Abstract
Securing the physical components of a city's interdependent critical infrastructure (ICI) such as power, natural gas, and water systems is a challenging task due to their interdependence and a large number of involved sensors. In this paper, using a novel integrated state-space model that captures the interdependence, a two-stage cyber attack on an ICI is studied in which the attacker first compromises the ICI's sensors by decoding their messages, and, subsequently, it alters the compromised sensors' data to cause state estimation errors. To thwart such attacks, the administrator of each critical infrastructure (CI) must assign protection levels to the sensors based on their importance in the state estimation process. To capture the interdependence between the attacker and the ICI administrator's actions and analyze their interactions, a Colonel Blotto game framework is proposed. The mixed-strategy Nash equilibrium of this game is derived analytically. At this equilibrium, it is shown that the administrator can strategically randomize between the protection levels of the sensors to deceive the attacker. Simulation results coupled with theoretical analysis show that, using the proposed game, the administrator can reduce the state estimation error by at least 50% compared to a non-strategic approach that assigns protection levels proportional to sensor values.
Introduction
Modern smart cities rely on Interdependent Critical Infrastructures (ICIs)—such as power grids, natural gas pipelines, and water distribution networks—that are deeply intertwined. For instance, power plants require water for cooling, while water pumps and gas compressors require electricity to function. To monitor these systems, administrators rely on state estimation algorithms that process data from thousands of distributed sensors.
This interconnectedness creates a massive attack surface. If an attacker compromises these sensors, they can inject false data to blind administrators to the true physical state of the grid. Most existing security research either treats these infrastructures in isolation or assumes the defender has unlimited resources. This research addresses the realistic challenge of how a defender should allocate strictly limited security resources across a massive, interdependent network of sensors to minimize potential damage.
Methods
The authors propose a comprehensive mathematical and game-theoretic framework to address these vulnerabilities:
- Integrated Physical Model: A novel continuous/discrete state-space model is developed to mathematically link the physical dynamics of power, gas, and water systems.
- Cyber Attack Model: The study models a two-stage attack. First, the attacker spends resources to break sensor security. Second, the attacker alters data to maximize the Cumulative Error Difference (CED) in the system's centralized Kalman Filter. The attacker designs an "$\alpha$-feasible" attack, ensuring false data is bounded using Kullback-Leibler (KL) divergence to avoid detection.
- Colonel Blotto Game: To model the resource allocation battle, the authors employ a Colonel Blotto game. The attacker and defender simultaneously distribute limited resources across multiple "battlefields" (sensor clusters). The player who allocates more resources to a specific cluster wins control of it.
- Mixed-Strategy Nash Equilibrium (MSNE): Because deterministic strategies are easily exploited, the authors derive the MSNE by relaxing the problem into a "General Lotto" game. This provides a probability distribution detailing how the defender should randomize resource allocation to keep the attacker guessing.
Key Insights and Ideas
- Interdependence as a Vulnerability: Securing utilities in silos is insufficient. The authors prove that if a defender only protects one infrastructure (e.g., the water grid), an attacker can exploit interdependencies by attacking the gas or power grid to cause massive estimation errors in the water grid.
- Strategic Randomization: If a defender statically allocates resources based purely on sensor importance, a smart attacker can predict and exploit the thresholds where the defender is weak. A mixed strategy (randomizing defenses based on the MSNE) forces the attacker into a mathematically disadvantageous position.
- Quantifying Sensor Value: The paper provides a rigorous method to calculate the "value" of any given sensor cluster. This value is defined as the maximum stealthy damage (CED) an attacker could cause to the entire interdependent grid if they controlled that specific cluster.
Results and Evaluation
The framework was evaluated using a simulated synthetic topology based on real-world parameters, consisting of 10 generators, 11 gas pipelines, 11 water pipelines, and 32 sensor clusters.
- 50% Error Reduction: By utilizing the MSNE strategy, the defender can reduce state estimation error by at least 50% compared to baseline approaches where resources are statically allocated proportional to sensor value.
- The Cost of Siloed Defense: In simulations, when the defender protected only the natural gas infrastructure, the average estimation error of natural gas state variables increased by a factor of 2.6 compared to a holistic defense strategy.
- Theoretical Alignment: Empirical simulations closely matched theoretical predictions. For instance, when the attacker had 20% of the defender's resources, the theoretical expected percentage of compromised sensors was 10%, and the simulation averaged 11%.
Conclusion
This work demonstrates that strategic randomization and a holistic view of infrastructure interdependencies are vital for smart city security. While the current model assumes a centralized administrator with perfect knowledge, future work could explore decentralized state estimation models where utility owners cooperate without sharing proprietary raw data, or games of incomplete information where resources and system knowledge are uncertain.
Summary updated