Distributed Virtual Environments (DVEs) must continue to perform well as users are added. However, DVE performance can become sensitive to user behavior in many ways their actions, their positions, and even the direction that they look. These behavioral elements are important for evaluating virtual terrains. While two terrains may be similar in terms of user experience, task efficiency, immersion, and even aesthetics, they may exhibit substantially different performance out of the DVE when many users are logged in.

Distributed Virtual Environment systems are complex systems that include graphics, physical simulation, and networked state synchronization. The system performance depends on all these components balancing the load and while sharing available computational, graphic, and network resources. We present an approach to analysis of the load balance of load and resource usage.

We describe a simple load-measure-model method for analyzing the scalability of Distributed Virtual Environments (DVEs). We use a load simulator and three metrics to measure a DVE's engine with varying numbers of simulated users. Our load simulator logs in as a remote client and plays according to how users played during the conducted user study. Two quality of virtuality metrics, fidelity and consistency, describe the user's experience in the DVE. One engine performance metric provides the cycle time of the engine's primary loop. Simulation results (up to 420 users) are discussed.

This paper describes a 3D, tangible user interface system and related interaction techniques for the CAVE environment. The developed system is based on off-the shelf software and hardware components to provide 3D input data for CAVE applications. A CAVE with four sides (three walls and a floor) is used as a display and interaction space. Interaction tasks (selection, manipulation, navigation) are performed using interaction techniques based on manipulation of physical objects (props). All virtual objects are directly manipulated using the corresponding props to which Augmented Reality physical markers are attached. Each physical marker corresponds to a specific virtual object. The floor projection (a white rectangle overlaid on top of the application generated video stream) or a directional light create illumination necessary for computer vision based marker detection using ARToolKit. Initial evaluation results are positive and provide directions for future research.

The Cognitive Agent Architecture (Cougaar) is a distributed agent architecture framework originally developed to build large scale logistics applications for the Department of Defense. In order to facilitate rapid development of complex military and commercial Cougaar applications, we are devising "recipes" on how to automate the Cougaar application development process. The envisioned Cougaar Model Driven Architecture (CMDA) system will capture requirements from the user and automatically generate related software artifacts such as requirement, design, code and test cases. The automated generation of software artifacts, even if partial in nature, should considerably increase the productivity of Cougaar developers. Recognizing the sophistication of the Cougaar framework and the limitations of current transformation technologies, we have systematically developed an approach that combines component assembly in the large with transformations in the small. The approach’s ability to generate accurate software artifacts depends largely on how user requirements are captured and transformed. The CMDA system employs innovative approaches and techniques to enable users represent their requirements properly. This paper briefly describes the CMDA system, its user interface and the various inputs captured from the user.