跳频信号参数估计相关技术研究

发布时间：2018-03-10 13:18

本文选题：跳频信号　切入点：参数估计　出处：《解放军信息工程大学》2014年硕士论文　论文类型：学位论文

【摘要】：跳频通信由于具有良好的抗干扰、抗截获及较强的组网能力,已经在军事和民用通信领域中得到了广泛的应用,同时也对第三方的通信监测提出了严峻的挑战。跳频信号参数估计作为跳频信号分析处理的主要任务之一,具有十分重要的现实意义。本文就跳频信号参数估计的相关技术展开深入研究,具体工作及创新点如下:1、研究了基于时频脊线的跳频信号参数估计方法。首先,针对传统时频分析复杂度较高的问题,给出了一种基于盒维数的跳周期快速估计方法,该方法利用盒维数与跳频频率正相关的规律,采用盒维数曲线代替时频脊线,将跳周期估计转化为盒维数变化周期的估计。仿真表明相对于短时傅里叶变换(STFT),该方法以较小的性能损失换取了计算量上的较大优势;然后,针对STFT频率估计精度低的缺点,给出了一种基于滑动旋转不变技术(ESPRIT)的跳频频率估计方法,该方法利用ESPRIT超分辨的优势提取时频脊线,提高了跳频频率的估计精度。2、研究了基于自混频的跳频信号跳周期和跳时估计方法。针对跳周期估计,由延时共轭相乘构造自混频信号,理论分析了低频分量、延迟时间及跳周期的关系。然后根据低频分量曲线拐点的位置估计跳周期,给出了一种基于延时自混频的跳周期估计方法;针对跳时估计,给出了一种基于分段自混频的跳时估计方法,按照不同偏移时间对信号进行分段自混频,根据低频分量、偏移时间及跳时的关系,得出了低频分量最大值对应的偏移时间含有跳时信息的结论。仿真结果表明上述两种方法均能有效估计出跳周期或跳时,且都适用于多跳频信号的情况。3、研究了跳频信号的实时跟踪问题。首先分析了以跳频频率为系统状态的状态空间模型,并引入粒子滤波算法对频率进行跟踪。针对跟踪性能欠佳的问题,给出了一种基于ESPRIT辅助的改进方法,通过ESPRIT算法为粒子更新提供参考信息,提高了频率跟踪性能。对于粒子滤波不适用于多跳频信号的情况,研究了一种基于稀疏重构的多跳频信号频率跟踪及DOA估计方法。该方法利用跳频信号在频域和空域的稀疏性,建立基于阵列接收的信号稀疏表示模型,采用稀疏贝叶斯学习(SBL)算法进行模型求解,通过频率估计和跳变时刻检测完成多跳频信号频率的实时跟踪和DOA估计。仿真结果验证了该方法的有效性。
[Abstract]:Frequency hopping communication has been widely used in military and civil communication fields because of its good anti-jamming, anti-interception and strong networking capability. At the same time, it also poses a severe challenge to the third party's communication monitoring. Estimation of frequency hopping signal parameters is the main factor in frequency hopping signal analysis and processing. It has very important practical significance. In this paper, the related techniques of frequency-hopping signal parameter estimation are deeply studied. The specific work and innovation are as follows: 1. The method of frequency hopping signal parameter estimation based on time-frequency ridge is studied. In order to solve the problem of high complexity in traditional time-frequency analysis, a fast estimation method of hopping period based on box dimension is presented. The method uses the law of positive correlation between box dimension and frequency hopping frequency, and uses box dimension curve instead of time-frequency ridge. The hopping period estimation is transformed into the estimation of the variation period of the box dimension. The simulation results show that this method gains a large advantage in computation with small performance loss compared with the STFT method, and then, in view of the disadvantage of low accuracy of STFT frequency estimation, A frequency hopping frequency estimation method based on sliding rotation invariant technique (Esprit) is presented. This method uses the advantage of ESPRIT super-resolution to extract time-frequency ridges. The precision of frequency hopping frequency estimation is improved. The frequency hopping period and time hopping estimation method based on self-mixing is studied. For the frequency hopping estimation, the low frequency component is theoretically analyzed by using the time-delay conjugate multiplication to construct the frequency hopping signal. The relationship between the delay time and the hopping period. Then, according to the position of the inflexion point of the low-frequency component curve, a method for estimating the hopping period based on the time-hopping self-mixing is presented. A method of time hopping estimation based on piecewise self-mixing is presented. According to the relationship of low-frequency component, offset time and hopping time, the signal is segmented self-mixing according to different offset time. It is concluded that the offset time corresponding to the maximum value of the low-frequency component contains time hopping information. The simulation results show that the two methods mentioned above can effectively estimate the hopping period or the hopping time. In this paper, the real-time tracking problem of frequency-hopping signals is studied. Firstly, the state space model with frequency-hopping frequency as the system state is analyzed. Aiming at the problem of poor tracking performance, an improved method based on ESPRIT is presented, which can provide reference information for particle updating through ESPRIT algorithm. The performance of frequency tracking is improved. For the case that particle filter is not suitable for multi-frequency hopping signals, a method of frequency tracking and DOA estimation for multi-frequency hopping signals based on sparse reconstruction is studied. The method makes use of the sparsity of frequency-hopping signals in frequency domain and spatial domain. The sparse representation model based on array reception is established, and the sparse Bayesian learning algorithm is used to solve the model. The real-time frequency tracking and DOA estimation of multi-frequency hopping signals are accomplished by frequency estimation and jump time detection. The simulation results show that the proposed method is effective.
【学位授予单位】：解放军信息工程大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：TN914.41

【参考文献】