Adaptive Information Security Management in Cloud-Oriented Intelligent Transportation Systems

Abstract

The paper investigates theoretical and applied aspects of adaptive information security management in cloud-oriented Intelligent Transportation Systems (ITS). We analyze architectural features of the “Cloud–Edge–Vehicle” environment, identify key cyber-threat vectors, and substantiate the shortcomings of traditional security approaches—specifically static access policies, centralized PKI solutions, and classical IDS/IPS—in multi-domain, dynamic transportation ecosystems. We develop a mathematical and algorithmic model of adaptive management that treats an ITS as a discrete dynamical system whose state evolves over time under the influence of external threat factors and adaptive security measures. The control problem is formalized as minimization of a combined risk-and-cost functional, ensuring a balance between the level of protection and system performance.
Within the proposed approach, mechanisms for collecting and analyzing telemetry and network data, risk assessment methods, algorithms for optimal selection of security policies, and feedback loops across architectural layers are integrated into a single closed-loop adaptive management system. For anomaly detection in streaming data, we apply machine-learning algorithms (in particular, LSTM/GRU and autoencoders), while model consistency across layers without data centralization is achieved via Federated Learning. We consider methods to counter adversarial attacks both during training and at inference time for ML components. The obtained results provide a scientific foundation for building scalable, threat-resilient, and resource-efficient protection mechanisms for cloud-oriented ITS. The proposed model can be used to enhance the security of transportation networks, ensure policy coherence across architectural layers, and automate real-time cyber-threat response.