稀疏系统辨识中最小均方算法的研究

Study of Least Mean Square Based Identification of Sparse Systems

长时间以来,将稀疏性应用于自适应滤波框架引起了很多人在理论和应用问题上的研究兴趣。由于稀疏系统的普遍存在,利用稀疏性特性改进辨识算法就具有重要的理论意义和实用价值。首先介绍基于最小均方算法的稀疏系统辨识的研究背景及研究现状,然后给出稀疏系统的辨识模型,最后重点介绍几种稀疏最小均方算法并指出今后进一步的研究方向。

Sparsity exploitation in adaptive filtering framework has attracted considerable research interest in both theoretical and applied issues for a long time. Sparse systems are ubiquitous in many scenarios, using the sparsity to improve the identification algorithm will have theoretical significance and practical value. The research background and the current situation of the least mean square（LMS）based identification of sparse systems was introduced in this paper. Then,the chan-nel model of sparse systems was developed. Further-more, we introduce a series of sparse LMS algorithm. At last,the further work was indicated.