Date of Award

3-11-2011

Document Type

Thesis

Degree Name

Master of Science

Department

Department of Electrical and Computer Engineering

First Advisor

Brett Borghetti, PhD.

Abstract

Military organizations need realistic training scenarios to ensure mission readiness. Developing the skills required to differentiate combatants from non-combatants is very important for ensuring the international law of armed conflict is upheld. In Simulated Training Environments, one of the open challenges is to correctly simulate the appearance and behavior of combatant and non-combatant agents in a realistic manner. This thesis outlines the construction of a data driven agent that is capable of imitating the behaviors of the Virtual BattleSpace 2 behavior classes while our agent is configured to advance to a geographically specific goal. The approach and the resulting agent promotes and motivates the idea that Opponent and Non-Combatant behaviors inside of simulated environments can be improved through the use of behavioral imitation.

AFIT Designator

AFIT-GCO-ENG-11-04

DTIC Accession Number

ADA540365

Download

Share

COinS