Issues in fidelity, architecture, and security for computer generated actors in network centric warfare simulations

The US military is in the midst of a far-reaching and in many ways unprecedented change in its philosophy, approach, and technologies used for warfare. The old days of massed formations, accurate but large weapons, and movement is being replaced by small formations (as small as one person), powerful, small but extremely accurate weapons, and high speed movement all augmented by rapid movement of information between and among all military components in the battlespace and from the battlespace back to rear areas and even to the logistics depots. For this future vision of high-speed, highly mobile warfare to be achieved and be effective there are a number of issues that must be addressed and solved in the real-world and in the world of simulation. In this paper, we will examine the impact that this switch in warfare paradigms will have on the world of simulation and focus particularly on the impact that this change will have on computer-generated actors and the new demands that will be placed upon them, both in the portrayal of friendly and enemy combatant forces. Clearly, these solutions will build upon much work performed before. Additionally, these solutions must draw upon experience and results in the fields of human behavior modeling, cultural modeling, modeling of adversarial decision making, and modeling of operations other than war in order to correctly model the friendly small units and enemy activity and responses within the context of simulations used for a variety of purposes ranging from training to acquisition. The new CGAs will have to be much more flexible in their planning and actions, more flexible in their software architectures, capable of more autonomous behavior, and be able to manage unexpected circumstances with minimal human intervention. In short, a new breed of CGAs are needed, ones that are more autonomous, capable, flexible, and "intelligent" than currently available CGAs. In addition, their knowledge bases require a degree of protection and security that is also unprecedented, especially when coupled with the need for ease of expansion and modification of the knowledge bases. The paper is organized as follows. Section one contains a discussion of and motivation for our research. Section two holds a brief summary of previous work. Se- ction three presents a discussion of the architecture and security issues to be addressed and potential solutions for simulation CGAs. Section four contains a brief summary of the paper and suggestions for further research.