Goal reasoning (GR) is the ability for AI agents to dynamically reason about and change their goals to better align their activities with real-world conditions during execution. In previous work we have integrated existing GR models onboard unmanned underwater vehicles (UUVs) and demonstrated their ability to interact with UUV control systems and direct the UUVs to accomplish different goals based on sensor feedback from the environment. However, these GR models were based on discrete models of action, and providing them with dynamic, up-to-date cost estimation for UUV maneuvers was not straightforward. Further, these models used reactive, single-goal selection and did not consider the impact of changing goals on the long-term success of the mission. The objective of this project was to address these needs by developing a new GR model for use in embodied agents, and to demonstrate its efficacy in the UUV domain.


    Zugriff

    Zugriff über TIB

    Verfügbarkeit in meiner Bibliothek prüfen


    Exportieren, teilen und zitieren



    Titel :

    Goal Deliberation for UUV Control


    Beteiligte:
    M. A. Wilson (Autor:in) / D. W. Aha (Autor:in)

    Erscheinungsdatum :

    2021


    Format / Umfang :

    9 pages


    Medientyp :

    Report


    Format :

    Keine Angabe


    Sprache :

    Englisch