An Agent Populated Testbed for C2 Experimentation

Abstract

Public reporting burden for this collection of information is estimated to average 1 hour per response, including the time for reviewing instructions, searching data sources, gathering and maintaining the data needed, and completing and reviewing the collection of information. i ABSTRACT Recent concepts in the field of command and control (C2), such as Power to the Edge and Network-Centric Warfare, have indicated the need for a testbed for experimentation. We describe a gaming testbed, populated by realistic synthetic agents, for modeling the complex human interactions comprising C2 structures, and for exploring the effectiveness of C2 concepts in a variety of tactical circumstances. SPONSOR/MONITOR'S ACRONYM(S) 9. SPONSORING/MONITORING AGENCY NAME(S) AND ADDRESS(ES)AFRLA testbed for experimentation with C2 concepts must be capable of reproducing the complex, rapid, dynamic, and unpredictable unfolding of events in battle. Scenarios can be carefully limited, and realism diluted to a practical level, but enough of the complex interactions of individual and collective behavior must be preserved to capture the essentially human nature of C2. Experimentation involves the correlation of tactical events and outcomes with a host of variable parameters, such as C2 structure, communication patterns and reliability, and implementation of commander's intent.This report describes the details of our testbed and experiment design and implementation, in addition to findings and lessons learned. Our approach included the integration and extension of three existing platforms: Counter-Strike, a multiplayer firstperson tactical shooter; SimBionic, a visual behavior authoring and execution engine developed at Stottler Henke; and ADaM, a data mining tool suite developed at the University of Alabama, Huntsville. The testbed produced by this integration allows for the planning and execution of missions by human or synthetic agents, including specifications of different C2 structures, and the mining of experimental data for the effectiveness of the C2 concepts. We discuss the applicability of our findings to other echelons, as well as its contribution to the evaluation of new C2 concepts.ii