time-frequency analysis in PMSM under dynamic conditions [A]. IEEE International Symposium on Intelligent Signal Processing [C],2007:1-6
[52]E. Swiercz. Recognition of radar signals by time-frequency peak filtering [A]. International Radar Symposium, IRS 2006. [C],2006:1-4
[53]L. Stankovic. Quadratic and higher order time-frequency analysis based on the short-time Fourier transform [A]. Sixth International, Symposium on Signal Processing and its Applications [C],2001:581-582 vol.582
[54]F. Hlawatsch, G. F. Boudreaux-Bartels. Linear and quadratic time-frequency signal representations [J]. Signal Processing Magazine, IEEE.1992,9(2):21-67
[55]P. Goncalves, R. G. Baraniuk. Pseudo affine Wigner distributions:definition and kernel formulation [J]. IEEE Transactions on Signal Processing.1998,46(6):1505-1516
[56]B. Boashash, P. O'Shea. Polynomial Wigner-Ville distributions and their relationship to time-varying higher order spectra [J]. IEEE Transactions on Signal Processing.1994, 42(1):216-220
[57]H. I. Choi, W. J. Williams. Improved time-frequency representation of multicomponent signals using exponential kernels [J]. IEEE Transactions on Acoustics, Speech and Signal Processing.1989,37(6):862-871
[58]B. Boashash. Time Frequency Signal Analysis [M]. Melbourne, Australia:Longman Cheshire,1992
[59]Y. Zhao, L. E. Atlas, R. J. Marks, Ⅱ. The use of cone-shaped kernels for generalized time-frequency representations of nonstationary signals [J]. IEEE Transactions on Acoustics, Speech and Signal Processing.1990,38(7):1084-1091
[60]L. Stankovic. S-class Of Time-frequency Distributions [J]. IEE Proceedings on Vision, Image and Signal Processing.1997,144(2):57-64
[61]P. Zavarsky, N. Fujii. Introduction of cross ambiguity function for elimination of crossterms in Wigner distribution of the third order [J]. Electronics Letters.1996,32(2): 94-95
[62]B. Ristic, B. Boashash. Kernel design for time-frequency signal analysis using the Radon transform [J]. Signal Processing, IEEE Transactions on.1993,41(5):1996-2008
[63]G. Jones, B. Boashash. Generalized instantaneous parameters and window matching in the time-frequency plane [J]. IEEE Transactions on Signal Processing 1997,45(5): 1264-1275
[64]L. Stankovic. A method for time-frequency analysis [J]. IEEE Transactions on Signal Processing.1994,42(1):225-229
[65]B. Boashash, B. Ristic. Polynomial time-frequency distributions and time-varying higher order spectra:Application to the analysis of multicomponent FM signals and to the treatment of multiplicative noise [J]. Signal Processing.1998,67(1):1-23
[66]C. Richard. Time-frequency-based detection using discrete-time discrete-frequency Wigner distributions [J]. IEEE Transactions on Signal Processing.2002,50(9): 2170-2176
[67]L. Stankovic. An analysis of some time-frequency and time-scale distributions [J]. Annals of Telecommunications.1994,49(9):505-517
[68]L. Stankovic. Highly concentrated time-frequency distributions:pseudo quantum signal representation [J]. IEEE Transactions on Signal Processing.1997,45(3):543-551
[69]T. J. Abatzoglou. Fast Maximnurm Likelihood Joint Estimation of Frequency and Frequency Rate [J]. IEEE Transactions on Aerospace and Electronic Systems.1986, AES-22(6):708-715
[70]B. Boashash, P. O'Shea, M. J. Arnold. Algorithms for instantaneous frequency estimation:a comparative study [A]. Advanced Signal Processing Algorithms, Architectures, and Implementations [C]. San Diego, CA, USA,1990. SPIE:126-148
