高机动微弱目标的检测方法研究

发布时间：2018-05-03 20:43

本文选题：距离徙动 + 多普勒徙动　；参考：《西安电子科技大学》2014年硕士论文

【摘要】：延长积累时间可增强目标的信号能量,进而改善雷达对目标的检测性能。但对于高速、高机动目标而言,随着时间的延长,雷达所接收到的目标回波会明显地产生距离走动和弯曲以及多普勒频偏、扩散和模糊等现象。对这类目标如果不能有效地消除距离徙动和多普勒徙动,那么其能量就得不到有效积累。本文主要针对天基雷达在对高机动微弱目标进行检测的过程中的碰到的这些问题,介绍了几种有效的长时间目标能量相参积累方法,并研究了在多目标环境下如何对不同目标的速度进行超分辨估计。首先,介绍了线性调频信号及其脉冲压缩原理,构建出高机动微弱目标的回波信号模型,并利用所建模型对目标作相对运动时所导致的距离徙动和多普勒徙动等问题进行了讨论。接着,介绍了最大似然法,通过构造匹配函数和设计合适的多普勒调频率搜索间隔和中心频率搜索间隔可有效地消除距离徙动和多普勒徙动。分析了基于一阶keystone变换的方法,它不需要目标的先验信息就可一次性矫正多目标的距离徙动现象,但不能解决二次距离弯曲和多普勒频率徙动问题。针对这个问题,介绍了基于二阶keystone变换与匹配滤波相结合的方法,该方法可有效地解决目标做机动运动时导致能量扩散的问题,并且当目标回波发生多普勒模糊时也能够很好地对其能量做出相参积累。其次,还介绍了一种基于二维频域匹配滤波的方法,该方法是先将经过脉冲压缩后的目标回波信号变换到距离-多普勒二维频率域上,然后通过构造相应的补偿函数进行匹配滤波处理。这种方法只需要对加速度进行一维的搜索估计,就能够很好地消除因目标径向速度所导致的距离走动和因目标径向加速度所导致的距离弯曲和多普勒扩散等问题。最后,研究了一种基于子空间投影的多目标速度超分辨估计的方法,该方法首先对距离频域-方位时域的数据进行一阶Keystone变换,从而矫正目标信号的线性包络走动,然后在距离时域和方位时域对目标所处的距离单元数据利用前后项空间平滑法进行解相参处理,接着对前后项空间平滑得到的相关矩阵进行特征分解,把分解得到的噪声子空间向与目标速度有关的搜索导向矢量进行投影,最后再根据不同峰值对不同目标的速度进行超分辨估计。实验仿真结果表明,本方法可对位于同一距离单元的多个目标对应的径向速度进行有效的估计,从而提高天基雷达对径向速度差异较小的多个目标的探测能力。
[Abstract]:Prolonging the accumulation time can enhance the signal energy of the target and improve the radar detection performance. However, for high speed and high maneuvering targets, with the extension of time, the radar echo will obviously produce the phenomena of range walking and bending, Doppler frequency offset, diffusion and blur. If the range migration and Doppler migration of these targets can not be eliminated effectively, the energy can not be effectively accumulated. In this paper, aiming at the problems encountered by space-based radar in the process of detecting high maneuvering weak targets, several effective methods of energy coherent accumulation for long time targets are introduced. And how to estimate the velocity of different targets in multi-target environment is studied. Firstly, the linear frequency modulation signal and its pulse compression principle are introduced, and the echo signal model of high maneuvering weak target is constructed. The range migration and Doppler migration caused by the relative motion of the target are discussed by using the established model. Then the maximum likelihood method is introduced. The range migration and Doppler migration can be effectively eliminated by constructing matching function and designing appropriate Doppler frequency search interval and center frequency search interval. The method based on first-order keystone transform is analyzed. It can correct the range migration phenomenon of multiple targets without prior information of target, but it can not solve the problem of secondary range bending and Doppler frequency migration. In order to solve this problem, a method based on second-order keystone transform and matched filter is introduced. This method can effectively solve the problem of energy diffusion caused by maneuvering motion of target. And when the target echo occurs Doppler ambiguity, it can also make coherent accumulation of its energy. Secondly, a method based on 2-D frequency domain matched filtering is introduced, in which the target echo signal after pulse compression is first converted to the range-Doppler 2-D frequency domain. Then the matching filter is processed by constructing the corresponding compensation function. This method only needs to estimate the acceleration in one dimension, and it can eliminate the range movement caused by the radial velocity of the target and the range bending and Doppler diffusion caused by the radial acceleration of the target. Finally, a multi-target velocity super-resolution estimation method based on subspace projection is studied. Firstly, the first order Keystone transform is applied to the range frequency-azimuth time domain data to correct the linear envelope walk of the target signal. Then in the time domain and azimuth time domain, the distance unit data of the target is solved by the spatial smoothing method of the front and rear terms, and then the correlation matrix obtained by the smoothing of the front and rear terms is decomposed by the characteristic. The decomposed noise subspace is projected to the search guidance vector related to the target velocity, and then the super-resolution estimation of the velocity of different target is carried out according to the different peak value. The experimental results show that the proposed method can effectively estimate the radial velocities of multiple targets located in the same range cell, thus improving the detection ability of space-based radar for multiple targets with little radial velocity difference.
【学位授予单位】：西安电子科技大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：TN957.51

【参考文献】