Automated Support for Human Decision and Control in Military Systems with Varying Levels of Autonomy

From the perspective of the human commander of a complex automated process, an important factor that influences the quality of the command dialog is the ability of the user interface to provide information that allows the commander to perform situation assessment. Another important factor is the ability of the user interface to provide accessible and usable mechanisms to support communication between the human and the contol system. We believe that much of the substrate for such human-machine dialog may be provided by a dynamic, model-based representation of the system components and processes, shared by all software and human agents participating in the control situation. In addition, we believe that the current method of providing user interactions to dynamic control systems, which involves static predesign of all user interface components and interactions, is inadequate to meet the needs of multi-agent systems requiring varying levels of participant autonomy and varying levels of mixed-initiative control by human commanders, especially those in dynamically changing situations such as those present in a military battlefield. We have developed an approach that overcomes these limitations, using model-based productive reasoning about situations and interaction design, and are successfully applying this approach in the domain of military command and control of teams of semi-autonomous vehicles.