Abstract
Physical cyber systems have introduced new threats and challenges in detection and immediate response. This study examines how Graph Neural Networks (GNNs) can aid cybersecurity and drone management within a physical cyber system involving cyber intrusions and unmanned aerial vehicles (UAVs). By bridging the structural understanding of graphical neural networks, this work presents an integrated procedure that allows intrusion detection systems to learn underlying network structures, identify malicious activities, and facilitate drone response measures.
Methodology
Based on an emulation-based case study, cyberattack models were created to provoke drone responses, proving that graph-based learning can enhance situational awareness, swarm coordination, and adaptive maneuvering. According to performance valuation, this method has a detection rate of 94.2%, an average area under the receiver operating characteristic (ROC) curve of 0.955, and an average response time of 1.4 seconds.
Comparative Experiments
Comparative experiments reveal that the proposed GraphSAGE network is more effective than Graph Convolutional Networks (GCNs) and Graph Attention Networks (GATs) in identical situations. Such findings demonstrate that graphical neural networks can be utilized to avert intrusion and respond within dynamic cyber-physical systems.
Blogger's Review: This article showcases the innovative application of Graph Neural Networks in cybersecurity, particularly in drone management. The combination of high detection rates and rapid response times highlights the adaptability of GNNs in complex environments, warranting further promotion in future cybersecurity strategies.