[71]R. L. Burden, J. D. Faires, A. C. Reynolds. Numerical Analysis [M]. MA:Prindle, Weber & Schmidt, Boston,1978
[72]D. Rife, R. Boorstyn. Single tone parameter estimation from discrete-time observations [J]. IEEE Transactions on Information Theory.1974,20(5):591-598
[73]E. J. Kelly, I. S. Reed, W. L. Root. The Detection of Radar Echoes in Noise [J]. Journal of the Society for Industrial and Applied Mathematics. 1960,8(3):481-507
[74]S. Peleg, B. Porat. Linear FM signal parameter estimation from discrete-time observations [J]. IEEE Transactions on Aerospace and Electronic Systems.1991,27(4): 607-616
[75]S. Peleg, B. Porat. Estimation and classification of polynomial-phase signals [J]. IEEE Transactions on Information Theory.1991,37(2):422-430
[76]S. Peleg, B. Friedlander. The discrete polynomial-phase transform [J]. IEEE Transactions on Signal Processing.1995,43(8):1901-1914
[77]B. Porat, B. Friedlander. Asymptotic statistical analysis of the high-order ambiguity function for parameter estimation of polynomial-phase signals [J]. IEEE Transactions on Information Theory,.1996,42(3):995-1001
[78]B. Porat. Digital processing of random signals:theory and methods [M]. Prentice-Hall, Inc.,1994
[79]S. Peleg, B. Friedlander. Multicomponent signal analysis using the polynomial-phase transform [J]. IEEE Transactions on Aerospace and Electronic Systems.1996,32(1): 378-387
[80]L. Cohen. What is a multicomponent signal? [A].1992 IEEE International Conference on Acoustics, Speech, and Signal Processing [C],1992:113-116
[81]S. Barbarossa. Detection and estimation of the instantaneous frequency of polynomial-phase signals by multilinear time-frequency representations [A]. IEEE Signal Processing Workshop on Higher-Order Statistics [C],1993:168-172
[82]S. Barbarossa, V. Petrone. Analysis of polynomial-phase signals by the integrated generalized ambiguity function [J]. IEEE Transactions on Signal Processing.1997, 45(2):316-327
[83]S. Barbarossa, A. Porchia, A. Scaglione. Multiplicative multi-lag high order ambiguity function [A].1996 IEEE International Conference on Acoustics, Speech, and Signal Processing [C],1996:3022-3025
[84]S. Barbarossa. Analysis of multicomponent LFM signals by a combined Wigner-Hough transform [J]. IEEE Transactions on Signal Processing.1995,43(6):1511-1515
[85]Y. Wang, G. Zhou. On the use of high-order ambiguity function for multi-component polynomial phase signals [J]. Signal Processing.1998,65(2):283-296
[86]Y. Wang, G. Zhou. On the use of high order ambiguity function for multicomponent polynomial phase signals [A].1997 IEEE International Conference on Acoustics, Speech, and Signal Processing [C],1997:3629-3632
[87]S. Barbarossa, A. Scaglione, G. B. Giannakis. Product high-order ambiguity function for multicomponent polynomial-phase signal modeling [J]. IEEE Transactions on Signal Processing.1998,46(3):691-708
[88]M. Z. Ikram, G. T. Zhou. Estimation of multicomponent polynomial phase signals of mixed orders [J]. Signal Processing.2001,81(11):2293-2308
[89]向敬成,张明友.雷达系统[M].电子工业出版社.2001
[90]李祖新.雷达对抗面临严重挑战[J].舰船电子对抗.1999(1)：5-8
[91]B. Boashash. Estimating and interpreting the instantaneous frequency of a signal. I. Fundamentals [J]. Proceedings of the IEEE.1992,80(4):520-538
[92]程柏林,陈蓓,赵国林.一种多分量lfm信号脉内调制特征的提取方法[J].电子信息对抗技术.2006(03)
[93]荣海娜,张葛祥,金炜东.基于S-method的多分量辐射源信号检测[J].四川大学学报.2009,41(1)：174～179
[94]荣海娜,张葛祥,金炜东.多分量雷达辐射源信号检测新方法[J].系统工程与电子技术.2009,31(9)：2096-2100
