As urban landscapes expand and the vehicular population surges, the resultant traffic congestion and accident rates present formidable challenges. Traditional traffic management, reliant on manual intervention, rudimentary surveillance, and static signage, is ill-equipped to manage the intricate dynamics of urban traffic networks, especially during peak traffic periods. This paper delves into the development of cooperative control algorithms that aim to refine traffic flow and curtail accidents within multi-agent settings. We concentrate on methodologies such as Cooperative Adaptive Cruise Control (CACC) and Multi-Agent Reinforcement Learning (MARL), which are adept at gauging environmental conditions and vehicular behaviors to enact optimal, collaborative decisions. The algorithms are designed to harmonize individual vehicular goals, such as travel time minimization, with overarching system objectives, including congestion alleviation and safety enhancement. Employing a simulation model grounded in SUMO (Simulation of Urban Mobility), this study offers a meticulous platform for vetting and dissecting diverse traffic management strategies. By embedding sophisticated cooperative control algorithms within the simulation, the research endeavors to manifest the capacity for superior traffic management in intricate urban contexts, thereby fostering safer and more efficient transportation ecosystems.
Cooperative control algorithms for traffic congestion mitigation and accident prevention in multi-agent environments
Workshop on Electronics Communication Engineering (WECE 2024) ; 2024 ; Wuhan, China
Proc. SPIE ; 13553 ; 1355308
12.03.2025
Aufsatz (Konferenz)
Elektronische Ressource
Englisch
Machine Learning Based Traffic Congestion and Accident Prevention Analysis
Springer Verlag | 2024
|A multi-agent architecture for cooperative inter-jurisdictional traffic congestion management
Online Contents | 2002
|Europäisches Patentamt | 2024
|