A Compact and Wireless Wearable System for Sleep Apnea Detection

Sleep apnea is a prevalent sleep disorder with serious health consequences. While polysomnography is the clinical gold standard for apnea diagnosis, it is costly, obtrusive, and impractical for continuous or home-based monitoring. Existing wearable solutions also often lack sufficient accuracy and reliability. This paper presents a novel and practical method for the accurate diagnosis of sleep apnea. A compact wireless wearable device is developed to capture bio-signals and transmit them to a smartphone. The device integrates an Inertial Measurement Unit (IMU) and a photoplethysmography (PPG) block to directly capture signals from the body. These signals are then modulated and wirelessly transmitted to the user’s smartphone via Bluetooth Low Energy (BLE). As the third block of the system, the smartphone’s internal microphone records the user’s breathing sounds throughout sleep. Together, the three blocks generate six signals: SpO₂, heart rate (both extracted from the PPG signal), breathing sound, and 3-axis acceleration, which are subsequently processed in MATLAB. By applying signal processing techniques, sleep apnea and hypopnea events, along with their timing, can be diagnosed with high accuracy. Based on an event-to-event comparison, the developed system achieves a sensitivity of 85.8% and an F1-score of 87.3% using a single signal. The use of a combined approach further enhances sensitivity, reliability, and the F1-score to 91%.