Lessons for the Construction of Military Simulators: A Comparison of Artificial Intelligence With Human-Controlled Actors

Abstract

Computer Generated Forces (CGF) are a significant component of military modelling and simulation. The creation of artificial agents with human-like decision-making is difficult. By using a Turing Test in a virtual environment that relates to military scenarios, areas in which CGF need to improve in order to help direct future studies are probed. The main areas determined for improvement are: environment awareness, human variance, persistence, vengeance anticipation, learning and teaming. The CGF show consistent weakness in these areas across all virtual environments and should be considered in future studies. They present a technical challenge in adoption of CGF as surrogate players in computer supported simulation activity for training mission rehersal, operations research and military experimentation.

Open PDF

Document Details

Document Type
Technical Report
Publication Date
Sep 01, 2004
Accession Number
ADA429112

Entities

People

  • Jennifer Sandercock

Organizations

  • Defence Science and Technology Group

Tags

Communities of Interest

  • Biomedical
  • Weapons Technologies

DTIC Thesaurus Topics

  • Artificial Intelligence
  • Combat Simulations
  • Computer Programs
  • Computer Science
  • Computer Simulations
  • Computers
  • Human Behavior
  • Human Systems Integration
  • Human-Machine Systems
  • Military Training
  • Operations Research
  • Personality
  • Simulators
  • Systems Science
  • Training
  • Video Games
  • Virtual Reality

Readers

  • Artificial Intelligence
  • Systems Analysis and Design
  • Team-Based Human-Centered Cognitive Task Decision Making and Information Performance.

Technology Areas

  • AI & ML
  • AI & ML - DoD AI Strategy