Vulnerability Mitigation of Urban Traffic Control Against Cyberattacks Using Secure Multi-Party Computation

Abstract

Traffic signal control systems are essential for safe and efficient urban traffic management. However, these systems increasingly rely on collecting and processing extensive data, posing confidentiality challenges that must be addressed to protect sensitive information and safeguard against cyber-threats. This article proposes a privacy-preserving traffic signal control strategy based on secure multi-party computation. The design uses multiple computing servers and a secret sharing scheme to encrypt traffic data into several shares, each corresponding to a server. The servers then securely compute the green time plans, working directly on the encrypted data. The proposed method protects traffic data from curious computing servers and external eavesdroppers, preventing the adversaries from gaining the required information to launch a complex attack. Additionally, a detection and recovery procedure from false data injection attacks on green time signals is presented using secret reconstruction from different server combinations. Simulation results for various scenarios demonstrate the proposed method’s effectiveness in maintaining data confidentiality while achieving the same performance as an unprotected control strategy, with a reasonable addition to computation time. It also depicts a significant reduction in vulnerability against falsified green time signal attacks in terms of total travel time in the network.