量子人工蜂群优化的盲源分离算法
Blind Source Separation Algorithm Based on Quantum Artificial Bee Colony Optimization
程静 1王荣杰2
作者信息
- 1. 集美大学轮机工程学院,福建 厦门 361021
- 2. 集美大学轮机工程学院,福建 厦门 361021;福建省船舶与海洋工程重点实验室,福建 厦门 361021
- 折叠
摘要
为了实现分离多种服从不同分布类型的源信号,将一种改进的量子人工蜂群方法用于优化盲源分离算法.在标准量子人工蜂群算法的基础上,引入混沌优化算子生成初始解,使初始种群的解均匀分布在可行解空间上;在搜索阶段引入动态的邻域因子和遗忘因子,控制寻优方向,提高算法的收敛速度和寻优能力;以信号峰度构造目标函数,利用改进的量子人工蜂群方法对目标函数寻优,获得分离矩阵,实现混合信号的分离.仿真结果表明,所提算法能够分离亚高斯分布、超高斯信号及两者的混合信号,且在收敛速度和分离精度上均优于传统算法.
Abstract
In order to achieve the separation of source signals subject to arbitrary distribution,an improved quantum artificial bee colony method was proposed for optimizing the blind source separation algorithm.First,on the basis of the standard quantum artificial bee colony algorithm,a chaotic optimization operator was intro-duced to generate the initial solution,so that the solutions of the initial population were uniformly distributed on the feasible solution space;Second,dynamic neighborhood factor and forgetting factor were introduced in the search stage to control the optimization direction,improving the convergence speed and optimization ability;Finally,the objective function was constructed based on signal kurtosis,and the separation matrix was obtained by optimizing the objective function using the improved quantum artificial bee colony method and hence one could realize the separation of mixed signals.The simulation results showed that the proposed algorithm was able to separate sub-Gaussian distribution,super-Gaussian signal and the mixed signal of both,and it outper-forms the traditional algorithm in terms of convergence speed and separation accuracy.
关键词
盲源分离/量子人工蜂群算法/峰度/超高斯分布/亚高斯分布Key words
blind source separation/quantum artificial bee colony optimization/kurtosis/super-Gaussian distribution/sub-Gaussian distribution引用本文复制引用
出版年
2024
国家科技期刊平台
NSTL
万方数据