Evolutionary Game Analysis of Security Strategies in Urban Traffic Control Systems

To investigate the security issues of Urban Traffic Control (UTC) systems, it is essential to understand the network structure. A typical UTC system comprises Traffic Signal Machines (TSMs) located at intersections and an Upper Computer (UC) housed in the traffic management center. TSMs receive data from the UC and traffic sensors to control traffic lights, while the UC manages and coordinates the TSMs.

Attackers may attempt to disrupt or disable the UTC system to achieve malicious goals. Both the UC and TSMs can employ various defensive strategies, such as information security measures and coordinated control mechanisms. Determining the optimal defensive strategies for the UC and TSMs is crucial. Moreover, attackers adapt their strategies in response to the defenses employed. This dynamic interaction leads to an evolving game, well-suited for analysis using Evolutionary Game Theory (EGT).

Hypothesis 1: We define the following:

• UC: Represents the Upper Computer. Its defensive strategy includes selecting defense with probability 'p' and not defending with probability '1-p'.* TSM: Represents the Traffic Signal Machine. Its defensive strategy includes selecting defense with probability 'q' and not defending with probability '1-q'.* A: Represents the attacker. Its strategy includes selecting attack with probability 'r' and not attacking with probability '1-r'.* α: Detection rate of the attack detection system.* β: False alarm rate of the attack detection system.* S: Security value of the traffic network.* C_a: Attack cost.* C_u: Individual defense cost of the UC.* C'_u: UC's defense cost when coordinating with the TSM.* C_t: Individual defense cost of the TSM.* C'_t: TSM's defense cost when coordinating with the UC.

Hypothesis 2: Assuming both players are rational, the security value 'S' must exceed any associated cost. Otherwise, there's no incentive for either party to prioritize security. Therefore: S > C_a, S > C_u, S > C'_u, S > C_t, S > C'_t.

Let's analyze the payoffs for different scenarios when the attacker targets the UC:

• Scenario 1: (p, q, r) = (1, 1, 1) - Attacker's payoff: αS - C_a; Defenders' payoff: (1-α)S.* Scenario 2: (p, q, r) = (1, 1, 0) - Attacker's payoff: -C_a; Defenders' payoff: S - C_u - C_t.* Scenario 3: (p, q, r) = (1, 0, 1) - UC's payoff: (1-α)S - C'_u; TSM's payoff: (1-α)S - C_t; Attacker's payoff: αS - C_a.* Scenario 4: (p, q, r) = (0, 1, 1) - UC's payoff: (1-α)S - C_u; TSM's payoff: (1-α)S - C'_t; Attacker's payoff: αS - C_a.* Scenario 5: No detection errors - Defenders' payoff: -C_u - C_t.

Analyzing these payoffs allows us to construct an evolutionary game payoff matrix (Table 1). This matrix illustrates the payoffs for all three parties under various strategy combinations at any given moment. Although specific values may change if the attacker adopts different tactics, the underlying principles and the matrix structure remain applicable.

This study employs EGT to analyze the dynamic interplay of strategies between the UC, TSMs, and attackers in UTC systems. Our findings offer valuable insights for developing effective defense mechanisms and enhancing the security and resilience of critical traffic infrastructure.