A computational framework for elderly residence interior design: Integrating intelligent health monitoring via SOM-optimized K-means and wireless sensor networks

Abstract

With the accelerating pace of population aging, elderly individuals living alone face heightened risks in home-based care, necessitating integrated solutions for real-time health monitoring and adaptive residential design. This study proposes a computational framework that combines intelligent health monitoring with interior spatial optimization, leveraging data mining and wireless sensor networks (WSNs) to enable data-driven personalization. The framework includes WSN-based data acquisition, SOM-optimized K-Means clustering for behavioral feature extraction (achieving a 19% reduction in Davies Bouldin Index to 0.38 and 89.6% clustering purity), and spatial-behavior fusion using trilateration-based indoor positioning (error <0.5 m) to guide layout decisions such as emergency device placement. A multi-feature fusion algorithm integrates motion and appliance usage data to support real-time behavior recognition with 92.1% accuracy. Experimental validation demonstrates a 37% reduction in false alarms and 91% alignment of adaptive design recommendations with user preferences. The WSN maintains sub-280 ms latency, ensuring timely alerts. This framework advances computational gerontechnology by offering a scalable and practical solution for integrating intelligent monitoring with aging-in-place living environments.