基于多外辐射源的被动雷达目标检测算法研究

发布时间：2018-06-10 11:36

本文选题：被动雷达 + 雷达组网　；参考：《中国科学技术大学》2017年硕士论文

【摘要】：被动雷达是使用空间环境中已有的军用或民用设备发射的信号而自身不发射信号的雷达系统。被动雷达系统通常由监督通道和参考通道两部分构成,监督通道用来接收目标反射信号,参考通道则接收外辐射源的直达波信号。被动雷达系统具有反隐身,抗反辐射导弹,低空探测,以及低成本,灵活性大等诸多特点。但是,目前的被动雷达系统多基于单个外辐射源信号进行空间目标的检测、定位和追踪,而对联合多种不同的外辐射源信号进行目标探测的研究甚少。基于多个外辐射源信号的被动雷达系统能够从不同的角度对目标进行照射,这将增大空间分集增益,避免因目标反射截面积闪烁带来的检测性能下降。本文基于利用多个外辐射源信号进行探测的被动雷达系统,进行了以下研究:1.在多个外辐射源信号非正交的条件下,使用单个监督通道进行目标检测。在假设监督通道里来自外辐射源的直线信号都已经被完全滤除,只有带噪的目标反射信号存在时,基于广义似然比准则,对未知的反射系数进行了极大似然估计,在噪声功率已知和未知两种条件下构造了对应的检测器。并且,在噪声功率已知的条件下对检测器性能进行了理论性能分析。通过实验发现,检测器能获得较好的检测性能和鲁棒性。2.对于真实情况下监督通道里存在的直线信号,研究基于梯度自适应的格形滤波器对直线干扰信号的滤除效果。利用参考通道信号来学习监督通道里的直线信号,在工作过程中将监督通道里的直线信号消除,进而进行检测。实验显示,在外辐射源数目较小时,能够有比较好的滤波效果,但是对于多个外辐射源混合的直线信号,滤波性能将下降。3.针对多个外辐射源信号联合使用时,监督通道直线信号滤波困难的问题,提出改进的信号模型。基于多个监督通道和参考通道信号的接收系统,在监督通道的接收信号模型中,将直线信号成分保留,利用广义似然比准则联合参考通道和监督通道对外辐射源信号进行估计,并对未知的反射系数进行估计,得到两种噪声功率下的检测器。并且对噪声功率已知条件下的检测器进行了理论性能分析。实验结果表明,当将直线信号视作一种强干扰,在模型中保留该成分,而不是通过滤波的方法进行消除时,所提出的检测器能够达到等价的滤波条件下很好的检测效果,并且具有更好的鲁棒性。
[Abstract]:Passive radar is a radar system which uses signals transmitted by military or civil equipments in space environment but does not transmit signals by itself. Passive radar system is usually composed of two parts: the supervised channel and the reference channel, which are used to receive the target reflected signal, and the reference channel to receive the direct wave signal from the external emitter. Passive radar system has many characteristics, such as anti stealth, anti radiation missile, low altitude detection, low cost, high flexibility and so on. However, most of the passive radar systems are based on single external emitter signal to detect, locate and track the space target. However, there is little research on target detection based on the combination of a variety of external emitter signals. The passive radar system based on multiple external emitter signals can irradiate the target from different angles, which will increase the spatial diversity gain and avoid the deterioration of detection performance caused by the reflection cross-sectional area flicker of the target. Based on the passive radar system which uses multiple external emitter signals to detect, the following research work is done: 1: 1. Under the condition that multiple external emitter signals are not orthogonal, a single supervised channel is used for target detection. Assuming that the linear signals from the external emitter in the supervised channel have been completely filtered out, the unknown reflection coefficient is estimated with maximum likelihood based on the generalized likelihood ratio criterion when only the noisy target reflection signal exists. The corresponding detector is constructed under two conditions: the noise power is known and the noise power is unknown. Furthermore, the theoretical performance of the detector is analyzed under the condition of known noise power. The experimental results show that the detector can obtain better detection performance and robustness. 2. The filtering effect of gradient adaptive lattice filter on linear interference signal is studied. The reference channel signal is used to learn the straight line signal in the supervised channel, and the linear signal in the supervised channel is eliminated in the working process, and then the line signal in the supervised channel is detected. The experimental results show that the number of external radiation sources is small and the filtering effect is better, but the filtering performance of the mixed linear signals of multiple external emitter sources will be decreased by .3. An improved signal model is proposed to solve the problem of the difficulty of filtering the supervised channel linear signals when the signals of multiple external emitter sources are used together. Based on the receiving system of multiple supervised channels and reference channel signals, in the received signal model of the supervised channel, the components of the linear signal are preserved, and the external emitter signals of the reference channel and the supervised channel are estimated by using the generalized likelihood ratio criterion. The unknown reflection coefficient is estimated and the detector under two kinds of noise power is obtained. The theoretical performance of the detector with known noise power is analyzed. The experimental results show that when the linear signal is regarded as a strong interference and the component is retained in the model rather than eliminated by filtering, the proposed detector can achieve a good detection effect under the condition of equivalent filtering. And it has better robustness.
【学位授予单位】：中国科学技术大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TN957.51

【相似文献】