Audio-based Wearable Multi-Context Recognition System for Apnea Detection

Apnea or sleep apnea syndrome is a condition in which a person unconsciously stops breathing during a sleeping state for longer than a certain timeframe. Multiple long-term apnea events induce various impairments. However, apnea detection in hospitals is an intensive and complicated procedure, resulting in a high rate of non-diagnosis and a low awareness of the disease. Existing wearable devices for apnea detection mostly use heartbeat signal patterns and SpO
2
levels to detect the disease; however, because apnea is a respiratory impairment, it is believed that using the breathing pattern is the most straightforward approach in apnea detection. Several recent studies investigated whether swallowing frequency during sleep can increase along with the apnea severity. However, the number of wearable devices using swallowing to detect apnea is extremely limited. Thus, this study proposes a wearable system to recognize human contexts such as breathing, heartbeat pattern, and swallowing using an audio sensor. Experiments were conducted to compare and obtain the most suitable parameters for the system such as the window sizes, types of audio feature values, and classification algorithms. A prototype of the device was built and is able to detect breathing, swallowing, heartbeat, oral sounds, and body movement. The results show the highest accuracy of 76.9% when using a 1-s window size and Mel's frequency cepstral coefficient features in contact microphone data.