A significant shift in contemporary medical treatment is represented by combining Artificial Intelligence (AI) and the Internet of Things (IoT) into telemedicine, virtual medical treatments, and remote patient monitoring. An increasing elderly demographic, the rise of chronic illnesses, and unequal access to healthcare treatment are driving up the need for worldwide medical structures, and AI is emerging as a key instrument in tackling these urgent problems. This investigation utilized a smart monitoring gadget with an IoT emphasis to monitor the heart rate of elderly people. The device tracks patients' GPS statistics in real-time and informs healthcare professionals of any confinement violations for any emergency. The smart sensor connects to a networking interface in the IoT cloud, which processes and analyzes data to determine bodily functioning. The suggested architecture has 3 levels of functionality: a cloud tier with Application Peripherals Interfaces for mobile gadgets, smart IoT detectors, and smartphone applications. Each layer has a particular role. Statistics from the IoT perception tier are initially gathered to detect heart attacks. This layer stores data in a cloud server for preventative measures, alerts, and speedy actions. The smartphone program alerts the patient and the family during emergencies. This study proposes CNN-UUGRU, a new deep neural network architecture that combines convolutional and upgraded gated recurrent segments for identifying human actions. This system outperformed other deep neural networks on the Kaggle database, accomplishing 97.8% accuracy, 96.9% precision, and an F-value of 97.85%. © 2025 Elsevier B.V., All rights reserved.