Development of a Wrist-Worn Sleep Monitoring Device
Sleep is one of the five basic survival needs of human beings and is required to maintain physical and mental balance. The health and living standard of millions of people are adversely being affected by various sleep disorders which are proven to be health hazardous for up to 45% of people worldwid
2025-06-28 16:32:03 - Adil Khan
Development of a Wrist-Worn Sleep Monitoring Device
Project Area of Specialization Wearables and ImplantableProject SummarySleep is one of the five basic survival needs of human beings and is required to maintain physical and mental balance. The health and living standard of millions of people are adversely being affected by various sleep disorders which are proven to be health hazardous for up to 45% of people worldwide. Looking at the severity of this situation, it becomes necessary to analyze sleep, sleep cycles and diagnose sleep disorders at the earliest. Although many gadgets have been designed in this regard and their use is valued all over the world, very few sleep-related studies have been carried out and there is limited availability and adoption of wearable sleep technology in Pakistan. Limited research reported in the literature indicates that more than 30% of the general adult population is suffering from insomnia and a major proportion is at high risk for obstructive sleep apnea. Moreover, people with financial crises are 59% more likely to be affected by sleep disorders. This research aims to develop a system that contributes to medical technology by carrying out sleep research on the Pakistani adult population. The research will be published in conference and journal papers for the benefit of other researchers. Efforts will be taken to commercialize the device, and the data obtained will be made available to sleep researchers in Pakistan. The developed hardware shall be the property of the university, so as to allow other students to generate more data and apply different algorithms to increase the system’s efficiency. The sleep monitoring system proposed in this paper comprises a Raspberry Pi controller and commercial off-the-shelf sensors for measurement of body movement, heart rate, and blood oxygen saturation. The designed system will be evaluated using Machine learning i.e., the data obtained from the proposed system will be processed using different machine learning (ML) algorithms so to classify sleep stages. Besides, the behavior of recorded data will be analyzed for making decisions in a real-time environment. Processed data will be stored on Cloud to provide ease for long-term monitoring. An android application will be developed to fetch results from the cloud. The results will be visualized on a smartphone app to aid convenience to the patient and to the warden/attendant in adverse circumstances. The suggested system will allow people to play an effective role in maintaining their physical and mental health and will permit them to experience healthier sleep.
Project ObjectivesThis project aims to design and develop a sleep monitoring system that is easy to use, inexpensive, and comparable in accuracy to similar commercially available systems.
The project objectives are mentioned below:
- To select sensors for the device.
- To design and develop smart wristband prototype.
- To store data on cloud-based service.
- To develop an android application to display results.
- To validate the prototype by comparing its performance with a commercial device.
This project is divided into segments starting from the selection of hardware and software for the wrist-worn prototype. A Tri-axis Accelerometer (ADXL345) and Pulse oximetry & heat rate sensor (MAX30100) for triple-axes motion detection and for sensing oxygen saturation in blood and heart rate signals, respectively, shall be integrated with the Raspberry Pi controller. The later phase involves the development of Python code for the measurement of body vitals and physical movements. Data will be obtained from the patient’s body via the sensors and will be pre-processed. This data then shall be used to classify sleep into stages using different machine learning algorithms. The processed data will be stored to some cloud-based service such as Firebase. An android based mobile application will be developed on Android studio to fetch the results from firebase and for efficient visualization of decision making. The system performance shall be measured by producing a confusion matrix for the prototype. The performance of the prototype shall be compared with the performance of commercially available accurate sleep monitoring device. Figure 1 shows the implementation workflow.
- Indigenous development of sleep monitoring system: This device may satisfy the need for home-based, inexpensive, comfortable, and convenient sleep monitoring. It may be used to report the wearer about the quality of sleep (restful or restless sleep) and can determine whether the person is getting enough amount of sleep each night. It may also assist the wearer to improve his/her sleep habits by using features such as bed-time alarms or wake-up alarm clock.
- Creation of Sleep data: Goals of national and personal development are achieved through healthy population. Proper sleep is essential for good health. A very limited sleep research is reported from Pakistan. There is a dire need to carry out sleep studies and to make people aware of the complications arising from sleep disorders. The indigenous development of wearable sleep monitoring device in Pakistan will help understand sleep-related issues within the adult population.
- Low-cost alternative: The proposed wrist-worn device will be less costly than the existing sleep monitoring devices. Figure 2 illustrates various sleep monitoring devices and their prices in PKR.
The project at the end would have a smart wrist band prototype consisting of Raspberry Pi 4 and integrated sensors for measurement of body movement in X, Y, Z axes, heart rate, and oxygen saturation in the blood. The Raspberry Pi will collect data from an Accelerometer (ADXL345) and Pulse oximetry and heart rate sensor (MAX30100). The data acquired will be processed using different Machine Learning algorithms and sleep stages will be classified accordingly. The processed data will be sent on Google firebase where it will be stored for long term monitoring. An android application will fetch the results from firebase and the outcomes will be effectively visualized on Smartphone. The caretaker will also be alerted/notified in case of any abnormal condition. The conceptual design of the project is illustrated in Figure 3.
| Item Name | Type | No. of Units | Per Unit Cost (in Rs) | Total (in Rs) |
|---|---|---|---|---|
| Total in (Rs) | 70500 | |||
| Raspberry PI 4 Model B + SD card + HDMI cable + Power supply | Equipment | 1 | 15000 | 15000 |
| MAX30100 | Equipment | 3 | 400 | 1200 |
| ADXL345 | Equipment | 3 | 300 | 900 |
| Fitbit sleep monitor | Equipment | 1 | 30000 | 30000 |
| Glue gun | Equipment | 1 | 1000 | 1000 |
| Soldering Iron | Equipment | 1 | 400 | 400 |
| Peripherals (Monitor, Keyboard, Mouse) | Equipment | 1 | 10000 | 10000 |
| Li-Ion Battery | Equipment | 3 | 500 | 1500 |
| Li-Ion Battery charger | Equipment | 2 | 50 | 100 |
| Jumper Wires | Equipment | 40 | 5 | 200 |
| Soft wires | Equipment | 2 | 100 | 200 |
| All Miscellaneous | Miscellaneous | 1 | 10000 | 10000 |