[95]张国柱.雷达辐射源识别技术研究[D].国防科学技术大学研究生院[博士].2005
[96]D. D. Vaccaro. Electronic warfare receiving systems [M]. Artech House on Demand, Boston. London,1993
[97]S. Barbarossa, A. Zanalda. A combined Wigner-Ville and Hough transform for cross-terms suppression and optimal detection and parameter estimation [A].1992 IEEE International Conference on Acoustics, Speech, and Signal Processing,1992. ICASSP-92., [C],1992:173-176 vol.175
[98]孙晓昶,皇甫堪.基于Wigner-Hough变换的多分量LFM信号检测及离散计算方法[J].电子学报.2003,31(2)：241-244
[99]L. Stankovic, J. F. Bohme. Time-frequency analysis of multiple resonances in combustion engine signals [J]. Signal Processing.1999,79(1):15-28
[100]L. J. Stankovic, T. Thayaparan, M. Dakovic. Signal Decomposition by Using the S-Method With Application to the Analysis of HF Radar Signals in Sea-Clutter [J]. IEEE Transactions on Signal Processing.2006,54(11):4332-4342
[101]H. N. Rong, G. X. Zhang, W. D. Jin, et al. AN IMPROVED TIME FREQUENCY ANALYSIS METHOD FOR NOISY MULTI-COMPONENT SIGNALS [A]. IEEE International Conference on Signal Processing and Communications [C]. Dubai, U ARAB EMIRATES,2007:500-503
[102]H. Rong, G. Zhang, W. Jin. An Improved Time Frequency Analysis Method for Noised Multi-component Signals [A]. Proceedings of 2007 IEEE International conference on signal processing and communications [C]. Taipei,2007:500-503
[103]F. Auger, P. Flandrin. Improving the readability of time-frequency and time-scale representations by the reassignment method [J]. IEEE Transactions on Signal Processing.1995,43(5):1068-1089
[104]荣海娜,张葛祥,金炜东.基于重排S-method的多分量辐射源信号分析方法[J].西南交通大学学报.2009,44(2)：195～200
[105]K. Kodera, C. d. Villedary, R. Gendrin. A new method for the numerical analysis of time-varying signals with small BT values [J]. Phys. Earth Planet. Interiors.1976(12): 142-150
[106]K. Kodera, R. Gendrin, C. Villedary. Analysis of time-varying signals with small BT values [J]. IEEE Transactions on Acoustics, Speech and Signal Processing.1978,26(1): 64-76
[107]I. Djurovic, L. Stankovic. Time-frequency representation based on the reassigned S-method [J]. Signal Processing.1999,77(1):115-120
[108]S. Stankovic, L. Stankovic. An architecture for the realization of a system for time-frequency signal analysis [J]. IEEE Transactions on Circuits and Systems Ⅱ: Analog and Digital Signal Processing.1997,44(7):600-604
[109]于凤芹,徐美华,曹家麟.基于时频重排-Hough变换的多分量Chirp信号的检测与参数估计[J].信号处理.2005,21(6)：585-588
[110]尉宇,孙德宝,吴江洪,et al.基于WHT的多分量LFM参数估计与分离[J].系统上程与电子技术.2003,25(25)：1472-1474,1546
[111]C. H. Messom, G. Sen Gupta, S. N. Demidenko. Hough Transform Run Length Encoding for Real-Time Image Processing [J]. IEEE Transactions on Instrumentation and Measurement.2007,56(3):962-967
[112]荣海娜,张葛祥,金炜东.基于PHAF的多分量辐射源信号检测[J].西南交通大学学报.2010(2)
[113]荣海娜,张葛祥,金炜东.多分量信号快速时频分析方法[J].电路与系统学报.录用
[114]B. Barkat, B. Boashash. A high-resolution quadratic time-frequency distribution for multicomponent signals analysis [J]. IEEE Transactions on Signal Processing.2001, 49(10):2232-2239
[115]王璞,扬建宇.基于乘积性模糊函数的核函数设计方法[J].电波科学学报.2007,22(6)：1056～1060
[116]S. Peleg, B. Porat. The Cramer-Rao lower bound for signals with constant amplitude and polynomial phase [J]. IEEE Transactions on Signal Processing.1991,39(3): 749-752
[2]M. I. Skolnik.雷达系统导论[M].左群声,徐国良,马林and王德纯.第三版.电子工业出版社,2006
[3]G. W. Stimson.机载雷达导论[M].吴汉平.第二版.电子工业出版社,2005
[4]赵国庆.雷达对抗原理[M].西安电了科技大学出版社,1999
[5]H. K. Mardia. New techniques for the deinterleaving of repetitive sequences [J]. IEE Proceedings on Radar and Signal Processing,.1989,136(4):149-154
