IoT-Enabled Wearable System for Real-Time Fall Detection and Elderly Monitoring

Falls are one of the leading causes of injury and loss of autonomy among the elderly, especially in the context of aging populations and the growing need for real-time health-monitoring technologies. This study develops and validates a low-cost wearable system that is designed to detect falls and monitor posture using embedded Internet of Things (IoT) infrastructure. The system architecture integrates an ESP32 microcontroller with an MPU6050 inertial sensor, wirelessly transmitting motion data via the MQTT protocol to a Raspberry Pi 3, which processes the information and activates an external camera when a fall is suspected. A threshold-based algorithm was implemented to classify user postures and detect abrupt motion variations associated with fall events. The entire system was validated through controlled experiments simulating daily activities—such as standing, walking, sitting, and lying down—as well as various types of falls. The results indicated reliable performance in detecting upright and supine postures and capturing acceleration and angular velocity patterns during simulated falls. However, the system presented difficulties in distinguishing sitting from fall events and identifying soft falls, achieving an overall classification accuracy of 60%. Hardware integration, wireless communication stability, and real-time visualization through the Node-RED dashboard were implemented, highlighting the feasibility of combining embedded sensing with lightweight communication protocols for wearable elderly-monitoring applications.