摘要
从独立分量分析的角度探讨了共址干扰抑制问题,并提出了基于互信息最小准则的干扰抑制算法。该算法利用了信号的高阶统计信息,兼顾了复杂度和精度,并且能够有效地抑制非高斯性干扰。为使算法稳定,对权值迭代公式进行了归一化处理;给出了新的共址干扰抑制滤波器结构;分析了算法稳定收敛的充分条件。仿真实验表明:该算法能够很好地将弱有用信号从强干扰信号中分离。
Abstract. The co-located interference suppression is discussed based on the principle of independent component analysis (ICA). The interference suppression algorithm based on the criterion of minimiza- tion of mutual information is proposed. This algorithm utilizes the higher-order statistics information of signals and makes a trade-off between complexity and accuracy; it is effective to suppress the non- Gaussian interference as well. In order to stabilize the algorithm, the iterative equation of the weights is normalized. A new structure of co-located interference suppression filter is proposed. The sufficient condition of the stable convergence of the algorithm is analyzed. The simulation results indicate that the weak signal of interest could be separated from the strong interference by using this algorithm.
出处
《海军工程大学学报》
CAS
北大核心
2016年第4期16-20,84,共6页
Journal of Naval University of Engineering
基金
国家自然科学基金资助项目(51422705
61201055)
关键词
独立分量分析
互信息
共址干扰抑制
收敛条件
independent component analysis
mutual information
co-located interference suppression
convergence condition