[6]D. J. Milojevic, B. M. Popovic. Improved algorithm for the deinterleaving of radar pulses [J]. IEE Proceedings on Radar and Signal Processing.1992,139(1):98-104
[7]P. S. Ray. A novel pulse TOA analysis technique for radar identification [J]. IEEE Transactions on Aerospace and Electronic Systems,.1998,34(3):716-721
[8]H. E. Hassan. Joint deinterleaving/recognition of radar pulses [A].2003. Proceedings of the International Radar Conference [C],2003:177-181
[9]孟建,胡来招.用于信号处理的重复周期变换[J].哈尔滨工业大学学报.1998,13(1)：1-7
[10]H. E. A. B. Hassan, F. Chan, Y. T. Chan. Queueing analysis of the deinterleaving of radar pulses in a dense emitter environment [A]. Canadian Conference on Electrical and Computer Engineering,2003. IEEE CCECE [C],2003:2015-2020 vol.2013
[11]N. Levanon, E. Mozeson. Rdar Signals [M]. John Wiley & Sons, Hoboken, New Jersey, 2004
[12]P. E. Pace. Detecting and classifying low probability of intercept radar [M]. Artech House, Norwood, MA,2004
[13]R. M. E. M. van Heijster. Universal precision ESM receiver based on software defined radio technology [A]. The European Conference on Wireless Technology,2005 [C], 2005:419-422
[14]林象平.雷达对抗原理[M].西北电讯工程学院出版社,西安,1985
[15]J. B. Moore, V. Krishnamurthy. Deinterleaving pulse trains using discrete-time stochastic dynamic-linear models [J]. IEEE Transactions on Signal Processing.1994, 42(11):3092-3103
[16]王乃和.关于ESM中雷达信号分选识别问题的探讨[J].电子对抗.1991.25(3)：44-51
[17]D. R. Wilkinson, A. W. Watson. Use of metric techniques in ESM data processing [J]. IEE Proceedings on Communications, Radar and Signal Processing.1985,132(4): 229-232
[18]J. Roe. A review of applications of artificial intelligence techniques to naval ESM signal processing [A].IEE Colloquium on Application of Artificial Intelligence Techniques to Signal Processing [C],1989:5/1-5/5
[19]J. Roe, S. Cussons, A. Feltham. Knowledge-based signal processing for radar ESM systems [J]. IEE Proceedings on Radar and Signal Processing 1990,137(5):293-301
[20]J. A. Anderson, M. T. Gately, P. A. Penz, et al. Radar signal categorization using a neural network [J]. Proceedings of the IEEE.1990,78(10):1646-1657
[21]A. L. Roe. Artificial neural networks for ESM emitter identification-an initial study [A]. IEE Colloquium on Neural Networks for Systems:Principles and Applications [C], 1991:4/1-4/3
[22]T. Jinsong, Z. Zhaoda. A kind of neural network similar to ART network with application to radar signal sorting [A]. Proceedings of the IEEE 1994 National Aerospace and Electronics Conference,1994. NAECON [C],1994:398-401 vol.391
[23]J. Matuszewski, L. Paradowski. The knowledge based approach for emitter identification [A].12th International Conference on Microwaves and Radar,1998. MIKON'98. [C],1998:810-814 vol.813
[24]E. Granger, Y. Savaria, P. Lavoie, et al. A comparison of self-organizing neural networks for fast clustering of radar pulses [J]. Signal Processing.1998,64(3): 249-269
[25]S. Ching-Sung, L. Chin-Teng. A vector neural network for emitter identification [J]. IEEE Transactions on Antennas and Propagation.2002,50(8):1120-1127
[26]E. Granger, M. A. Rubin, S. Grossberg, et al. A What-and-Where fusion neural network for recognition and tracking of multiple radar emitters [J]. Neural Networks.2001, 14(3):325-344
[27]E. Granger, M. A. Rubin, S. Grossberg, et al. Classification of incomplete data using the fuzzy ARTMAP neural network [A]. Proceedings of the IEEE-INNS-ENNS International Joint Conference on Neural Networks,2000. IJCNN 2000 [C],2000: 35-40 vol.36
