Creating a Tactical Pilot’s Assistant for Combat Operations in Contested Denied Environments: An Overview of Three Different Approaches and Results
As our adversaries become more capable in their development of advanced systems for combat, those systems themselves are becoming increasingly complex as well. The future battlespace is expected to be one where the mechanisms and tools we depend on today for the execution of a combat mission will not be reliable in terms of the real world representations they provide, given creative manipulation of sources and data. Less elegant approaches to contesting and denying operations merely take a given source of data or capability away. A far more sophisticated approach is one that does very subtle manipulation of key data and representations, such that even the ops systems may not detect miscorrelation or suspect data, thus requiring the human operator, under stress to do those detections and assessments. There may also be such a variety of information that the human operator will not be able to evaluate the quality or “truth” of the data in real time, and may be forced to make decisions on deficient or suspect sources and data. Further, the density of data from the variety of sensors and sources at hand will potentially overwhelm human operators just dealing with the regular influx of data. With this in mind, a major goal of this effort is to examine the extent to which we can use emerging human behavior and machine learning models to create software-based pilot’s assistance that can help offload some of the source and data monitoring from the human operator and to alleviate cognitive bottlenecks in the midst of a combat engagement. Three different approaches to the development of an agent-based assistant will be described and discussed across this and the three other companion papers. This initial paper will serve to frame the other presentations and provide a practical foundation for the approaches that have been undertaken. This presentation will highlight the key activities, challenges and development efforts and will describe the applications spaces we used for demonstration. The capstone demonstration where each of the approaches will be evaluated will be described and discussed. Our future plans for integrating the approaches and developed models into an F16 tactical training environment will be discussed as well.