Quantum-Driven Anomaly Detection Framework for Consumer IoT Cyber-Physical Systems

초록

This study aims to enhance the security of Consumer IoT (CIoT) systems by addressing the limitations of traditional anomaly detection approaches. To achieve this, we propose the Quantum-Driven Adaptive Anomaly Detection Framework (QADAPT), a novel model designed to enable real-time anomaly detection through a quantum-inspired adaptive cognitive mapping function. The framework is built upon a multilayered architecture consisting of a Quantum-State Convolutional Layer, Synthetic Verification Layer, and Adaptive Mapping Layer, allowing simultaneous data state analysis and validation against synthetic signals. Q-ADAPT uses an adaptive deep learning model to recognize evolving CIoT behavior patterns, enhancing detection accuracy and resilience under varying noise conditions. The simulation environment spans a time frame of 340 minutes, designed to evaluate the robustness of the model in six distinct scenarios under Gaussian noise. Performance results reveal that Q-ADAPT achieves a detection accuracy of 97.8% in lowcomplexity environments and maintains 91.3% under high-noise conditions. © 1975-2011 IEEE.

키워드