This paper proposes a method which estimates the relationships between learner's behaviors and other agents' ones in the environment through interactions (observation and action) using the method of system identification. In order to identify the model of each agent, Akaike's Information Criterion is applied to the results of Canonical Variate Analysis for the relationship between the observed data in terms of action and future observation. Next, reinforcement learning based on the estimated state vectors is performed to obtain the optimal behavior. The proposed method is applied to a soccer playing situation, where a rolling ball and other moving agents are well modeled and the learner's behaviors are successfully acquired by the method. Computer simulations and real experiments are shown and a discussion is given.


    Access

    Check access

    Check availability in my library

    Order at Subito €


    Export, share and cite



    Title :

    State space construction for behavior acquisition in multi agent environments with vision and action


    Contributors:
    Uchibe, E. (author) / Asada, M. (author) / Hosoda, K. (author)


    Publication date :

    1998-01-01


    Size :

    824231 byte




    Type of media :

    Conference paper


    Type of material :

    Electronic Resource


    Language :

    English



    State Space Construction for Behavior Acquisition in Multi Agent Environments with Vision and Action

    Uchibe, E. / Asada, M. / Hosoda, K. et al. | British Library Conference Proceedings | 1998


    State and Action Space Construction Using Vision Information

    Kobayashi, Y. / Ota, J. / Inoue, K. et al. | British Library Online Contents | 2000


    Action-based state space construction for robot learning

    Asada, M. / Noda, S. / Hosoda, K. | British Library Online Contents | 1997


    Vision-Based Robot Learning for Behavior Acquisition

    Asada, M. / Nakamura, T. / Hosoda, K. et al. | British Library Conference Proceedings | 1995


    Vision based driving agent for race car simulation environments

    Bári, Gergely / Palkovics, László | ArXiv | 2025

    Free access