Urban traffic congestion requires intelligent solutions beyond traditional rule-based systems. This paper presents Neuro-Adaptive Swarm Control, a decentralized AI approach that combines swarm intelligence with adaptive neural networks to improve smart traffic management. Inspired by self-organizing biological systems, it enables real-time coordination between traffic signals, autonomous vehicles, and smart infrastructure without relying on centralized control. Using multi-agent reinforcement learning (MARL) and neural adaptation, the system allows traffic entities to learn, predict congestion, and optimize responses dynamically. Unlike conventional AI-based traffic control, which relies on precollected data, this approach continuously adapts to live traffic conditions, enhancing scalability and efficiency. Theoretical analysis and simulations demonstrate that this model reduces congestion, improves traffic flow, and minimizes energy consumption, making it a scalable and robust solution for future smart cities.


    Zugriff

    Zugriff prüfen

    Verfügbarkeit in meiner Bibliothek prüfen

    Bestellung bei Subito €


    Exportieren, teilen und zitieren



    Titel :

    Neuro-Adaptive Swarm Control: A Decentralized AI Approach for Intelligent Traffic Management


    Beteiligte:
    Shindhe, Nanditha (Autor:in) / M, Nandish (Autor:in) / Roy, Moumita (Autor:in)


    Erscheinungsdatum :

    26.06.2025


    Format / Umfang :

    451309 byte




    Medientyp :

    Aufsatz (Konferenz)


    Format :

    Elektronische Ressource


    Sprache :

    Englisch



    Intelligent agents in decentralized traffic control

    Ferreira, E.D. / Subrahmanian, E. / Manstetten, D. | IEEE | 2001


    Intelligent Agents in Decentralized Traffic Control

    Ferreira, E. D. / Subrahmanian, E. / Manstetten, D. et al. | British Library Conference Proceedings | 2001



    Multimodal adaptive traffic signal control: A decentralized multiagent reinforcement learning approach

    Kareem Othman / Xiaoyu Wang / Amer Shalaby et al. | DOAJ | 2025

    Freier Zugriff