[28]李广彪,张剑云.独立分量分析在雷达信号分选中的应用[J].航天电子对抗.2005,21(4)：18-21
[29]胡建伟,扬绍全.基于小波互信息的辐射源脉冲分类[J].系统工程与电子技术.2005,27(11)：1895-1898
[30]G. X. Zhang, L. Z. Hu, W. D. Jin. Complexity feature extraction of radar emitter signals [A].3rd Asia-Pacific Conference on Environmental Electromagnetics [C]. Hangzhou, PEOPLES R CHINA,2003:495-498
[31]G. X. Zhang, W. D. Jin, L. Z. Hu. Fractal feature extraction of radar emitter signals [A]. 3rd Asia-Pacific Conference on Environmental Electromagnetics [C]. Hangzhou, PEOPLES R CHINA,2003:161-164
[32]张葛祥,胡来招,金炜东.基于熵特征的雷达辐射源信号识别[J].电波科学学报.2005(04)
[33]G. X. Zhang, W. D. Jin, L. Z. Hu. Resemblance coefficient based intrapulse feature extraction approach for radar emitter signals [J]. Chinese Journal of Electronics.2005, 14(2):337-341
[34]G. X. Zhang, H. N. Rong. Improved quantum-inspired genetic algorithm based time-frequency analysis of radar emitter signals [A].2nd International Conference on Rough Sets and Knowledge Technology [C]. Toronto, CANADA,2007:484-491
[35]朱明,金炜东,普运伟,et al.基于Chirplet原子的雷达辐射源信号特征提取[J].红外与毫米波学报.2007(04)
[36]普运伟,金炜东,朱明,et al.雷达辐射源信号模糊函数主脊切而特征提取方法[J].红外与毫米波学报.2008(02)
[37]程柏林,马晓岩,陈蓓.基于时频重排多分量辐射源信号分析研究[J].系统工程与电子技术.2006,28(5)：684-686,778
[38]邹兴文,张葛祥.基于图像处理技术的雷达辐射源信号时频特性分析[J].电路与系统学报.2009,14(3)：135-140
[39]余志斌,金炜东.多分量LFM雷达辐射源信号的经验模式分解[J].西南交通大学学报.2009,44(1)：49-54
[40]Q. Shie, C. Dapang. Joint time-frequency analysis [J]. Signal Processing Magazine, IEEE.1999,16(2):52-67
[41]L. Cohen. Time-frequency distributions-a review [J]. Proceedings of the IEEE.1989, 77(7):941-981
[42]D. Gabor. Theory of communication [J]. Proceedings of the IEE.1946,93(111): 429～457
[43]C. H. Page. Instantaneous Power Spectra [J]. Journal of Applied Physics.1952,23(1): 103-106
[44]E. Wigner. On the Quantum Correction For Thermodynamic Equilibrium [J]. Physical Review.1932,40(5):749
[45]D. Devedeux, J. Duchene, C. Marque. Use of synthetic uterine signals for an optimum choice of time/frequency representation [A]. Engineering in Medicine and Biology Society,1994. Engineering Advances:New Opportunities for Biomedical Engineers. Proceedings of the 16th Annual International Conference of the IEEE [C],1994: 1246-1247 vol.1242
[46]E. Gil, M. Mendez, J. M. Vergara, et al. Detection of obstructive sleep apnea in children using decreases in the amplitude fluctuations of PPG signal and HRV [A]. Engineering in Medicine and Biology Society,2008. EMBS 2008.30th Annual International Conference of the IEEE [C],2008:3479-3482
[47]V. X. Afonso, W. J. Tompkins. Detecting ventricular fibrillation [J]. Engineering in Medicine and Biology Magazine, IEEE.1995,14(2):152-159
[48]S. Pola, M. Varanini, A. Macerata, et al. Bivariate spectral analysis of nonstationary cardiovascular variability series through cross Wigner distribution [A]. Computers in Cardiology 1996 [C],1996:81-84
[49]R. Bailon, P. Laguna, L. Mainardi, et al. Analysis of Heart Rate Variability Using Time-Varying Frequency Bands Based on Respiratory Frequency [A].29th Annual International Conference of the IEEEEngineering in Medicine and Biology Society [C], 2007:6674-6677
[50]R. Burnett, J. F. Watson, S. Elder. The application of modern signal processing techniques to rotor fault detection and location within three phase induction motors [A]. Instrumentation and Measurement Technology Conference,1995. IMTC/95. Proceedings.'Integrating Intelligent Instrumentation and Control'., IEEE [C],1995:
