An Operational Model for Finite State Machine Replanning in MODSAF
Abstract
The United States Army is tasked by the nation to provide an army that is disciplined, well trained, well-equipped, and well-led, capable of deploying anywhere in the world and winning decisively. At the tactical level of war, this translates to individual commanders preparing their units to fight and defeat an enemy. In order to prepare to accomplish this task, commanders must train their units under realistic conditions. Given the fiscal constraints placed on the United States military, simulations provide an opportunity for leaders to utilize current training methodology to train themselves and their staffs for wartime missions at minimum cost. An essential component of a simulation is the Computer-Generated Force (CGF) that attempts to replicate realistic human and physical behaviors in the synthetic battlefield. Although there are many different CGFs in existence today, this thesis will focus on how ModSAF replicates and implements human behavior. Currently, there is no organic behavioral model in ModSAF that conducts self-modification of tasks based purely on observation of the synthetic battlefield. Existing models are adequate to train perfunctory, rote tasks but do not challenge military leaders under conditions that are found during combat. A human operator sitting at a computer workstation is necessary to make behavioral decisions for subordinate computer generated entities. This does not represent how a military leader would conduct these tasks in a live tactical situation. This thesis will propose an operational model for ModSAF that will allow a tank platoon to autonomously replan its behavior based on observation of its synthetic battlespace.
Document Details
- Document Type
- Technical Report
- Publication Date
- Jan 01, 2000
- Accession Number
- ADA376477
Entities
People
- John S. Kolasheski
Organizations
- University of Central Florida