An Adaptive Quantum-Resistant Intrusion Detection and Secure Communication Framework for Next-Generation Cyber-Physical Systems

Abstract

Next-generation cyber-physical systems need to be protected by security mechanisms that are able to adapt to new attacks and to withstand the new threats that are possible in the future with quantum technologies. This study designed an adaptive quantum-resistant intrusion-detection and secure-communication framework that combines the elements of behaviour monitoring, concept-drift adaptation, post-quantum key establishment, digital signature, policy control, and automatic response. CPS dataset with 1,500 observations and 38 variables was generated, along with 480 post-quantum cryptographic benchmark trials, for a simulation-based design. The CPS data covered normal traffic, 6 different categories of attacks, 94 zero-day attacks, and 218 concept-drift observations. The baseline classifiers were logistic regression, decision tree, random forest, SVM, and gradient boosting, which were evaluated, and the proposed adaptive IDS was compared with a static one. The adaptive model achieved 95.67% accuracy, 98.99% precision, 90.88% recall, a 94.76% F1-score, 99.35% ROC-AUC, and a 0.70% false-positive rate. The recall accuracy on zero-day improved from 28.72% to 77.66% and concept-drift adaptation accuracy improved from 55.96% to 81.65%. Post-quantum profiles offered more long-term security, but necessitated more handshake time, memory, energy, and communications overhead. The results indicated that the most efficient and quantum-resistant combination was the use of ML-KEM-768 and ML-DSA-65. The results show the potential of adaptive intrusion detection combined with post-quantum communication for next-generation CPS, but testing on physical testbeds and real operational datasets is required.