A Federated Cyber-Physical Platform for Real-Time Coordination of Emergency Medical Service via 112 Integrations

This paper presents a federated cyber-physical platform for real-time coordination of Emergency Medical Services (EMS) through seamless integration with 112 emergency infrastructure. The proposed system combines physiological signals from wearable devices with metadata derived from 112 calls, such as urgency level, location, and time of incident, to enable decentralized, privacy-preserving triage at the edge. Leveraging federated learning, the platform ensures that patient data remains local while model intelligence is globally shared across EMS units, ambulances, and hospitals. A simulation involving 10,000 synthetic patient records demonstrated that the system achieves high triage accuracy (72%), sub-millisecond inference latency (0.10 ms), and significant bandwidth savings (98.5%) without compromising model fidelity (KL divergence = 0.0000). Furthermore, the platform reduced average emergency dispatch time by 27% compared to baseline workflows. These results confirm that integrating edge AI, federated learning, and emergency call metadata can transform traditional EMS infrastructures into intelligent, privacy-compliant, and response-optimized cyber-physical systems. DOI: 10.61416/ceai.v27i3.9635