From Sensors to Edge Intelligence: A Critical Review of AI-Enabled IoT for Healthcare

Artificial intelligence is now deeply integrated into the Internet of Medical Things (IoMT), supporting applications from wearable sensors to hospital servers. However, current deployments face three persistent issues: reliance on cloud processing that delays urgent responses, fragmented security measures that leave system gaps, and evaluation based on small or single-site datasets. This paper critically reviews recent advances across five key domains: sensing technologies, lightweight security mechanisms, edge and federated computing, data foundations, and new interaction channels such as mixed reality. The review shows that improved sensors and energy-aware radios enhance accuracy and battery life, compact intrusion detectors secure low-power devices, and edge gateways reduce cloud dependence. At the same time, gaps remain in multi-hospital benchmark datasets, fast and resource-efficient federated updates, and explainable decision-making for clinicians. By consolidating these findings, the paper highlights both the progress achieved and the open challenges, pointing to opportunities for future research in energy-aware privacy schemes, interoperable standards, and transparent AI models that can be trusted in real clinical practice.