强混沌背景下弱谐波信号检测算法研究

发布时间：2018-06-13 14:52

本文选题：混沌 + 弱信号检测　；参考：《电子科技大学》2017年硕士论文

【摘要】：研究证明,很多工程问题可以归结为强混沌背景下的弱谐波信号检测问题。混沌理论因为其普适性和自身丰富的动力学特性而被广泛应用于弱信号检测领域。因此,强混沌背景下的弱谐波信号检测问题是一个研究热点。本文利用混沌理论结合信号检测技术对强混沌背景下的弱谐波信号检测算法展开研究,主要工作包括:1、研究了基于Duffing混沌振子方程的弱谐波信号检测算法,该算法利用混沌系统对初值比较敏感,以及混沌运动过程会随着Duffing方程的外加周期策动力的改变而呈现出不同相空间状态图的特点实现对弱谐波信号的检测。当在系统中加入含有微弱谐波信号的待检测信号以后,调节外加周期策动力,系统运动状态会从混沌状态进入大尺度周期状态,2、研究对比了三种均匀混沌背景下的弱谐波信号检测算法。其中,ESN检测算法利用参考序列训练回声状态网络,并构建混沌信号一步预测矩阵,最后利用一步预测误差检测弱谐波信号。然而该算法对高斯白噪声比较敏感,当混沌背景中混有高斯白噪声时,ESN检测算法失效。针对噪声敏感这一问题,本文提出了基于SOCP的最优化检测算法,该算法利用混沌信号具有二阶统计特性不变性的特点,利用参考序列构建数据矩阵提取混沌背景二阶统计特性,并设计最优滤波器,最后根据输出信干噪比(SINR)检测弱谐波信号。然而基于SOCP的最优化理论算法需要求解一个凸优化问题,算法计算复杂度较高。针对上述特点,本文进一步提出了一种基于EKF的迭代算法,该算法在传统EKF算法的基础上进行了优化,降低了计算复杂度。3、探索了非均匀混沌背景下弱谐波信号的检测算法,给出了一种基于APES的非均匀检测算法。非均匀检测算法不需要参考序列,首先构建Hankel矩阵提取自身混沌特性,并通过设计最优滤波器求解最优权值,最后通过输出信干噪比(SINR)检测弱谐波信号。仿真实验验证了非均匀检测算法可以检测均匀混沌背景下和非均匀混沌背景下的弱谐波信号,并具有良好的抗噪声性能。
[Abstract]:It is proved that many engineering problems can be attributed to weak harmonic signal detection under strong chaotic background. Chaos theory is widely used in weak signal detection field because of its universality and rich dynamic characteristics. Therefore, weak harmonic signal detection in strong chaotic background is a hot topic. In this paper, the weak harmonic signal detection algorithm based on duffing chaotic oscillator equation is studied by using chaos theory and signal detection technology. The main work includes: 1. In this algorithm, the chaotic system is sensitive to the initial value, and the chaotic motion process presents different phase space state diagrams with the change of the external periodic driving force of the duffing equation to realize the detection of weak harmonic signals. When the signal with weak harmonic signal is added to the system, the external periodic force is adjusted. The motion state of the system changes from chaotic state to large scale periodic state. Three weak harmonic signal detection algorithms in homogeneous chaotic background are studied and compared. The echo state network is trained by reference sequence, and the one-step prediction matrix of chaotic signal is constructed. Finally, the weak harmonic signal is detected by one-step prediction error. However, the algorithm is sensitive to the white noise of Gao Si. When the white noise is mixed with Gao Si in the chaotic background, the algorithm fails. In order to solve the problem of noise sensitivity, an optimal detection algorithm based on SOCP is proposed in this paper. The algorithm extracts the second-order statistical properties of chaotic background by constructing a data matrix of reference sequences, which makes use of the invariance of the second-order statistical properties of chaotic signals. The optimal filter is designed and the weak harmonic signal is detected according to the output signal-to-noise ratio (SINR). However, the optimization algorithm based on SOCP needs to solve a convex optimization problem, and the computational complexity of the algorithm is high. In view of the above characteristics, an iterative algorithm based on EKF is proposed in this paper. The algorithm is optimized on the basis of the traditional EKF algorithm, which reduces the computational complexity of .3. and explores the weak harmonic signal detection algorithm in the non-uniform chaotic background. A non-uniform detection algorithm based on apes is presented. The nonuniform detection algorithm does not need a reference sequence. Firstly, the Hankel matrix is constructed to extract its own chaotic characteristics, and the optimal weight is solved by designing an optimal filter. Finally, the weak harmonic signal is detected by output signal-to-noise ratio (SINR). The simulation results show that the non-uniform detection algorithm can detect the weak harmonic signal in the homogeneous chaotic background and the non-uniform chaotic background, and has good anti-noise performance.
【学位授予单位】：电子科技大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TN911.23

【相似文献】