Autonomous vehicles (AVs) enhance driving efficiency and reduce accidents but require robust risk assessment methods due to dense traffic and uncertainties. Existing methods rely on predefined rules, which lack generalization. This paper presents a novel risk quantification method without expert rules, leveraging reinforcement learning and adversarial agents. The proposed model uses Gated Transformer Networks for multivariate time series regression, analyzing historical traffic data to generate continuous risk assessments. Simulation experiments validate the method's efficacy, demonstrating its precision and robustness.


    Access

    Check access

    Check availability in my library

    Order at Subito €


    Export, share and cite



    Title :

    A Data-Driven Risk Assessment Method for Autonomous Vehicles Without Expert Rule Design


    Contributors:
    Wang, Caojun (author) / Yang, Shuo (author) / Huang, Yanjun (author)


    Publication date :

    2024-09-24


    Size :

    1245304 byte





    Type of media :

    Conference paper


    Type of material :

    Electronic Resource


    Language :

    English



    Expert-driven Rule-based Refinement of Semantic Segmentation Maps for Autonomous Vehicles

    Manibardo, Eric L. / Lana, Ibai / Del Ser, Javier et al. | IEEE | 2023


    Risk assessment system for autonomous vehicles and method thereof

    SHIN IN SUP / HWANG EUN JI | European Patent Office | 2024

    Free access

    Expert systems, a tool for autonomous underwater vehicles

    Blidberg, D.R. / Westneat, A.S. / Corell, R.W. | Tema Archive | 1983


    False discrimination of rule violations for autonomous vehicles for data processing

    KLEEN ANDREAS / DUINJER TEBBENS RICHARD / BIN-NUNN AVI et al. | European Patent Office | 2025

    Free access

    DISCERNING FAULT FOR RULE VIOLATIONS OF AUTONOMOUS VEHICLES FOR DATA PROCESSING

    COLLIN ANNE / TEBBENS RADBOUD DUINTJER / BIN-NUN AMITAI et al. | European Patent Office | 2023

    Free access