For autonomous vehicles to safely share the road with human drivers, autonomous vehicles must abide by specific "road rules" that human drivers have agreed to follow. "Road rules" include rules that drivers are required to follow by law -- such as the requirement that vehicles stop at red lights -- as well as more subtle social rules -- such as the implicit designation of fast lanes on the highway. In this paper, we provide empirical evidence that suggests that -- instead of hard-coding road rules into self-driving algorithms -- a scalable alternative may be to design multi-agent environments in which road rules emerge as optimal solutions to the problem of maximizing traffic flow. We analyze what ingredients in driving environments cause the emergence of these road rules and find that two crucial factors are noisy perception and agents' spatial density. We provide qualitative and quantitative evidence of the emergence of seven social driving behaviors, ranging from obeying traffic signals to following lanes, all of which emerge from training agents to drive quickly to destinations without colliding. Our results add empirical support for the social road rules that countries worldwide have agreed on for safe, efficient driving.


    Zugriff

    Download


    Exportieren, teilen und zitieren



    Titel :

    Emergent Road Rules In Multi-Agent Driving Environments


    Beteiligte:
    Pal, Avik (Autor:in) / Philion, Jonah (Autor:in) / Liao, Yuan-Hong (Autor:in) / Fidler, Sanja (Autor:in)

    Erscheinungsdatum :

    2020


    Anmerkungen:

    International Conference on Learning Representations (2021)
    International Conference on Learning Representations, 2021


    Medientyp :

    Preprint


    Format :

    Elektronische Ressource


    Sprache :

    Englisch




    Driving risk monitoring method for emergent target

    KONG HUIFANG / WANG CHENSHUN / ZHANG QIAN et al. | Europäisches Patentamt | 2024

    Freier Zugriff


    Intelligent driving model considering vehicular dynamics and heterogeneous road environments

    Zhang, Jiahao / Qi, Hongsheng | Taylor & Francis Verlag | 2025


    Road Traffic Signs Recognition in Automatic Driving under Extreme Environments

    Chi, Juntian / Feng, Zhuangzhou / Zhang, Weipeng | VDE-Verlag | 2022


    Road Traffic Management With a Multi-agent System for Truck and Car Driving Simulators

    Tran, J.-M. | British Library Conference Proceedings | 1995