Design of a Smart Home System for Independent Elderly Integrating Environmental Sensing and AI-Based Safety Monitoring
Year: 2025
Type: article
Abstract: With accelerating global population aging and more elderly living alone, ensuring their safety, comfort, and rapid emergency response is an urgent societal challenge. This study introduces a smart home integrating a digital temperature-and-humidity sensor, a light intensity sensor, a combustible-gas sensor, an air-quality sensor, and a camera-based You Only Look Once version 5 small (YOLOv5s) model for non-intrusive fall detection, all driven by an STM32F103 microcontroller unit (MCU). Hardware architecture and control logic were validated in Proteus, and the YOLOv5s algorithm was trained and evaluated on a PC. Simulations demonstrate reliable real-time regulation of heating, ventilation, and air conditioning, lighting, and gas-exhaust actuators with millisecond-level alarm triggering; the fall-detection module achieved a mean average precision at an intersection-over-union threshold of 0.5 (mAP\@0.5) of 0.848 and an accuracy of 0.830 for the “down” class, outperforming more computationally intensive detectors under constrained hardware resources.
Author: Zefeng Liu
Institution: Beijing University of Posts and Telecommunications
Language: English
Subfield: Computer Vision and Pattern Recognition
Field: Computer Science
Domain: Physical Sciences
Open Access status: diamond
OpenAlex ID: W4414199867