基于瞬时频率估计的抗窄带干扰技术研究

In spite of the nonstationarity inherent in many signals of practical interest and the frequency is fast time-varying over a short data record, most of the initial work on instantaneous frequency estimation and modeling of nonstationary signals is based on the assumption that the signal is stationary or only slowly time-varying over a short data record. Under such assumptions, it is possible to use stationary sinusoidal frequency estimation techniques or adaptive algorithms to estimate the instantaneous frequency. Also, when the narrowband interference is nonstationary, the interfering frequency can vary rapidly over short data records and narrowband interference mitigation based on stationarity frequency estimation cannot be used to effectively mitigate the interference. And the traditional adaptive algorithm to do the same may not converge fast enough to track the rapidly varying frequency.. In this project, the modeling of nonstaionary signals will be achieved through time-varying autoregressive (TVAR) model by the use of basic functions. And based on TVAR model, the method for the estimation of parameters will be analyzed over a short data record. Then, we will analyze the influence on the iteration direction of the normalized least mean square with orthogonal correction factors algorithm under the narrowband interference. Over a short data record, we will analyze the new iteration direction for the adaptive filtering and construct a new adaptive algorithm that has a good narrowband interference rejection, and the convergence will be improved. Also, the statistical convergence and tracking behaviors of the proposed algorithm will be given. At last, we will construct the two-stage filter structure, where the main role of the first stage filter mitigate the narrowband interference based on the parameters of the TVAR model, and the second stage filter perform mainly channel equalization based on the balanced reduced order model.. The research of this project will further expand the concept of narrowband anti-interference in communication systems. It will achieve the development theory of the modeling of nonstationary signals, instantaneous frequency estimation, communication countermeasures.

在非稳定的窄带干扰环境下，信号频率在较短的数据记录中快速变化，基于稳定信号模型的瞬时频率估计方法不能够有效的消除窄带干扰，同时传统的自适应方法也不能够快速跟踪频率的变化。本项目拟基于时变自回归模型，在较短信号的数据记录中，分析基函数同以特征值定义的“模态”的关系，建立时变自回归模型参数的估计方法，实现瞬时频率的估计。分析非稳定窄带干扰信号对带有正交化校正因子的归一化最小均方算法迭代方向的影响，研究自适应滤波器新的迭代方向及其迭代方向上的误差，建立一种抗窄带干扰的自适应方法，并分析其收敛性和跟踪性的随机统计模型。研究构建二级滤波结构，其中第一级基于估计的时变自回归模型参数实现对非稳定窄带干扰信号的抑制，第二级利用平衡降维结构实现对无线信道的自适应均衡。本项目的研究成果将进一步拓展无线通信系统中抗窄带干扰装置的设计方法，对非稳定信号建模、瞬时频率估计、无线通信对抗等方面理论的发展具有重要意义。

在非稳定的窄带干扰环境下，干扰信号的频率在较短的雷达、无线通信等系统主瓣数据记录中快速变化，基于稳定信号模型的瞬时频率估计方法不能够有效的消除窄带干扰，同时传统的自适应方法也不能够快速跟踪频率的变化，这是国内一个典型的“卡脖子”技术。本项目基于时变自回归模型，在较短信号的雷达、无线通信等系统主瓣数据记录中，分析了基函数同以特征值定义的“模态”的关系，建立了时变自回归模型参数的估计方法，实现了非平稳干扰信号瞬时频率的估计。分析了非稳定窄带干扰信号对带有正交化校正因子的归一化最小均方算法迭代方向的影响，研究了自适应滤波器新的迭代方向及其迭代方向上的误差，建立了一种抗窄带干扰的自适应方法，并分析了其收敛性和跟踪性的随机统计模型。研究并构建了一种二级滤波结构，其中第一级基于估计的时变自回归模型参数实现对非稳定窄带干扰信号的抑制，第二级利用平衡降维的自适应滤波结构实现对无线信道的自适应均衡。本项目的研究成果进一步拓展了无线通信、雷达等系统中抗窄带干扰装置的设计方法，对非稳定信号建模、瞬时频率估计、无线通信对抗等方面理论的发展具有重要意义。本项目研究成果利用某型号的实际数据进行了验证，说明了所建立方法的有效性。