摘要
在人体睡眠实时监测中,采集到的胸腔微动信号是一种包含呼吸、心跳信息的复杂混叠信号,识别和分离该混叠信号对分析睡眠质量具有重要意义。提出一种具有自适应性的胸腔微动信号实时分离和提取算法,使用两次基于固定"筛"数量的聚类经验模态分解(EEMD)进行信号分离,结合相关性分析方法进行信号识别,利用希尔伯特-黄变换(HHT)进行信号时频分析。进行了仿真分析和实验研究,结果均表明,该算法具有很强的自适应性,能够准确、高效的完成睡眠信号的实时处理。
In real-time sleep monitoring, the collected chest signal is a complex aliasing signal, which contains information of breath and heartbeat. The identification and the separation of the signal are of great significance for analyzing the quality of sleep. An adaptive algorithm used for the chest signal' s real-time separation and extraction was proposed in this paper, in which two times of the ensemble empirical mode decomposition (EEMD) based on fixed sifting number criterion and the correlation analysis method were used for the signal' s separation and recognition respectively, while the Hilbert-Huang transform (HHT) was used to analyze the time-frequency characteristics of the signal. The feasibility of the algorithm was proved by the simulation signal. The experimental data proves that the algorithm has strong adaptivity and it can process the real-time sleep signal accurately and efficiently.
出处
《电子测量与仪器学报》
CSCD
北大核心
2016年第10期1497-1505,共9页
Journal of Electronic Measurement and Instrumentation