Route choice modeling is an important issue in transportation planning. In recent years, the importance of comfortable urban spaces not only for vehicles but also for pedestrians has been increasing toward low-carbon society. Accordingly, efforts have been made to model the pedestrians' characteristic behaviors. However, most of the existing studies about route choice modeling are limited to the representation within a single traffic mode. In a typical urban street space, each traffic mode is interacting with each other in a complex manner. Describing multimodal interaction is necessary to efficiently allocate the limited road space and to maximize the value of the road space as a result of the network effect. However, there are two challenges to describing the interaction of multiple transportation modes in a route choice model: the computational cost of model estimation and the limited expressive power of the model. In this study, we construct a route choice model based on the AIRL method which can rapidly estimate the user equilibrium under multi-agent Markov game and can take into account the interaction among multiple traffic modes explicitly. The use of highly expressive machine learning models is also expected to lead the high reproducibility of traffic behavior.


    Access

    Check access

    Check availability in my library

    Order at Subito €


    Export, share and cite



    Title :

    Enhancing Joint Behavior Modeling with Route Choice Using Adversarial Inverse Reinforcement Learning


    Contributors:
    Ogawa, Daichi (author) / Hato, Eiji (author)


    Publication date :

    2023-09-24


    Size :

    247511 byte





    Type of media :

    Conference paper


    Type of material :

    Electronic Resource


    Language :

    English



    Modeling Driver Behavior using Adversarial Inverse Reinforcement Learning

    Sackmann, Moritz / Bey, Henrik / Hofmann, Ulrich et al. | IEEE | 2022



    Multiagent modeling of pedestrian-vehicle conflicts using Adversarial Inverse Reinforcement Learning

    Nasernejad, Payam / Sayed, Tarek / Alsaleh, Rushdi | Taylor & Francis Verlag | 2023


    Predicting Driver Behavior on the Highway with Multi-Agent Adversarial Inverse Reinforcement Learning

    Radtke, Henrik / Bey, Henrik / Sackmann, Moritz et al. | IEEE | 2023