基于时频分析的跳频信号盲侦察算法研究

发布时间：2019-02-24 20:07

【摘要】：跳频通信系统是扩频通信系统的一种实现方式,难以被识别干扰,是一种常用的军事通信技术。随着跳频通信系统的性能指标越来越高,针对跳频通信进行电子侦察的难度也越来越大,因此迫切需要研究新的跳频信号盲侦察算法。本文以时频分析为基础,研究了针对同步正交组网方式的跳频信号盲侦察算法。本文简要介绍了跳频通信和时频分析的原理,研究了跳频信号的非平稳特性。针对跳频信号的时频特征,基于聚类的方法,提出了一种跳频信号的盲检测算法。仿真结果表明,本算法能够有效地从混有干扰信号的接收信号中提取跳频信号。针对跳频信号的盲源分离问题,本文重点研究线性延迟混合模型。在欠定条件下,本文采用稀疏分量分析算法,首先估计与跳频信号混合过程相关的时频矩阵,然后根据时频矩阵分离跳频信号。本文通过检测单源点,以时频比算法为基础估计时频矩阵;然后估计源信号的入射方向,对时频矩阵进行重排;在源信号分离中,采用改进的子空间投影分离算法,通过改善算法流程,放宽了稀疏假设条件,增大了算法的适用范围。本文主要对跳频信号的跳频载波频率以及跳频速率进行估计,并设计了两种不同的参数盲估计算法。同时结合理论推导和性能仿真,本文分析了两种算法的性能特点与适用范围。本文在FPGA开发平台上,对跳频信号的盲侦察算法进行了硬件实现。在FPAG内部搭建了嵌入式软核,采用C语言开发,加速了开发进度。测试结果表明,该系统能够实现基本的参数估计功能并满足预设的性能指标。
[Abstract]:Frequency-hopping communication system is a kind of realization method of spread spectrum communication system, which is difficult to identify interference and is a commonly used military communication technology. With the performance of FH communication system becoming higher and higher, it is more and more difficult to conduct electronic reconnaissance for FH communication. Therefore, it is urgent to study a new blind FH signal reconnaissance algorithm. Based on time-frequency analysis, a blind reconnaissance algorithm of frequency hopping signal for synchronous orthogonal network is studied in this paper. This paper briefly introduces the principle of frequency hopping communication and time frequency analysis, and studies the non-stationary characteristic of frequency hopping signal. According to the time-frequency characteristics of frequency-hopping signals, a blind detection algorithm for frequency-hopping signals is proposed based on clustering method. The simulation results show that the proposed algorithm can extract frequency hopping signals from the received signals mixed with interference signals. In order to solve the problem of blind source separation for frequency hopping signals, this paper focuses on the linear delay mixing model. Under underdetermined conditions, a sparse component analysis algorithm is used to estimate the time-frequency matrix associated with the mixing process of frequency-hopping signals, and then to separate the frequency-hopping signals according to the time-frequency matrix. In this paper, the time-frequency matrix is estimated based on the time-frequency ratio algorithm by detecting the single source point, then the direction of incident signal is estimated, and the time-frequency matrix is rearranged. In source signal separation, an improved subspace projection separation algorithm is adopted. By improving the algorithm flow, the sparse assumption condition is relaxed and the applicable range of the algorithm is enlarged. In this paper, the frequency hopping carrier frequency and the frequency hopping rate of the frequency hopping signal are estimated, and two different blind estimation algorithms are designed. At the same time, combined with theoretical derivation and performance simulation, this paper analyzes the performance characteristics and applicable range of the two algorithms. In this paper, the blind reconnaissance algorithm of frequency hopping signal is implemented on FPGA platform. The embedded soft core is built in FPAG, and C language is used to develop it, which speeds up the development progress. The test results show that the system can realize the basic function of parameter estimation and meet the preset performance index.
【学位授予单位】：哈尔滨工业大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TN914.41

【参考文献】