Fall Detection with Artificial Intelligence and IoT Health Monitoring System

The elderly population is constantly increasing, facing physical and cognitive limitations that require attention and care. Consequently, caregivers of older adults must maintain constant attention to their physical well-being. Falls are common and can produce serious consequences. The constant monitoring of vital signs is crucial to monitor health status and respond quickly to emergencies. Therefore, the design of a device that accurately detects risk situations in an older adult is a challenge. This paper presents the development of a bracelet-type device with Internet of Things (IoT) architecture for automatic fall detection and vital signs monitoring. Fall detection is performed by a machine learning algorithm trained on data obtained from an accelerometer. The Edge Impulse platform was used where the data processing and the learning block are configured. The trained model was implemented in an ESP32 microcontroller that detects falls with an accuracy of 91%. The device incorporates sensors for monitoring temperature, heart rate, and oxygen saturation that are measured on the wrist of the elderly. A mobile application for the caregiver of the elderly was implemented that issues alerts if the system detects falls, notifies vital sign values if these are outside the normal range, and presents the results in real time.