Artificial Intelligence Driven Intrusion Detection for Cybersecure Smart Buildings Using Internet of Things Sensors

Abstract

Presented research addresses escalating cyber risk within smart buildings driven by dense Internet of Things sensing and automated control. The work targets timely detection and containment of malicious activity affecting building services and preserves operational continuity. A data driven security architecture integrates intelligent sensing, automated anomaly recognition, closed loop response, and tamper resistant logging. Evidence derives from controlled simulations using realistic attack injections, long duration sensor streams, and quantitative performance evaluation. Results indicate high detection accuracy, low false alarm rates, rapid alert latency, and limited computational overhead under sustained adversarial pressure. Automated isolation and recovery actions maintain service availability during denial and device compromise events. The findings extend prior research by coupling operational telemetry with security monitoring and by demonstrating feasible real time protection under resource constraints. Practical implications include deployable protection for building management systems and guidance for secure sensor network design. Future research directions include adversarial adaptation analysis, scalability across larger deployments, and refinement of learning driven defenses under heterogeneous device conditions.