雷达抗干扰波形优化设计的研究

发布时间：2018-06-04 14:40

本文选题：雷达 + 波形设计　；参考：《哈尔滨工业大学》2014年博士论文

【摘要】：传统的雷达系统一旦确定发射波形以后,便无法在运行时再对波形进行更改。因此雷达波形设计算法需要考虑各种不同的性能指标,以便设计出完美的波形,使其适用于可能遭遇到的雷达场景中。然而,由于波形设计的自由度有限,当所有指标添加到同一个波形优化的目标函数之后,难免顾此失彼。随着在线波形设计技术日趋成熟,越来越多的研究者投入到波形设计的研究中。一旦雷达可以在运行时改变发射波形,就没有必要去设计在各个方面都“完美”的波形;波形设计算法可以针对当前的雷达场景量身定做发射波形;此时只需要考虑当前环境中的目标和干扰。也就是说,在保证波形设计自由度不变的情况下减少优化指标之后,可以使设计波形相比传统波形在某些特定的方面具有更好的性能,能够更有效地减小当前环境中的干扰。在线波形设计技术主要由两个部分构成:1)用于估计当前雷达场景参数的估计器;2)波形设计算法。近年来,雷达的研究者们已经提出了众多可用于在线波形设计的优化算法,但是这些算法都或多或少存在一些特定的问题和缺陷,使它们无法应用于一些复杂的雷达场景中。本课题针对这些问题及缺陷,着重研究了波形设计算法。首先根据雷达场景中常见的距离旁瓣遮蔽问题,提出了迭代的功率谱逼近算法(Iterative Spectral Approximation Algorithm,ISAA)框架。在引入了脉冲压缩的概念之后,距离旁瓣遮蔽一直是雷达设计者需要考虑的问题。匹配滤波器输出的距离像可以建模为雷达场景与波形自相关函数的卷积;在这种模型下,波形的自相关函数类似成像模型中的点扩散函数,将一个距离单元中的回波能量散布到了其它的距离单元之中,从而导致了距离像的模糊。但不同于光学系统,对于主动雷达系统而言,发射波形是可以掌控的要素。这意味着一旦获取了当前距离像的粗略信息,并了解了强散射体存在的距离单元,就可以通过设计在指定区间具有低自相关幅值的波形,以抑制强散射体距离旁瓣对特定距离区间的干扰;而ISAA框架正是在这种背景下为了设计波形而提出的。ISAA框架根据波形的恒模特性构造约束条件,利用相关函数与谱的傅立叶变换关系在频域构造设计目标,通过交替投影的方法实现了波形优化。本课题提出了一种新颖的动态理想自相关构造方法(Dynamic Ideal Autocorrelation Construction,DIAC),该方法与ISAA框架结合之后,可以得到一种高效的具体算法,相比国内外其它研究中提出的同类算法,减小了波形优化的时间消耗。其次,提出了名为纯相位非线性规划(Phase-Only Nonlinear Programming,PONLP)的算法框架。PONLP与ISAA框架类似,也可用于波形自相关的优化。然而与ISAA框架相比,PONLP有一个独特的优势:在设计波形时,可以在其自相关序列上设置数量较多,宽度较窄的低相关幅值区间。具有这种特点的波形可以用来抑制海杂波尖峰产生的距离旁瓣。不同于ISAA,PONLP是一种基于梯度的优化方法。不同于传统的基于梯度的方法,PONLP使用纯相位导数替代导数,使用纯相位一维搜索替代直线型的一维搜索。通过这种方式,算法的每一次的迭代都能保证设计波形位于可行域之中,从而提高了收敛速度。第三,针对雷达场景中可能存在的大量窄带有源干扰,提出了秩亏傅立叶变换的概念,并将其应用于交替投影框架之中,得到了ISAA框架的另一类具体算法。其设计波形在指定的频率区间具有极低的能量分布,从而降低了窄带有源干扰对信号的影响。传统的雷达系统可以在接收端设置带阻滤波器以抑制外部的有源干扰。然而,如果干扰和发射波形在频率轴有重叠,那么带阻滤波器在滤除干扰的同时,也会消除一部分来自目标的回波能量,进而降低信噪比。而波形设计的方法可以保证波形与干扰尽可能在频率轴没有重叠,使接收端的滤波器在滤除干扰的同时不会削弱回波。为了让设计波形在具有特定的功率谱形状的同时对波形的自相关函数形状进行优化,分别提出了基于多集合交替投影的ISAA框架和基于多目标的PONLP框架,由这两种框架导出的具体算法均可完成既定的设计目标。第四,根据多波形设计和应用的特点,对ISAA框架进行了扩展。分别提出了基于ISAA的互相关优化-内积约束(Cross-Correlation Optimization-Inner Product Constraint,CSOIPC)算法和多维ISAA算法(Multi-Dimensional ISAA,MDISAA),基于广义ISAA(Generalized-ISAA,GISAA)的APLOWP框架。其中CSOIPC可用于发射-接收联合优化,APLOWP可为瞬态极化雷达系统设计发射波形,而MDISAA则是一种通用的多波形设计框架,可应用于多输入-多输出(Multi-Input,Multi-Output,MIMO)雷达的正交波形设计。最后,针对实际工作中遇到的各种雷达场景的特点,通过数学建模的方法构造了若干假想的雷达场景;在这些场景中,利用计算机数值仿真对本课题提出的各种波形设计算法进行了测试。数值仿真的结果验证了这些算法的有效性。
[Abstract]:The traditional radar system can not change the waveform again when the waveform is confirmed, so the radar waveform design algorithm needs to consider various performance indicators so as to design a perfect waveform to apply to the possible radar scene. However, because the degree of freedom of the waveform design is limited, when the radar waveform design is limited, When the target function is added to the same waveform optimization, it is difficult to avoid it. As the online waveform design technology is mature, more and more researchers have put into the research of waveform design. Once radar can change the waveform of the launch, it is not necessary to design the "perfect" waveform in all aspects; The shape design algorithm can tailor the waveform for the current radar scene. At this time, only the target and the interference in the current environment are considered. That is to say, the design waveform can have better performance than the traditional waveform in certain aspects after reducing the optimization index under the condition that the free degree of the waveform design is kept constant. It can reduce the interference in the current environment more effectively. The online waveform design technology consists of two parts: 1) the estimator for the current radar scene parameters; 2) the waveform design algorithm. In recent years, the radar researchers have proposed a number of optimization algorithms that can be used for online wave design, but these algorithms are all or more or more. There are few specific problems and defects so that they can not be applied to some complex radar scenes. This topic focuses on the research of waveform design algorithms for these problems and defects. Firstly, according to the common distance sidelobe occlusion problem in radar scenes, an iterative work rate spectrum approximation algorithm (Iterative Spectral Approximation) is proposed. Algorithm, ISAA) frame. After the introduction of the concept of pulse compression, the distance sidelobe occlusion has been a problem that radar designers need to consider. The range image of the output of the matched filter can be modeled as the convolution of the autocorrelation function of the radar scene and the waveform; in this model, the autocorrelation function of the waveform is similar to the point diffusion in the imaging model. The function is spread the echo energy in a distance unit into