[51]J. Rosero, J. Cusido, A. Garcia, et al. Fault detection of eccentricity by means of joint time-frequency analysis in PMSM under dynamic conditions [A]. IEEE International Symposium on Intelligent Signal Processing [C],2007:1-6
[52]E. Swiercz. Recognition of radar signals by time-frequency peak filtering [A]. International Radar Symposium, IRS 2006. [C],2006:1-4
[53]L. Stankovic. Quadratic and higher order time-frequency analysis based on the short-time Fourier transform [A]. Sixth International, Symposium on Signal Processing and its Applications [C],2001:581-582 vol.582
[54]F. Hlawatsch, G. F. Boudreaux-Bartels. Linear and quadratic time-frequency signal representations [J]. Signal Processing Magazine, IEEE.1992,9(2):21-67
[55]P. Goncalves, R. G. Baraniuk. Pseudo affine Wigner distributions:definition and kernel formulation [J]. IEEE Transactions on Signal Processing.1998,46(6):1505-1516
[56]B. Boashash, P. O'Shea. Polynomial Wigner-Ville distributions and their relationship to time-varying higher order spectra [J]. IEEE Transactions on Signal Processing.1994, 42(1):216-220
[57]H. I. Choi, W. J. Williams. Improved time-frequency representation of multicomponent signals using exponential kernels [J]. IEEE Transactions on Acoustics, Speech and Signal Processing.1989,37(6):862-871
[58]B. Boashash. Time Frequency Signal Analysis [M]. Melbourne, Australia:Longman Cheshire,1992
[59]Y. Zhao, L. E. Atlas, R. J. Marks, Ⅱ. The use of cone-shaped kernels for generalized time-frequency representations of nonstationary signals [J]. IEEE Transactions on Acoustics, Speech and Signal Processing.1990,38(7):1084-1091
[60]L. Stankovic. S-class Of Time-frequency Distributions [J]. IEE Proceedings on Vision, Image and Signal Processing.1997,144(2):57-64
[61]P. Zavarsky, N. Fujii. Introduction of cross ambiguity function for elimination of crossterms in Wigner distribution of the third order [J]. Electronics Letters.1996,32(2): 94-95
[62]B. Ristic, B. Boashash. Kernel design for time-frequency signal analysis using the Radon transform [J]. Signal Processing, IEEE Transactions on.1993,41(5):1996-2008
[63]G. Jones, B. Boashash. Generalized instantaneous parameters and window matching in the time-frequency plane [J]. IEEE Transactions on Signal Processing 1997,45(5): 1264-1275
[64]L. Stankovic. A method for time-frequency analysis [J]. IEEE Transactions on Signal Processing.1994,42(1):225-229
[65]B. Boashash, B. Ristic. Polynomial time-frequency distributions and time-varying higher order spectra:Application to the analysis of multicomponent FM signals and to the treatment of multiplicative noise [J]. Signal Processing.1998,67(1):1-23
[66]C. Richard. Time-frequency-based detection using discrete-time discrete-frequency Wigner distributions [J]. IEEE Transactions on Signal Processing.2002,50(9): 2170-2176
[67]L. Stankovic. An analysis of some time-frequency and time-scale distributions [J]. Annals of Telecommunications.1994,49(9):505-517
[68]L. Stankovic. Highly concentrated time-frequency distributions:pseudo quantum signal representation [J]. IEEE Transactions on Signal Processing.1997,45(3):543-551
[69]T. J. Abatzoglou. Fast Maximnurm Likelihood Joint Estimation of Frequency and Frequency Rate [J]. IEEE Transactions on Aerospace and Electronic Systems.1986, AES-22(6):708-715
