摘要
提出了一种基于遗传算法的盲源分离算法,该算法直接从信号的样本序列估计信号的概率分布,解决了信号间互信息的求解问题,通过遗传算法最小化信号的互信息,实现对线性混叠信号的分离。将该算法用于轴承声音信号的分离,取得了较好的效果。
A new blind source separation algorithm based on the genetic algorithm is proposed, in which the probability distribution of the separated signals can be directly estimated from the sample signals, so the mutual entropy solution problem can be solved, and through mutual entropy of minimizing signals, the separation of the linear mixture signals can be realized. The genetic algorithm can be used to separate the sound signals of the bearing,achieving better results.
出处
《轴承》
北大核心
2005年第9期31-34,共4页
Bearing
基金
国家自然科学基金重点资助项目(50335030)
关键词
滚动轴承
诊断
盲源分离
遗传算法
互信息
rolling bearing
diagnosis
blind source separation
genetic algorithm
mutual entropy
