基于多测量矢量模型的压缩感知雷达时延-多普勒估计

摘要雷达的时延-多普勒估计是雷达研究的重要问题之一。压缩感知（Compressed Sensing, CS）雷达可以用低于奈奎斯特的采样率对雷达回波进行采样，并通过稀疏重构算法估计时延和多普勒信息。当前，压缩感知雷达时延-多普勒估计方法大部分是基于单测量矢量模型(Single Measurement Vector, SMV)模型，具有耗时长、受噪声干扰大等缺点。而在实际雷达应用中，雷达以一定的脉冲重复周期发射连续的脉冲串信号，雷达回波不仅具有稀疏性，还具有一定的内在结构，更加符合压缩感知多测量矢量（Multiple Measurement Vector, MMV）模型。因此本文研究基于MMV模型的压缩感知雷达时延-多普勒估计。本文首先利用MMV模型恢复具有相同稀疏结构的系数矩阵，根据该矩阵每列矢量非零幅值所在位置估计目标时延，继而对估计出的非零幅值做DFT变换来估计多普勒频率值。本文通过仿真实验分析基于MMV模型的压缩感知雷达的时延-多普勒估计性能，仿真结果表明基于MMV模型的时延-多普勒估计方法在耗时上、估计精度上均优于基于SMV模型的估计方法。28614
关键字：压缩感知；单测量矢量模型；多测量矢量模型；时延估计；多普勒估计
毕业论文设计说明书外文摘要
Title Time Delay and Doppler Estimation of Compressed Sensing Radar
Based on Multiple Measurement Vectors Model
Abstract
Time delay and Doppler estimation is important for radar applications. Compressed sensing (CS) radar can acquire the radar echo with the sub-Nyquist sampling and estimate time delay and Doppler by using sparse reconstruction algorithms. Most of time delay and Doppler estimation for compressed sensing radar is conducted by single measurement vector (SMV) model. However, the SMV model used to estimate the time delay and Doppler information is time-consuming and noise-affected. In radar applications, radar emits continuous pulse train signal with certain pulse repetition interval. The acquired radar echo signal is not only sparse but also has some inherent structure. It is more suitable to use the Multiple Measurement Vector (MMV) model-based method. In this paper, we apply the MMV model to the time delay and Doppler estimation for compressive sensing radar. According to the MMV model, we firstly recover the coefficient matrix with the same sparse structure by using the MMV-based method. Then the time delay can be estimated by choosing positions of the nonzero coefficients. After that, we estimate Doppler by performing DFT transform on the nonzero coefficients. The performance of the MMV-based time delay and Doppler estimation is analyzed by experiment. It is shown that proves that the estimation based on MMV is superior to that based on SMV both on time-consuming and estimation accuracy.
Keywords: Compressed Sensing(CS); Single Measurement Vector(SMV)；Multiple Measurement Vector(MMV)；Time Delay Estimation；Doppler Estimation
目次
1 引言1
2 压缩感知模型3
2.1 压缩感知理论的SMV模型3
2.2 压缩感知理论的MMV模型4
3 雷达信号正交压缩采样处理8
3.1 雷达信号模型8
3.2 正交压缩采样9
4 基于MMV模型的雷达时延-多普勒估计 12
4.1 多脉冲雷达信号的MMV模型12
4.2 基于MMV模型的时延-多普勒估计 13
5 仿真15
5.1 仿真环境和参数设置15
5.2 估计性能15
结论 27
致谢 28
参考文献 29
1 引言
时延-多普勒估计[1,2]是雷达信号处理中目标检测和目标信息提取等一系列问题的重要组成部分，其本质是依据接受到的目标信号，快速、精准地估计出接收信号相对于基准时间的延迟以及因为该目标径向运动而产生的多普勒频移。基于多测量矢量模型的压缩感知雷达时延-多普勒估计:http://www.751com.cn/tongxin/lunwen_23532.html