[70]B. Boashash, P. O'Shea, M. J. Arnold. Algorithms for instantaneous frequency estimation:a comparative study [A]. Advanced Signal Processing Algorithms, Architectures, and Implementations [C]. San Diego, CA, USA,1990. SPIE:126-148
[71]R. L. Burden, J. D. Faires, A. C. Reynolds. Numerical Analysis [M]. MA:Prindle, Weber & Schmidt, Boston,1978
[72]D. Rife, R. Boorstyn. Single tone parameter estimation from discrete-time observations [J]. IEEE Transactions on Information Theory.1974,20(5):591-598
[73]E. J. Kelly, I. S. Reed, W. L. Root. The Detection of Radar Echoes in Noise [J]. Journal of the Society for Industrial and Applied Mathematics. 1960,8(3):481-507
[74]S. Peleg, B. Porat. Linear FM signal parameter estimation from discrete-time observations [J]. IEEE Transactions on Aerospace and Electronic Systems.1991,27(4): 607-616
[75]S. Peleg, B. Porat. Estimation and classification of polynomial-phase signals [J]. IEEE Transactions on Information Theory.1991,37(2):422-430
[76]S. Peleg, B. Friedlander. The discrete polynomial-phase transform [J]. IEEE Transactions on Signal Processing.1995,43(8):1901-1914
[77]B. Porat, B. Friedlander. Asymptotic statistical analysis of the high-order ambiguity function for parameter estimation of polynomial-phase signals [J]. IEEE Transactions on Information Theory,.1996,42(3):995-1001
[78]B. Porat. Digital processing of random signals:theory and methods [M]. Prentice-Hall, Inc.,1994
[79]S. Peleg, B. Friedlander. Multicomponent signal analysis using the polynomial-phase transform [J]. IEEE Transactions on Aerospace and Electronic Systems.1996,32(1): 378-387
[80]L. Cohen. What is a multicomponent signal? [A].1992 IEEE International Conference on Acoustics, Speech, and Signal Processing [C],1992:113-116
[81]S. Barbarossa. Detection and estimation of the instantaneous frequency of polynomial-phase signals by multilinear time-frequency representations [A]. IEEE Signal Processing Workshop on Higher-Order Statistics [C],1993:168-172
[82]S. Barbarossa, V. Petrone. Analysis of polynomial-phase signals by the integrated generalized ambiguity function [J]. IEEE Transactions on Signal Processing.1997, 45(2):316-327
[83]S. Barbarossa, A. Porchia, A. Scaglione. Multiplicative multi-lag high order ambiguity function [A].1996 IEEE International Conference on Acoustics, Speech, and Signal Processing [C],1996:3022-3025
[84]S. Barbarossa. Analysis of multicomponent LFM signals by a combined Wigner-Hough transform [J]. IEEE Transactions on Signal Processing.1995,43(6):1511-1515
[85]Y. Wang, G. Zhou. On the use of high-order ambiguity function for multi-component polynomial phase signals [J]. Signal Processing.1998,65(2):283-296
[86]Y. Wang, G. Zhou. On the use of high order ambiguity function for multicomponent polynomial phase signals [A].1997 IEEE International Conference on Acoustics, Speech, and Signal Processing [C],1997:3629-3632
[87]S. Barbarossa, A. Scaglione, G. B. Giannakis. Product high-order ambiguity function for multicomponent polynomial-phase signal modeling [J]. IEEE Transactions on Signal Processing.1998,46(3):691-708
[88]M. Z. Ikram, G. T. Zhou. Estimation of multicomponent polynomial phase signals of mixed orders [J]. Signal Processing.2001,81(11):2293-2308
[89]向敬成,张明友.雷达系统[M].电子工业出版社.2001
[90]李祖新.雷达对抗面临严重挑战[J].舰船电子对抗.1999(1)：5-8
[91]B. Boashash. Estimating and interpreting the instantaneous frequency of a signal. I. Fundamentals [J]. Proceedings of the IEEE.1992,80(4):520-538
[92]程柏林,陈蓓,赵国林.一种多分量lfm信号脉内调制特征的提取方法[J].电子信息对抗技术.2006(03)
