基于体动射频信号的睡眠分期

为了减轻传统接触式睡眠生理监测系统对人体造成的负担,设计了一种基于微波技术的非接触式睡眠生理信号采集与分析系统,提出一种基于体动射频信号的睡眠分期识别算法.通过小波变换对射频运动传感器(RFMS)采集的体动信号进行预处理,再计算出体动信号的能量值,最后通过判别式处理和阈值法实现了睡眠分期:醒觉期、浅睡期、中睡期、深睡期.实验采集分析了8个实验者为期46天的睡眠生理信号,同时同步采集视频信息、TANITA水床睡眠信息、接触式呼吸脉搏信号.与视频结果比较发现醒觉期正确率达到90%;与TANITA水床睡眠结果相比,本系统的结果与其吻合程度达到70%;与不同睡眠状态下呼吸率、心率的变化相比,本系统的结果吻合度达到80%.

Identification of Sleep Stages Based on Body Movement Signals from RFMS

To reduce the burden of the traditional sleep monitoring system on human body， a non-contact sleep monitoring system based on microwave technique was developed， and a sleep stages recognition algorithm was proposed based on body movement signals from radio frequency motion sensors (RFMSs).The RFMS signals were processed by wavelet transform， then， the energy of signals was calculated by integral algorithm. Finally， the sleep stages such as wakefulness， light sleep， moderate sleep， deep sleep were recognized by discriminant analysis and threshold processing. The RFMS signals from 8 normal subjects， whose sleep were monitored for 46 days， were acquired， with their video information， TANITA sleep information， pulse wave， and respiration signals gathered synchronously. Compared with videos， the staging results of this system match 90% in wakefulness stage. In comparison with the results of TANITA， the results of the RFMS match 70% in other sleep stages.The results match 80% in comparison with the change modes of respiratory rate and heart rate at different sleep stages.