IoT-Integrated Fall Detection Using YOLO and Servo-Camera System

Falls among elderly individuals remain a critical public health concern, particularly in indoor environments where most incidents occur. While many existing fall detection systems rely on static camera configurations with limited fields of view (FOV) and generic pre-trained models, few have explored the integration of dynamic FOV camera control via mobile devices or addressed practical deployment challenges such as thermal performance and real-time IoT communication. This paper presents an IoT-integrated fall detection system that utilizes a wide-angle RGB camera dynamically repositioned via a smartphone application. The system incorporates a servo motor mechanism for remote camera adjustment and leverages the YOLOv8 deep learning model for real-time image-based fall detection. It is deployed on a Raspberry Pi 4 platform, with a Telegram-based interface enabling both system control and alert notifications. Experimental results show that the system achieves a detection accuracy of 88.5 % under optimal conditions. Comparative evaluations of camera angles and enclosure thermal designs emphasize the significance of field of view and hardware reliability. This work makes a practical contribution to fall detection system design for robust, real-world deployment.