Scalable IoT-Based Architecture for Continuous Monitoring of Patients at Home: Design and Technical Validation

This article presents a scalable IoT-based architecture for continuous and passive monitoring of patient behavior in home environments, designed as a technical founda-tion for future dementia risk assessment systems. The architecture integrates: (1) wearable BLE sensors with infrared room-level localization capability; (2) edge-computing gateways with local preprocessing and machine learning capability; (3) a three-channel data stream optimizing latency, bandwidth, and information com-pleteness; and (4) a federated learning framework enabling model development with-out data sharing between multiple institutions. Technical validation in two apartments (3 participants, 7 days) demonstrated: 97.6% room-level localization accuracy using infrared beacons; 41.2% network bandwidth reduction through intelligent compres-sion; less than 7 seconds end-to-end latency for 99.5% of critical events; and 98.5% deduplication accuracy in multi-gateway configurations. A proof-of-concept federated learning simulation confirms architectural feasibility of collaborative model training while preserving privacy, achieving convergence in five rounds with 1.4 MB commu-nication per institution. Cost analysis shows ~€490 for initial implementation and ~€55 monthly operation, representing 5-10 times lower costs than existing research systems (ORCATECH, SENDA). The development ensures the technical and economic feasibil-ity of continuous home monitoring for behavioral analysis. Clinical validation of di-agnostic capabilities through longitudinal studies with validated cognitive assess-ments remains a task for future work.