We present a novel hybrid learning method, named HyLEAR, for solving the collision-free navigation problem for self-driving cars in POMDPs. HyLEAR leverages interposed learning to embed knowledge of a hybrid planner into a deep reinforcement learner to faster determine safe and comfortable driving policies of the car. In particular, the hybrid planner combines pedestrian path prediction and risk-aware path planning with driving-behavior rule-based reasoning such that the determined safe trajectories also take into account, whenever possible, the ride comfort and a given set of driving-behavior rules. Our experimental performance analysis over the CARLA-CTS benchmark of critical traffic scenarios revealed that HyLEAR can significantly outperform the selected baselines in terms of safety and ride comfort.


    Zugriff

    Zugriff prüfen

    Verfügbarkeit in meiner Bibliothek prüfen

    Bestellung bei Subito €


    Exportieren, teilen und zitieren



    Titel :

    HyLEAR: Hybrid Deep Reinforcement Learning and Planning for Safe and Comfortable Automated Driving


    Beteiligte:


    Erscheinungsdatum :

    04.06.2023


    Format / Umfang :

    3771999 byte





    Medientyp :

    Aufsatz (Konferenz)


    Format :

    Elektronische Ressource


    Sprache :

    Englisch



    Hybrid Deep Reinforcement Learning and Planning for Safe and Comfortable Automated Driving

    Gupta, Dikshant / Klusch, Mathias | ArXiv | 2022

    Freier Zugriff

    Minimizing Safety Interference for Safe and Comfortable Automated Driving with Distributional Reinforcement Learning

    Kamran, Danial / Engelgeh, Tizian / Busch, Marvin et al. | DataCite | 2021

    Freier Zugriff


    Safe, Efficient, and Comfortable Velocity Control based on Reinforcement Learning for Autonomous Driving

    Zhu, Meixin / Wang, Yinhai / Pu, Ziyuan et al. | ArXiv | 2019

    Freier Zugriff

    Safe and Comfortable Motion Planning for Bumpy Road Driving

    Zhang, Jiajie / Xiao, Bin / Tang, Binbin et al. | IEEE | 2023