Artificial Intelligence for Continuous Health Monitoring and Predictive Diagnostics

The healthcare sector is undergoing a profound transformation driven by the integration of Internet of Things (IoT) technologies and Artificial Intelligence (AI), leading to the emergence of the Smart Health paradigm. This new approach shifts healthcare from a traditional reactive and hospital centered model to a proactive, patient centered system that leverages continuous monitoring through wearable sensors and intelligent data analysis. Miniaturized devices collect real time physiological and behavioral data, which are transmitted securely via IoT infrastructures to enable continuous assessment outside clinical settings. AI plays a crucial role in transforming this massive, complex sensor data into actionable knowledge for accurate diagnosis, adaptive decision-making, and personalized care. This thesis explores the synergy of IoT and AI, known as Artificial Intelligence of Things (AIoT), in Smart Health systems, which rely on edge, fog, and cloud computing architectures to enable local processing, data aggregation, and large-scale analysis. Despite these advances, challenges such as data privacy, security vulnerabilities, bias in AI models, and the opaque nature of DL hinder clinical adoption. Moreover, the literature highlights significant gaps: there is limited application of DL techniques to IoT health data in real-world conditions, few end-to-end AIoT healthcare applications operate in clinical workflows, and the diagnostic potential of multimodal data fusion remains underexplored. To address these shortcomings, the thesis undertakes a systematic review of current research, an exploratory analysis of authentic IoT health datasets, and the design of an AI model tailored for Smart Health applications. The central objectives are to critically evaluate the state of the art, identify technological challenges and opportunities, and assess the effectiveness of DL methods in extracting robust, clinically relevant insights from multimodal physiological signals collected by wearable IoT devices. This work aims to contribute to the advancement of predictive, preventive, and personalized medicine by highlighting how artificial intelligence can effectively support clinicians in the decision making process. Keywords: Internet of Things (IoT), Artificial Intelligence, Hybrid Model, Smart Health