Research And Implementation Of Sleep Monitoring System Based On Electrocardiogram And Respiratory Signals

As the pace of life gradually accelerates and the pressure increases day by day,more and more people have sleep problems in varying degrees.Polysomnography is the gold standard for evaluating sleep quality.By monitoring various physiological signals such as EEG,electrooculogram,electrocardiogram,and breathing,it can comprehensively evaluate the sleep status throughout the night.However,polysomnography has unavoidable shortcomings.First,it must be completed in a hospital.Second,it needs to wear a variety of physiological monitoring equipment during the monitoring process,which seriously affects the sleep quality of the subjects.To this end,this thesis designs and implements a portable sleep monitoring system based on ECG and respiratory signals for home monitoring application scenarios,and conducts research on sleep staging algorithms.The specific work of this thesis is as follows:1)Sleep monitoring system hardware design and implementation.Specifically,it includes the design of the main control module,the design of the ECG acquisition module,the design of the respiratory acquisition module,the design of the data storage and transmission module,and the design of the power management module.For each part of the hardware modules,the device selection,simulation design and circuit implementation were carried out.The PCB layout and wiring between each hardware module were rationally optimized to reduce mutual interference between each module and realize the miniaturization of the equipment.The real thing is presented in the form of a chest strap,which supports two-electrode collection and has good portability.2)Software design and implementation of sleep monitoring system.The system software mainly includes two parts,embedded and PC.Among them,the embedded terminal is mainly responsible for completing ECG and respiratory sensor data collection,data storage and data transmission,and the PC terminal is mainly responsible for completing data visualization,data processing and subsequent sleep staging.Program implementation of the algorithm.3)Research and implementation of sleep staging algorithm based on deep learning.A sleep staging model based on convolutional neural network and long-term shortterm memory network was designed,and data preprocessing,model training,and evaluation were completed using the Sleep Heart Health Study dataset.The evaluation results showed that it was cdomparable to domestic Compared with the external correlation model method,it has good classification performance.4)Testing and validation of sleep monitoring systems.Firstly,with reference to the JJG 1041-2008 digital ECG machine verification regulations,the performance test is carried out for the system functional hardware,and the results show that the system input voltage range,polarization withstand voltage,input noise level,common-mode rejection ratio and power consumption meet practical requirements,among which the input noise level is less than 20μVpp and the common-mode rejection ratio is greater than 86 d B.Secondly,the signal acquisition test shows that the communication works normally and the signal quality meets the requirements; Finally,the current mainstream portable commercial products with sleep monitoring function were selected for comparative verification of sleep staging function,and the results showed that the sleep staging results of 10 people had good consistency.