MediTrack: a smart medication adherence system using OCR, IoT, and cloud-based centralized health database

Abstract

Medication non-adherence is a critical healthcare challenge, particularly among elderly and chronically ill patients in developing countries where limited healthcare infrastructure, fragmented medical records, and lack of caregiver support increase the risk of missed or incorrect doses. This paper presents MediTrack, an integrated smart medication management system designed to improve medication adherence through automated dispensing, sensor-based intake verification, and cloud-enabled prescription management. The proposed system combines optical character recognition (OCR) and natural language processing to digitize handwritten and printed prescriptions in both Bangla and English, enabling automatic schedule generation and centralized electronic health record storage. A microcontroller-driven dispensing unit utilizes stepper motors for precise dosage delivery, while infrared sensors and load-cell–based weight measurement verify pill dispensing and patient intake in real time. To address missed doses, a multi-level alert mechanism is implemented, incorporating audible alerts, visual notifications, and a mobility-enabled line-following robot that physically delivers medication to patients when necessary. Environmental monitoring ensures proper medicine storage by controlling temperature and humidity. Experimental validation demonstrates reliable dispensing accuracy, high intake detection performance, and effective missed-dose handling, while maintaining low power consumption and affordability. The system is designed to operate under limited internet connectivity and supports scalable deployment in home-care and institutional settings. MediTrack offers a cost-effective, localized, and scalable solution that enhances patient safety, reduces caregiver burden, and supports the transition toward digital healthcare systems in resource-constrained environments.