| Sleep is essential to human beings,and it is of great significance to people’s psychological health and physical health.In recent years,with the increasing pressure of various aspects,more and more sleep disorders,sleep problems caused widespread concern.Sleeping stage is the basic for studying sleep disorders,through it we can know the process of sleep in great detail,which has important clinical significance.Traditional artificial methods for sleep staging is costly,laborious and low efficiency.According to the characteristics on different stage of sleeping EEG,we divide the sleep into distinct period.This paper studied the following contents:1.Since the EEG signal is weak,random and susceptible to noise,we selected the chip ADS1299 which is dedicated to bioelectric potential measurement with high precision,and low noise.The dry electrodes as sensors,and STM32F407VGT6 as the main control chip,collect EEG through a three-lead single channel approach,store the collected data in SD card and sent to the host computer via bluetooth.Use rechargeable lithium battery to power the hardware platform.2.Extract the seven kinds of rhythmic waves of sleeping EEG,including K-complex waves,spindle waves δ,θ,α,β-1 waves and β-2 waves by wavelet packet decomposition and reconstructing method after denoise the collected EEG signals pre-treated.Extract the energy characteristics of the rhythm wave,including the total energy of each sleep phase(30 seconds as a unit),and the ratio of the K-complex waves,the δ waves,the θ waves,the α waves,the spindle waves,the β1 waves and the β2 waves occupy the total energy,and Eα?Eθ、Eδ?Eθ、Eβ?Eθ,,11 kinds of energy feature in total,and the nonlinear dynamic characteristics of sleep EEG were extracted after pretreatment,including complexity,multi-scale entropy and fuzzy entropy.3.Using MIT sleep database to extract the variety of EEG features,and LS-SVM to train and test these features.The accuracy rate of sleep staging is 92.88%.Finally,staging test the sleep EEG which is collected on the sleep EEG acquisition platform. |