摘要
独立分量算法是一种应用非常广泛的盲信号处理算法。而峭度作为一种重要的信号分析工具,可以有效地进行优化分析。然而,对于各种不同类型的算法的对比分析目前还少有介绍,所以有必要对基于峭度的FastICA和RobustICA两种独立分量算法进行对比分析研究。理论分析及实验结果表明,鲁棒独立分量法RobustICA在鲁棒性、收敛性和复杂度方面整体优于快速定点独立分量法FastICA,从而为实际应用提供一定的参考价值。
Independent component analysis algorithm is a widely used algorithm for blind signal processing. As an important signal analysis tool, kurtosis can be effective for optimization and analysis. However, there is little introduction to the comparison and analysis for various types of algorithms currently. Therefore, it is necessary to research and analyze the comparison between FastlCA and RobustlCA based on kurtosis. The theoretical analysis and the simulation results indicate that the robustness, convergence and complexity of RobustlCA are better than that of FastlCA on the whole. Thus it provides the reference for practical applications.
出处
《四川理工学院学报(自然科学版)》
CAS
2014年第4期43-47,共5页
Journal of Sichuan University of Science & Engineering(Natural Science Edition)
基金
四川省杰出青年基金项目(2011JQ0034)
四川省省属高校科研创新团队建设计划基金项目(13TD0017)
人工智能四川省重点实验室基金项目(2012RYJ05)