A U.S.-based client was developing a smart sleep apnea device named Jasper to monitor and manage sleep apnea conditions. The device needed to detect apnea episodes and regulate air pressure automatically. Additionally, snore detection was crucial for comprehensive patient monitoring.

Semphos provided comprehensive firmware development services for Jasper, leveraging the ESP32-S3 platform for its robust processing capabilities. The work included:

Custom Firmware Development: Implemented firmware that controlled the device’s air pressure system, allowing automatic adjustment based on apnea detection.

Snore Detection Model: Developed and trained a snore detection model using an enhanced dataset to accurately identify and count snores, providing real-time data and insights for patient monitoring.

Real-Time Data Processing: Ensured the firmware was optimized for efficient, real-time processing of sensor data to interface with a smartphone app via BLE.

The implementation of Semphos’s solutions, while promising, encountered challenges in porting the snore detection model from PC to the ESP32-S3 platform. Despite these difficulties, the project achieved significant progress. The model, which performed effectively on a PC, highlighted the complexities of adapting advanced machine learning algorithms to embedded systems. Through hard work and dedication, Semphos provided a robust foundation and valuable technical insights that will aid in refining the device’s performance in future iterations. This experience underscored the importance of continuous innovation and adaptation in IoT development, contributing to future advancements in sleep health technology.