DEAN (Digital Ear and Nose) Node: Non-Intrusive Resident ADL Recognition at Home

초록

As aging populations increase, Ambient Assisted Living (AAL) systems are essential for promoting independent living. Human Activity Recognition (HAR) plays a key role in these systems, but traditional approaches using wearable devices and sensor networks suffer from issues such as user discomfort, battery limitations, high installation costs, and scalability constraints. This study introduces the DEAN node, a non-intrusive solution that integrates sound and environmental data for improved indoor activity recognition. By leveraging multi-data fusion, including temperature, humidity, illuminance, and air quality, the system enhances detection accuracy beyond sound-based recognition alone. Furthermore, edge computing ensures real-time processing, reducing network load and enhancing privacy by minimizing data transmission to external servers. The system incorporates a hub device that consolidates recognition results from multiple nodes, facilitating seamless coordination and scalable deployment in smart living environments. This research demonstrates the potential of DEAN nodes in AAL applications and provides a foundation for future enhancements, including anomaly detection and broader deployment in diverse environments.

키워드