Sleeping Posture Monitoring And Sleep Quality Analysis Based On BCG Signals In Smart Medical Systems

Sleeping posture is a significant indicator for the diagnosis and treatment of the position-related sleep disorder,cardiovascular disease,and the identification and adjustment of sleeping posture is of great importance to improve the sleep apnea hypopnea index of patients with positional obstructive sleep apnea syndrome.The traditional sleep monitoring systems rely on professional medical equipment and personnel,and have the disadvantage of high cost and strong professional operation which are difficult to popularize household daily use.Therefore,it requires a convenient,human-free sensory and low-cost sleeping system to monitor continuously and automatically the posture and quality during sleeping in daily life.This paper focuses on the impact of different sleeping posture on the characteristics of BCG signals based on a non-invasive sleep monitoring device constructed by the smart medical system,and researches on the determination of sleeping posture and sleep monitoring through BCG.The non-contact system has the inherent shortcoming of weak stability,and the collected BCG signal strength is low and it is susceptible to noise,body movement and other disturbance.Hence,recognizing sleeping posture by the BCG signal characteristics has certain challenge.In addition to the BCG signal,the collected signal also contains mixed signals of body movement and environmental noise.this paper proposes a noise reduction method of an adaptive Butterworth band-pass filter which adaptively adjust the filter passband through the user’s own heartbeat state,so as to extract BCG signals more accurately.And according to the actual BCG periodic signal characteristics,this paper proposes an optimization measure of J-peak fine-tuning on the basis of the general template matching method to solve the problem of positioning deviation due to non-100%template matching,so as to identify the J peaks in BCG signal sequence more accurately.The average of heart rate and respiration rate are calculated respectively according to the principle of action between BCG signal and heartbeat and respiration.Meanwhile,the BCG signal characteristics are input into the random forest classifier to study the mapping relationship between the signal characteristics and four postures:supine,prone,left-side and right-side postures,and thus realizes automatic monitoring of sleeping posture.In order to verify the recognition effect,this paper collected data from multiple subjects for an experimental analysis.The results show that the sleeping posture monitoring model based on the BCG signal in this paper is up to an accuracy of 81%in sleeping posture recognition.And this paper also analyzes the user’s sleep quality based on heartbeat,breathing and sleeping posture.This paper develops the smart medical system combined the sleeping posture monitoring and sleep quality analysis model,builds the software management system and the clients,so as to store and feed the data such as heartbeat,breathing,sleeping posture,and sleep quality back to the users,providing reference for further intervention and treatment.