【作者】 刘旭波；

【导师】 司锡才；

【作者基本信息】 哈尔滨工程大学 ， 通信与信息系统， 2011， 博士

【摘要】 低截获概率(LPI:Low Probability of Intercept)雷达的出现,对现代被动侦察接收机和反辐射导弹导引头提出了新的挑战。脉冲压缩雷达是一种LPI雷达,该体制雷达可将接收到的长脉冲雷达信号压缩为窄脉冲,在降低了截获概率的同时,提高了雷达的距离分辨率。传统接收机已不能有效截获此类雷达信号,因此,研制具有保存脉冲压缩雷达信号频率以及相位信息功能的宽带数字接收机势在必行。数字信道化接收机作为一种较为成熟的宽带数字接收机体制,可以有效实现对不同体制雷达信号的“盲接收”。完整接收到信号后,需要应用一些数字信号处理算法对脉冲压缩雷达信号的调制参数进行提取。脉冲压缩雷达信号最为普遍采用的信号类型包括相位编码(PSK:Phase Shift Keying)信号和调频(FM:Frequency Modulation)信号,PSK信号中最常用的是二相编码(BPSK:Binary Phase Shift Keying)信号和四相编码(QPSK:Quadrature Phase Shift Keying)信号；FM信号包括线性调频(LFM:Linear Frequency Modulation)及非线性调频(NLFM:Non Linear Frequency Modulation)信号。PSK信号的参数估计包括提取信号载频、码元速率以及编码规律；LFM信号参数估计主要完成信号初始、终止频率以及调频斜率的提取；而对NLFM信号的处理,主要完成信号相位多项式系数的估计。信号分选部分主要针对从信道化接收机输出的雷达信号参数,首先进行雷达信号预分选,去除不符合要求的雷达信号,对威胁信号进行去交错,达到稀释脉冲信号流的目的；再通过单脉冲的到达时间计算脉冲串的重复间隔,完成主分选。论文围绕宽带数字信道化接收机平台截获和识别脉冲压缩雷达信号,对PSK雷达信号和FM雷达信号的参数估计算法进行了深入的研究,对雷达信号分选算法进行了细致的探讨。在宽带信道化接收技术方面,推导出了基于多相滤波与快速傅里叶变换(FFT:Fast Fourier Transform)相结合的高速高效数字均匀信道化结构。在信道化后续处理中,采用CORDIC算法实现了对信号瞬时幅度和相位的提取,针对信道化结构第0个信道输出数据为实数,不能正确提取信号相位和幅度的问题,提出基于Hilbert变换的0信道复数化处理方法。在雷达信号参数估计方面,首先,应用频谱分析法对雷达信号脉内调制类型进行分类。应用循环自相关法实现对PSK信号的调制类型识别,可区分BPSK信号和QPSK信号,针对此类信号的参数提取,提出了基于积分包络的PSK信号参数估计方法,结合信号载频,可估计信号的子码宽度及编码规律；针对LFM信号参数估计问题,仍可应用积分包络法,可实现对此类信号的调频斜率、初始以及终止频率的估计；采用三次相位函数(CPF:Cubic Phase Function)法可实现对LFM及NLFM雷达信号的参数估计,可提取此类信号相位多项式的各系数,针对该算法计算量大的问题,提出调频斜率搜索范围设定准则,另外,提出双尺度调频斜率搜索法,可减少大量计算量,同时保持了算法精度。在信号分选方面,为了解决传统方法不能有效分选高密度复杂雷达信号,以及现有聚类方法达不到实时性要求的问题,提出一种利用雷达脉冲信号特征参数,基于改进的模糊聚类的雷达信号预分选方法。在改进的算法中,提出了距离加权系数的概念,依据信息熵使各参数距离加权系数的设置更加合理化。采用追踪法计算等价模糊矩阵,并提取聚类后的脉冲序号更具有实际意义。在雷达信号主分选方面应用修正的脉冲重复间隔(PRI:Pulse Repetition Interval)算法测量雷达信号的PRI。在硬件实现上提出一种非均匀设定PRI箱宽度的方法,可节省大量的算法时间,提高系统实时性。该方法既解决了以往直方图统计算法中的子谐波问题,又克服了传统PRI算法不能有效分选抖动PRI脉冲序列的缺点。更多还原

