English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Plus annual

Presents the access of premium content as premium feature

Premium Content

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Plus monthly

Global: Zendy Tools annual

Global: Zendy Tools monthly

Global: Zendy Free Plan

The rapid advancements of Internet of Things (IoT) and artificialintelligence (AI) have catalyzed the development of adaptive traffic signalcontrol systems (ATCS) for smart cities. In particular, deep reinforcementlearning (DRL) methods produce the state-of-the-art performance and have greatpotentials for practical applications. In the existing DRL-based ATCS, thecontrolled signals collect traffic state information from nearby vehicles, andthen optimal actions (e.g., switching phases) can be determined based on thecollected information. The DRL models fully "trust" that vehicles are sendingthe true information to the signals, making the ATCS vulnerable to adversarialattacks with falsified information. In view of this, this paper first timeformulates a novel task in which a group of vehicles can cooperatively sendfalsified information to "cheat" DRL-based ATCS in order to save their totaltravel time. To solve the proposed task, we develop CollusionVeh, a generic andeffective vehicle-colluding framework composed of a road situation encoder, avehicle interpreter, and a communication mechanism. We employ our method toattack established DRL-based ATCS and demonstrate that the total travel timefor the colluding vehicles can be significantly reduced with a reasonablenumber of learning episodes, and the colluding effect will decrease if thenumber of colluding vehicles increases. Additionally, insights and suggestionsfor the real-world deployment of DRL-based ATCS are provided. The researchoutcomes could help improve the reliability and robustness of the ATCS andbetter protect the smart mobility systems.

Attacking Deep Reinforcement Learning-Based Traffic Signal Control Systems with Colluding Vehicles