[93]荣海娜,张葛祥,金炜东.基于S-method的多分量辐射源信号检测[J].四川大学学报.2009,41(1)：174～179
[94]荣海娜,张葛祥,金炜东.多分量雷达辐射源信号检测新方法[J].系统工程与电子技术.2009,31(9)：2096-2100
[95]张国柱.雷达辐射源识别技术研究[D].国防科学技术大学研究生院[博士].2005
[96]D. D. Vaccaro. Electronic warfare receiving systems [M]. Artech House on Demand, Boston. London,1993
[97]S. Barbarossa, A. Zanalda. A combined Wigner-Ville and Hough transform for cross-terms suppression and optimal detection and parameter estimation [A].1992 IEEE International Conference on Acoustics, Speech, and Signal Processing,1992. ICASSP-92., [C],1992:173-176 vol.175
[98]孙晓昶,皇甫堪.基于Wigner-Hough变换的多分量LFM信号检测及离散计算方法[J].电子学报.2003,31(2)：241-244
[99]L. Stankovic, J. F. Bohme. Time-frequency analysis of multiple resonances in combustion engine signals [J]. Signal Processing.1999,79(1):15-28
[100]L. J. Stankovic, T. Thayaparan, M. Dakovic. Signal Decomposition by Using the S-Method With Application to the Analysis of HF Radar Signals in Sea-Clutter [J]. IEEE Transactions on Signal Processing.2006,54(11):4332-4342
[101]H. N. Rong, G. X. Zhang, W. D. Jin, et al. AN IMPROVED TIME FREQUENCY ANALYSIS METHOD FOR NOISY MULTI-COMPONENT SIGNALS [A]. IEEE International Conference on Signal Processing and Communications [C]. Dubai, U ARAB EMIRATES,2007:500-503
[102]H. Rong, G. Zhang, W. Jin. An Improved Time Frequency Analysis Method for Noised Multi-component Signals [A]. Proceedings of 2007 IEEE International conference on signal processing and communications [C]. Taipei,2007:500-503
[103]F. Auger, P. Flandrin. Improving the readability of time-frequency and time-scale representations by the reassignment method [J]. IEEE Transactions on Signal Processing.1995,43(5):1068-1089
[104]荣海娜,张葛祥,金炜东.基于重排S-method的多分量辐射源信号分析方法[J].西南交通大学学报.2009,44(2)：195～200
[105]K. Kodera, C. d. Villedary, R. Gendrin. A new method for the numerical analysis of time-varying signals with small BT values [J]. Phys. Earth Planet. Interiors.1976(12): 142-150
[106]K. Kodera, R. Gendrin, C. Villedary. Analysis of time-varying signals with small BT values [J]. IEEE Transactions on Acoustics, Speech and Signal Processing.1978,26(1): 64-76
[107]I. Djurovic, L. Stankovic. Time-frequency representation based on the reassigned S-method [J]. Signal Processing.1999,77(1):115-120
[108]S. Stankovic, L. Stankovic. An architecture for the realization of a system for time-frequency signal analysis [J]. IEEE Transactions on Circuits and Systems Ⅱ: Analog and Digital Signal Processing.1997,44(7):600-604
[109]于凤芹,徐美华,曹家麟.基于时频重排-Hough变换的多分量Chirp信号的检测与参数估计[J].信号处理.2005,21(6)：585-588
[110]尉宇,孙德宝,吴江洪,et al.基于WHT的多分量LFM参数估计与分离[J].系统上程与电子技术.2003,25(25)：1472-1474,1546
[111]C. H. Messom, G. Sen Gupta, S. N. Demidenko. Hough Transform Run Length Encoding for Real-Time Image Processing [J]. IEEE Transactions on Instrumentation and Measurement.2007,56(3):962-967
[112]荣海娜,张葛祥,金炜东.基于PHAF的多分量辐射源信号检测[J].西南交通大学学报.2010(2)
[113]荣海娜,张葛祥,金炜东.多分量信号快速时频分析方法[J].电路与系统学报.录用
[114]B. Barkat, B. Boashash. A high-resolution quadratic time-frequency distribution for multicomponent signals analysis [J]. IEEE Transactions on Signal Processing.2001, 49(10):2232-2239
[115]王璞,扬建宇.基于乘积性模糊函数的核函数设计方法[J].电波科学学报.2007,22(6)：1056～1060
[116]S. Peleg, B. Porat. The Cramer-Rao lower bound for signals with constant amplitude and polynomial phase [J]. IEEE Transactions on Signal Processing.1991,39(3): 749-752
© 2004-2018 中国地质图书馆版权所有 京ICP备05064691号 京公网安备11010802017129号 地址：北京市海淀区学院路29号 邮编：100083 电话：办公室：(+86 10)66554848；文献借阅、咨询服务、科技查新：66554700 |