Obstructive sleep apnea (OSA) affects over 936 million adults worldwide, leading to significant morbidity and reduced quality of life. Traditional diagnostic methods such as polysomnography (PSG), although effective, rely on hospital equipment and human resources and are not widely accessible. Emerging technologies such as the jaw tracking device developed by Sunrise have appeared, using gyroscopes and accelerometers to monitor jaw movements during sleep, showing high diagnostic accuracy compared to PSG. These devices use AI algorithms, providing a practical alternative for OSA diagnosis and monitoring. However, due to the high hardware cost of Sunrise and its rental-based business model, there are logistical challenges and limited monitoring capabilities. This prospective clinical study aims to offer the same capabilities through a novel ultrasonic transmitter for tracking jaw movements during sleep, which can be directly received by using a smartphone microphone, reducing the complexity of setup operations. This study will evaluate the sensitivity of the ultrasonic jaw tracking device in detecting apnea and compare it with PSG standards, developing a deep learning AI-driven algorithm to analyze data from jaw movements, breathing sounds, and apnea-related arousal events, as a basis for continuous home monitoring and evaluation of treatment effects for OSA patients.