Abstract
Creating coordinated interactions for multi-agent systems, such as physical robots or virtual characters, can present challenges in synchronizing expressive dialogue with coherent physical actions and adapting to unscripted events. The disclosed technology describes a multi-agent orchestration engine with a decoupled architecture that can employ a tiered timeline system to manage a global narrative and individual agent states. A dynamic interruption handler can leverage a large language model to reason about real-time inputs and determine if an agent should deviate from a scripted path. The engine can then generate synchronized dialogue and high level behavioral tokens, which may be subsequently translated into specific kinematic actions. This approach can provide a framework for driving responsive and coordinated agent behaviors across physical, virtual, and simulated environments, potentially allowing for adaptation to changing conditions while maintaining narrative structure.
Creative Commons License
This work is licensed under a Creative Commons Attribution 4.0 License.
Recommended Citation
Garcia, Allison; Cunningham, Corbin; and Dory, Mike, "Multi-Agent Orchestration Engine for Synchronized Generative Dialogue and Kinematics", Technical Disclosure Commons, ()
https://www.tdcommons.org/dpubs_series/11774