other distance units, which leads to the blurring of the distance image. But unlike the optical system, the emission waveform is a controllable element for the active radar system. This means that once the rough information of the current distance image is obtained and the strong scatterer is known. In the distance unit, the waveform with low autocorrelation amplitude can be designed in the specified interval to suppress the interference of the strong scatterer's distance sidelobe to a specific distance interval; and the ISAA frame is the.ISAA frame proposed to design the waveform in this background to construct the constraint conditions based on the constant modeling of the waveform, and use the correlation function and the correlation function. The Fu Liye transform relation in the spectrum is designed in the frequency domain, and the waveform optimization is realized by alternating projection. A novel dynamic ideal autocorrelation construction method (Dynamic Ideal Autocorrelation Construction, DIAC) is proposed in this paper. After the method is combined with the ISAA framework, a kind of efficient and specific algorithm can be obtained. Compared with the similar algorithms proposed in other studies at home and abroad, the time consumption of waveform optimization is reduced. Secondly, the algorithm framework named Phase-Only Nonlinear Programming (PONLP) is similar to ISAA framework, and can also be used for waveform autocorrelation optimization. However, compared with the ISAA framework, PONLP has a unique feature. Advantage: when designing a waveform, a low correlation amplitude range is set on its autocorrelation sequence. The waveform with this characteristic can be used to suppress the distance sidelobe generated by the sea clutter peak. Unlike ISAA, PONLP is a gradient based optimization method. Different from the traditional gradient based method, PONLP makes The pure phase derivative is replaced by the derivative, and the pure phase one-dimensional search is used to replace the linear one-dimensional search. By this way, each iteration of the algorithm can ensure the design waveform is in the feasible domain, thus improving the convergence speed. Third, the rank loss Fu Li is proposed for a large number of narrow band active interference which may exist in the radar scene. The concept of leaf transformation is applied to the alternate projection frame, and another specific algorithm of the ISAA frame is obtained. The design waveform has a very low energy distribution in the specified frequency range, thus reducing the influence of the narrow band active interference on the signal. However, if the interference and emission waveforms overlap in the frequency axis, the band resistance filter also eliminates a part of the echo energy from the target, and then reduces the signal to noise ratio while filtering the interference. And the waveform design method can ensure that the waveform and interference are not overlapped at the frequency axis as much as possible, so that the receiver is filtered. It does not weaken the echo while filtering the interference. In order to optimize the shape of the autocorrelation function of the waveform at the same time that the design waveform has a specific power spectrum shape, the ISAA framework based on multi set alternating projection and the multi-objective PONLP framework are proposed respectively. The specific algorithms derived from these two frameworks can complete both. Fourth, according to the characteristics of multi waveform design and application, the ISAA framework is extended. The cross correlation optimization based on ISAA (Cross-Correlation Optimization-Inner Product Constraint, CSOIPC) and multidimensional ISAA algorithm (Multi-Dimensional ISAA, MDISAA) are respectively proposed, based on the generalized ISAA. The APLOWP framework of d-ISAA, GISAA). In which CSOIPC can be used for the joint optimization of transmit and receive, APLOWP can design the transmitting waveform for the transient polarization radar system, and MDISAA is a universal multi waveform design framework, which can be applied to the orthogonal waveform design of the multiple input multiple output (Multi-Input, Multi-Output, MIMO) radar. Finally, the actual work is done. A number of imaginary radar scenes are constructed by mathematical modeling. In these scenarios, various waveform design algorithms proposed by this subject are tested by numerical simulation in these scenes. The results of numerical simulation verify the effectiveness of these algorithms.
【学位授予单位】：哈尔滨工业大学
【学位级别】：博士
【学位授予年份】：2014
【分类号】：TN974

【参考文献】