四次样条插值的改进Fast-ICA算法  

Improved Fast-ICA algorithm based on quadruple spline interpolation

作　　者：郭松林 丘斯帆 Guo Songlin;Qiu Sifan(School of Electrical&Control Engineering,Heilongjiang University of Science&Technology,Harbin 150022,China)

机构地区：[1]黑龙江科技大学电气与控制工程学院,哈尔滨150022

出　　处：《黑龙江科技大学学报》2023年第6期917-921,共5页Journal of Heilongjiang University of Science And Technology

摘　　要：为提升盲源分离算法的性能和精度,将传统Fast-ICA算法中的二阶牛顿迭代法改进为四阶收敛的牛顿迭代法,对其进行四次样条插值,构造四次样条插值函数,通过误差分析验证该函数的收敛特性,采用Matlab软件仿真分析原算法与改进算法分离同一组信号的能力。结果表明,改进四阶收敛的牛顿迭代法具有四阶收敛特性。改进算法与原算法相比精度和收敛速度分别提升了24%和35%。This paper aims to improve the performance and accuracy of the blind source separation algorithm.The study involves improving the second-order Newton iteration method of the traditional Fast-ICA algorithm to the fourth-order convergence Newton iteration method with the four-spline interpolation,and constructing four-spline interpolation function;verifying the convergence characteristics of the function through error analysis;simulating and analyze the strength of the original algorithm and the improved algorithm by using Matlab software.The results show that the improved fourth-order convergence Newton iteration method has the fourth-order convergence with the increase of the accuracy and convergence rate by 24% and 35% respectively.

关 键 词：盲源分离 Fast-ICA 四次样条插值 

分 类 号：TN911[电子电信—通信与信息系统]