机动目标参数估计方法研究

发布时间：2018-06-12 21:55

本文选题：机动目标 + 参数估计　；参考：《西安电子科技大学》2014年硕士论文

【摘要】：近年来,对复杂运动目标的检测和跟踪成为了雷达信号处理领域的前沿课题和紧迫任务。在一个短的雷达信号处理周期内,复杂运动目标可以一阶近似为匀加速度运动目标,其回波模型为线性调频信号。目标机动性将使雷达常规信号处理时,能量在多普勒域发生扩散,目标积累增益降低,分辨性能和信噪比下降,且机动性越大,目标能量越发散,检测性能越差。因此,在信号处理阶段研究对机动目标的径向初速度和加速度参数信息进行估计并提取,对改善机动目标检测和跟踪性能具有重要影响。论文围绕机动目标雷达信号参数估计方法进行了一些探索和研究。首先由匀速运动目标的回波特性推导得出机动目标的回波信号模型,并分析了机动目标加速度的存在对传统谱分析的性能影响,以便进一步研究机动目标雷达信号处理或补偿方法。本文主要介绍了典型的三种常规机动目标参数估计方法:应用最大似然思想的解线性调频法,时频分析方法中分数阶傅里叶变换法和拉冬-魏格纳变换法。首先详细介绍了这些常规方法的基本原理及主要内容;其次通过大量实验仿真并得出了不同参数对各个常规方法性能的影响,同时,简要概述了各方法的适用条件以及不足之处;最后重点分析了这些方法对处理多分量机动目标参数估计中存在的诸多不足以及这些方法的相互转换关系。针对常规方法处理多机动目标存在的局限性,本文结合最新的高分辨压缩感知理论和“分级逼近法”思想,提出了一种高分辨的多机动目标参数估计方法,它是利用CS理论和信号的稀疏特性建立超完备原子库,得到信号在超完备原子库上的分解系数投影,再通过1范数约束条件下的最优化算法和“分级逼近”的频率细化手段进行参数估计。本文简要描述了压缩感知系统的基本理论框架及主要内容,包括信号的稀疏表示、信号的观测采样、信号的重构等,并详细介绍了该方法的设计算法和流程,最后通过大量仿真实验验证了该方法能实现多机动目标的超分辨能力以及超高的参数估计精度。
[Abstract]:In recent years, the detection and tracking of complex moving targets has become the frontier and urgent task in the field of radar signal processing. In a short period of radar signal processing, the complex moving target can be approximated as a moving target with uniform acceleration, and its echo model is a linear frequency modulation signal. Target maneuverability will make the energy diffuse in the Doppler domain, decrease the target accumulation gain, decrease the resolution performance and SNR, and the greater the maneuverability, the more dispersed the target energy is and the worse the detection performance is. Therefore, in the signal processing stage, the estimation and extraction of radial initial velocity and acceleration parameters of maneuvering targets play an important role in improving the detection and tracking performance of maneuvering targets. This paper focuses on the method of radar signal parameter estimation for maneuvering target. Firstly, the echo signal model of maneuvering target is derived from the echo characteristics of moving target, and the influence of the acceleration of maneuvering target on the performance of traditional spectral analysis is analyzed. In order to further study the maneuvering target radar signal processing or compensation methods. This paper mainly introduces three typical methods of parameter estimation for maneuvering targets: the linear frequency modulation method using the maximum likelihood method, the fractional Fourier transform method and the Lato Wigner transform method in the time-frequency analysis method. Firstly, the basic principles and main contents of these conventional methods are introduced in detail. Secondly, the effects of different parameters on the performance of each conventional method are obtained through a large number of experiments and simulations, at the same time, The application conditions and shortcomings of these methods are briefly summarized, and the shortcomings of these methods in the estimation of multi-component maneuvering target parameters are analyzed. In view of the limitations of conventional methods in dealing with multiple maneuvering targets, this paper proposes a high-resolution method for estimating the parameters of multi-maneuvering targets by combining with the latest theory of high-resolution compression perception and the idea of "hierarchical approximation". It uses the CS theory and the sparse characteristic of the signal to establish the supercomplete atomic library, and gets the decomposition coefficient projection of the signal on the supercomplete atomic library. Then the parameters are estimated by the optimization algorithm under 1-norm constraint and the frequency thinning method of "hierarchical approximation". This paper briefly describes the basic theoretical framework and main contents of the compressed sensing system, including the sparse representation of the signal, the observation and sampling of the signal, the reconstruction of the signal, and introduces the design algorithm and flow chart of the method in detail. Finally, a large number of simulation experiments show that the method can achieve the super-resolution of multiple maneuvering targets and the ultra-high precision of parameter estimation.
【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

【相似文献】