
Capturing Military Doctrine in Behavior Trees Using LLMs | Unreal Fest Orlando 2025
In this session recorded at Unreal Fest Orlando 2025, Paul Foster of Lockheed Martin and Gita Sukthankar of UCF discuss in-progress work of an Unreal Engine plugin that provides an interactive chat that works alongside the Behavior Tree editor in the Unreal Editor.
Military doctrine refers to the guidelines that govern how military forces participate in various types of operations, from campaigns and battles to engagements. By leveraging large language models (LLMs), developers can create a model that understands these guidelines and assists developers in creating behavior trees or state trees that accurately capture military doctrine within the Unreal Engine Game AI framework, including real-world scenarios for training.
By building on research from the University of Central Florida and utilizing behavior tree examples from the robotics industry, we can make the authoring of behavior trees for individual agents and multi-agent systems more efficient and realistic, and less labor-intensive in terms of testing and validation.
Watch this video for an understanding of how LLMs assist in the generation of behavior trees in Unreal Engine, and learn about techniques for getting the most from trained LLMs.
Learn more about Unreal Engine for simulation here: unrealengine.com/uses/simulation
#MilitarySimulation, #LockheedMartin, #UCF, #LargeLanguageModels, #Robotics, #BehaviorTreeEditor, #GameAIFramework, #UnrealEngine, #UE5, #UnrealFestOrlando
Military doctrine refers to the guidelines that govern how military forces participate in various types of operations, from campaigns and battles to engagements. By leveraging large language models (LLMs), developers can create a model that understands these guidelines and assists developers in creating behavior trees or state trees that accurately capture military doctrine within the Unreal Engine Game AI framework, including real-world scenarios for training.
By building on research from the University of Central Florida and utilizing behavior tree examples from the robotics industry, we can make the authoring of behavior trees for individual agents and multi-agent systems more efficient and realistic, and less labor-intensive in terms of testing and validation.
Watch this video for an understanding of how LLMs assist in the generation of behavior trees in Unreal Engine, and learn about techniques for getting the most from trained LLMs.
Learn more about Unreal Engine for simulation here: unrealengine.com/uses/simulation
#MilitarySimulation, #LockheedMartin, #UCF, #LargeLanguageModels, #Robotics, #BehaviorTreeEditor, #GameAIFramework, #UnrealEngine, #UE5, #UnrealFestOrlando
Unreal Engine
Welcome to the official Unreal Engine channel!
Unreal Engine is a complete suite of real-time 3D tools made by developers, for developers. Create and take your content anywhere, from PC, console, mobile and VR games to cinematic experiences, visualizatio...