Introduction
Modern Intelligent Transportation Systems (ITS) are increasingly reliant on autonomous connected vehicles (ACVs) that depend on real-time control signals from communication base stations. This creates a critical chain of dependency: the ITS relies on Communications Infrastructure (CI), which in turn depends on the Power Grid (PG). This paper addresses the security implications of these interdependencies, specifically focusing on how stealthy attacks on the power grid can disrupt urban traffic flow.
The Multi-Layer Dependency Problem
An attacker can cause significant traffic disruptions without directly hacking vehicles or communication towers. By targeting the power grid, an adversary can degrade the quality of service of the cellular network, leading to non-optimal vehicle speeds and traffic congestion. Given limited budgets for backup power, city administrators must strategically allocate resources to protect the most critical components of this triple-layer system.
Methodology: Interdependence and Game Theory
The authors propose a two-tier interdependence model and a game-theoretic defense strategy:
- Interdependence Modeling: A mathematical mapping calculates how failures in power generators or lines ripple through the cellular network to cause traffic flow deviations on specific streets.
- Stealthy Attack Model: Attackers are modeled as "risk-averse," choosing targets that cause disruption while remaining difficult to detect.
- Stackelberg Game: The problem is framed as a leader-follower game where the defender (leader) allocates Backup Power Sources (BPS) first, and the attacker (follower) responds optimally.
- Optimization: The Stackelberg Equilibrium (SE) is characterized and solved using linear programming, ensuring computational efficiency for large-scale applications.
Experimental Setup and Modeling
The framework was validated using synthetic simulations of city topologies. The ITS is modeled as a grid-based city with $N$ intersections, while the CI uses standard hexagonal cell layouts. The PG is represented as a graph-based network of generators and load nodes.
Key Insights and Results
- Analytical Mapping: The core contribution is the precise calculation of how reduced communication coverage translates into increased fuel consumption and traffic jams.
- Sensitivity to Infrastructure: Simulations revealed that the system is more sensitive to attacks on power lines than on generators, as line attacks can be more targeted and stealthier.
- Performance: The proposed SE strategy reduced traffic flow deviation by up to 40% compared to naive equal-distribution strategies.
- The 50% Plateau: Findings show that once power generation is reduced by 50%, traffic deviation plateaus. At this level, base stations lack the bandwidth for high-priority control packets, making further power cuts redundant for the attacker.
- Scalability: Despite the complexity, the solution scales in linear time, making it suitable for metropolitan deployment.
Conclusion
This research provides a robust framework for securing ITS against indirect infrastructure attacks. By understanding the precise interdependencies between power, communications, and traffic, administrators can prioritize high-impact base stations for backup power. Future work will explore dynamic games to account for real-time strategy adjustments and human driver behavior in mixed-autonomy environments.
Summary updated