Leveraging Artificial Intelligence and Wearable Technologies in Glucose Managenment: A Comprehensive Scoping Review of Current Trends and Future Directions

Abstract

The integration of artificial intelligence (AI) and machine learning (ML) with wearable health monitoring systems presents transformative opportunities for advancing diabetes management. This paper delivers an in-depth review of state-of-the-art AI/ML
methodologies applied to wearable sensor technology, focusing on noninvasive techniques for real- time blood glucose monitoring. In
particular, we investigate the utility of wrist-worn photoplethysmography (PPG) signals for blood glucose estimation and assess IoTenabled continuous glucose monitoring (CGM) systems. A systematic literature search was conducted across prominent databases such as IEEE Xplore, PubMed, Scopus, and Web of Science. Studies were meticulously selected based on stringent criteria including technological innovation, methodological rigor, and clinical relevance. Our synthesis reveals significant progress in the accuracy, computational efficiency, and practical deployment of AI/ML algorithms in diabetes monitoring, while also identifying persistent challenges related to signal noise reduction, data heterogeneity, and sensor integration. The review further discusses the potential of these advanced technologies to enable early diagnosis, facilitate personalized treatment strategies, and ultimately improve patient outcomes. These insights underscore the critical need for continued interdisciplinary research and collaboration to overcome existing limitations and to drive the successful translation of these innovations into clinical practice.