A Smart RFID-Driven System for Dementia Patient Tracking: A Machine Learning Approach for Monitoring and Localization

The global increase in the elderly population, which is expected to reach 2.1 billion by 2050, has highlighted the need for reliable monitoring systems to assist elderly individuals, especially those with dementia. More than 55 million people worldwide live with dementia, a disease liable to induce fatal incidents such as wandering off and falling, resulting in almost 30% of injury-related deaths in the elderly. Current technologies, including GPS and camera-based systems, face severe limitations in indoor environments, such as privacy intrusions, high costs, and dependence on line-of-sight visibility. This study introduces a novel, cost-effective radio frequency identification (RFID)-based tracking system optimized for indoor settings to address these gaps. Leveraging the Internet of Things (IoT) architecture and cloud computing, our solution employs battery-less RFID tags embedded in unobtrusive wearable devices (e.g., anklets or bracelets) to enable real-time, multi-individual tracking without compromising privacy or relying on external power. Our proposed system uniquely integrates the efficiency of a low-complexity, cost-effective fingerprint-based localization framework with real-time data analytics and optimized ML models to achieve the accuracy, affordability, and scalability required for smart home applications for dementia. Extensive evaluation in simulated smart home environments demonstrates a 98% localization accuracy with NN and a modified KNN algorithm, outperforming existing approaches. As a proof of concept, the developed RFID-based localization system is capable of accurately tracking multiple elderly individuals within the home setting. Overall, the proposed system showed excellent accuracy results with only off-the-shelf components. The proposed system addresses scalability and cost barriers, offering a robust alternative to the more expensive and often hard-to-use commercial systems. This developed system not only enhances the safety of patients with dementia but also establishes a robust, adaptable framework for future IoT-driven healthcare applications.