Abstract
Sleep is very important for people to preserve their physical and mental health. The development of the ballistocardiography (BCG) sensor enables the possibility of day-to-day and portable monitoring at home. The goal of this study is to develop an IoT sleep quality monitoring system using BCG sensors, microcontrollers and cloud servers. The BCG sensor produces ECG data from the physical activity of the patient. The data is sent to the sensor and is read by the microcontroller. The sensor data is collected and pre-processed in the microcontroller. The microcontroller then transmits the data obtained from the BCG sensor to the cloud server for further analysis, i.e. to assess the sleep quality. The assessment of data transmission efficiency and resource consumption is carried out in this paper. The findings of the evaluation show that the proposed method achieves higher efficiency, lower response time and decreases memory usage by up to 77% compared to the conventional method.
Highlights
Sleep is a daily rhythm or circadian rhythm in humans that is governed by the human biological clock in the brain's core hypothalamus [1]
Of the 4 sleep stages, the patient's sleep quality will be classified using a combination of the Weight Extreme Learning Machine (WELM) method and the Particle Swarm Optimization (PSO) method used by Utomo et al (2019) [9] with an accuracy rate of 78.78% in three sleep classes (NREM, rapid eye movement (REM), awake) and an accuracy rate of 73.09% in four sleep classes
BCG sensor generates data stored on the microcontroller in the form of a text file printed by the Python program and stored on a cloud server database using the PostgreSQL database
Summary
Sleep is a daily rhythm or circadian rhythm in humans that is governed by the human biological clock in the brain's core hypothalamus [1]. The system is designed to collect patient sleep data by recording directly using a ballistocardiography monitor, the output of which comes in the form of a dataset for heart rate variability (HRV) Those data will be used to classify sleep cycles in humans, such as waking, light sleep, deep sleep, and rapid eye movement (REM) [8]. Of the 4 sleep stages, the patient's sleep quality will be classified using a combination of the Weight Extreme Learning Machine (WELM) method and the Particle Swarm Optimization (PSO) method used by Utomo et al (2019) [9] with an accuracy rate of 78.78% in three sleep classes (NREM, REM, awake) and an accuracy rate of 73.09% in four sleep classes This scientific study has been carried out in response to previous research which still needs studies in the construction of a portable, high-performance and low-resource-consumption health monitoring system to monitor sleep quality using BCG sensors.
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
