Smart ICU Monitoring System Using IoT and Edge AI for Real-Time Patient Risk Prediction

This paper presents a Smart ICU Monitoring System that integrates Internet of Things (IoT) devices with Edge Artificial Intelligence (Edge AI) to enable real-time, predictive patient monitoring. Traditional ICU systems rely on reactive, threshold-based alarms that often lead to alarm fatigue and delayed clinical responses. The proposed system leverages continuous data acquisition from biomedical sensors and processes it locally using edge devices to perform low-latency inference. Advanced deep learning models such as LSTM and CNN are utilized to analyze time-series and waveform data, enabling early detection of critical conditions such as sepsis and cardiac arrest. The system enhances patient safety, reduces false alarms, and ensures data privacy by minimizing cloud dependency. Experimental evaluation using benchmark datasets demonstrates improved prediction accuracy and reduced response time compared to conventional approaches.