Securing 5G-Enabled Cyber-Physical Systems: An Optimized Reflection Equivariant Quantum Neural Network Approach to DDoS Attack Detection
Beekanahalli Harish Swathi, R Arvind, Shalet Benvin, Ramji Gupta, Ramya Maranan, R. Jayanthi · 2025
The exponential increase in interconnectivity of CPS due to the rapid evolution of 5G networks makes it more susceptible to Distributed Denial-of-Service attacks. Real-time and accurate detection of DDoS attacks is required to ensure the reliability and security of CPS. This paper introduces an advanced approach to DDoS attack detection, including advanced preprocessing, feature extraction, and classification techniques. Data consistency and noise reduction were performed through a correlation coefficient min-max normalization approach. For feature extraction and classification, proposed a Reflection Equivariant Quantum Neural Networks (REQNN), which is designed to maintain symmetry in input transformations so that predictions are consistent across different representations of data. In order to optimize the hyperparameters of REQNN, the research adopted the Fennec Fox Optimization (FFO) algorithm inspired by the adaptive behaviors of Fennec foxes to achieve improved performance. Our proposed framework is evaluated on call detail records collected from the Telecom Italia site, The method demonstrated high recall, specificity, accuracy and F-Score, achieving 99.5% accuracy, 99.5% precision, 99.4% recall, and a balanced performance between precision and recall in the detection of DDoS malwares in the CPS across 5G networks. It significantly increases CPS resilience and provides a more robust solution to emerging cybersecurity requirements in next-generation networks.