Liu, Xi (Tongji University), Yu, Chunhui (Tongji University), Su, Zicheng (Tongji University), Ma, Wanda (University of Shanghai for Science and Technology)

A Semi-Black-Box Approach for Falsified Trajectory Injection Attack against Network Signal Control

Scheduled for presentation during the Video Session "On-Demand Video Presentations" (VP-VP), Saturday, November 22, 2025, 08:00−18:00, On-Demand Platform

2025 IEEE 28th International Conference on Intelligent Transportation Systems (ITSC), November 18-21, 2025, Gold Coast, Australia

This information is tentative and subject to change. Compiled on April 2, 2026

Abstract

Trajectory data from connected vehicles (CV) has shown remarkable potential for adaptive traffic control. However, the inherent wireless nature of vehicle-to-infrastructure (V2I) communication exposes CV-based traffic control systems to cyberattacks, significantly impeding their large-scale deployment. Notably, False Data Injection attacks stand out as particularly stealthy and easy to implement, yet research on their impact in multi-intersection scenarios remains limited. To bridge this gap, this study proposes a Semi-Black-Box Attack targeting CV trajectory data. We adopt a sequential framework leveraging rolling horizon optimization and adversarial example generation, to enable effective attacks against adaptive network traffic control systems. Specifically, we first formulate the worst-case signal control planning as a nominal optimization problem, leveraging a rolling-horizon framework to enhance the long-term impact of these malicious decisions. Moreover, a black-box falsified trajectory generation algorithm is developed, realizing reverse mapping from malicious decisions to falsified trajectories using surrogate-model-based adversarial attack. Simulation experiments based on CityFlow demonstrate that the proposed semi-black-box attack achieves outstanding performance in metrics of efficacy and rapidity against both Max-Pressure (MP) and Reinforcement Learning (RL)-based signal control.