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

Sensing and communication technologies have enhanced learning-based decisionmaking methodologies for multi-agent systems such as connected autonomousvehicles (CAV). However, most existing safe reinforcement learning basedmethods assume accurate state information. It remains challenging to achievesafety requirement under state uncertainties for CAVs, considering the noisysensor measurements and the vulnerability of communication channels. In thiswork, we propose a Robust Multi-Agent Proximal Policy Optimization with robustSafety Shield (SR-MAPPO) for CAVs in various driving scenarios. Both robustMARL algorithm and control barrier function (CBF)-based safety shield are usedin our approach to cope with the perturbed or uncertain state inputs. Therobust policy is trained with a worst-case Q function regularization modulethat pursues higher lower-bounded reward in the former, whereas the latter,i.e., the robust CBF safety shield accounts for CAVs' collision-freeconstraints in complicated driving scenarios with even perturbed vehicle stateinformation. We validate the advantages of SR-MAPPO in robustness and safetyand compare it with baselines under different driving and state perturbationscenarios in CARLA simulator. The SR-MAPPO policy is verified to maintainhigher safety rates and efficiency (reward) when threatened by both stateperturbations and unconnected vehicles' dangerous behaviors.

Safe and Robust Multi-Agent Reinforcement Learning for Connected Autonomous Vehicles under State Perturbations