【Abstract】 It presented new challenges to the modern passive investigation receiver and anti-radiation missile seeker with the appearance of low probability of intercept (LPI) radar. Pulse compression radar is a kind of LPI radar. This kind of radar can compress the return long pulse into a short one. The pulse compress technology can reduce the probability of intercept. Also, it can improve radar’s range resolution. The traditional receiver can’t receive this kind of radar signals, so, the wide-band digital receiver which can store frequency and phase information must be developed. As a mature wide-band digital receiver system, the digital channelized receiver can achieve "blind receiver" to different system radar signals.The pulse compression radar signal modulation parameters can be extracted by using modern signal processing algorithms, once the signals are intercepted. Phase shift keying (PSK) signals and frequency modulation (FM) signals are two kinds of widely used pulse compression signals. The PSK signals include binary phase shift keying (BPSK) signals and quadrature phase shift keying (QPSK) signals. Linear frequency modulation (LFM) and non-linear frequency modulation (NLFM) signals are the kinds of FM signals. To achieve parameters estimation of PSK, the carrier frequency, coding sequence and code rate must be extrated. The LFM signals parameters which need to be extracted include the initial frequency, the end frequency and the chirp rate. For processing to the non-linear FM signal, the phase polynomial coefficients must be estimated.In the signals sorting, the steps are that, at first, achieve pre-sorting by using the signal parameters obtained from digital channelized receiver, then, remove the radar signals which don’t meet the requirements to achieve the purpose of diluting pulse flow, at last, in the main sorting, the pulse repetition interval of pulse train was calculated by the arrival time of a single pulse.The pulse compression radar signals were intercepted and identificated in digital channelized receiver. In the paper, the parameter estimation algorithms of PSK and FM signals were researched in deep, and the signal sorting algorithms were discussed in detail.The derivation of high-speed and efficient structure of the uniform digital channelized receiver based on polyphase filter and FFT was given in the paper. For processing of channelized subsequent, CORDIC algorithm was adopted to extract instantaneous amplitude and phase of signals. The output of the zero channel is real, to this question, a method called complex processing to the zero channel based on Hilbert transform was proposed. In the terms of signal parameters estimation, at first, the spectral analysis can be used to complete pulse modulation classification. The cyclic autocorrelation can achieve recognition to PSK signals, this method can distinguish between BPSK and QPSK signals, to extract this kind of signal parameters, a method called integral-envelope has been proposed, combined with the carrier frequency, the subcode width and coding rule can be estimated. This algorithm can also estimate the chirp-rate, initial and terminal frequency of LFM signals. The method called cubic phase function (CPF) can be used to achieve parameters estimation of LFM and NLFM signals, this method can extract the phase function coefficients. The computation of this algorithm is very big, for this question, the search range criteria was proposed, in addition, a method called dual-scale search of frequency slope was brought forward to reduce the computation, also, this method can maintain the accuracy of the algorithm.In the terms of signals sorting, To solve the problems that the traditional methods can’t effectively sort high-density and complex radar-signals, and the existing clustering methods can’t achieve real-time requirement. A new method of sorting radar-signals based on modified fuzzy clustering by using radar-pulse parameters has been proposed. In this modified algorithm, the concept of distance weighted coefficient was put forward, it’s more reasonable to determine the distance weighted coefficients of the parameters based on information entropy. By using the tracking method to calculate the equivalent fuzzy matrix and get the serial number of the pulses has more practical significance. The modified PRI algorithm can be used to calculate the PRI of signal. In the hardware, a method called non-uniform setting PRI-box width was proposed, it can save much algorithm running time to enhance the system real-time performance. This sorting method solved the sub-harmonic problem in the past histogram statistical method and overcame shortcomings existed in the traditional PRI algorithm which can’t effectively sort jitter pulse trains.更多还原