A pyramid-structured platform for integrated military command and full-cycle course-of-action development
"At the heart of a national AI defense strategy: an ontology knowledge base of field manuals"
Ai-Commander is a command platform that combines ProActive Multi-Agent LLM/VLM technology with an ontology knowledge base.
The AI recognizes the objective for itself and runs a repeating loop of
plan, execute, re-plan
to arrive at the best course of action.
To move beyond what a single AI model can do, multiple agents work together, distributing the task among themselves.
The ontology knowledge base is used throughout, in both collaboration and autonomous planning.
It provides a
high-performance real-time wargame simulation environment in which red and blue forces face each other.
AI planning sets out the operation order (OPORD) for subordinate units as a concrete sequence of actions over time, supporting the commander's decision.
That same sequence of actions applies equally to soldiers and to combat robots.
It guarantees high accuracy with no hallucination
in the courses of action it generates, and stability in the system itself.
It draws on a flawless ontology knowledge base, built by extracting the essentials automatically from 192 army field manuals.
It was designed to give the same unit action plan
to manned-unmanned teams as well.
Beyond commanding soldiers, it can produce action plans that take in unmanned aerial vehicles (UAVs) and combat robots.
It stands to change how dedicated units are taught to develop courses of action.
