Thomas C. Hughes and Cynthia D. Martin
Veridian, Veda Operations
Cynthia Dominguez and Michael Patzek
Air Force Research Laboratory/Human Effectiveness Directorate
John M. Flach
Wright State University
For several years we've been working on integrating domain information into the design process. There has been increased emphasis on "up-front" analysis to gain a more complete understanding of the operational domain so subsequent design decisions can be justified based on their impact on system performance within a given mission context. Additionally, pressures to automate new systems to reduce manpower costs and increase system reliability have resulted in design of systems that are extremely difficult for operators to understand and control. Such conditions are often the result of insufficient insight into operational domains and inadequate consideration of the relative contributions to successful operation that human operators and intelligent automation techniques might provide. With the emergence of Uninhabited Air Vehicles (UAV) we run the risk of failing into the same trap. Some suggest that the default philosophy for UAVs is to automate everything. This leaves the operator to monitor the system and intervene as mission situations demand. Such approaches are fundamentally flawed because they fail to recognize and capitalize upon the unique contribution that the human operator provides, particularly in highly complex and dynamic work domains such as those anticipated with UAVS. The present paper describes our approach toward a design process that attempts to incorporate detailed domain information into the design that will hopefully lead to an integrated, highly coupled collaborative system interface for UAV system operators.
One Air Force mission in which UAVs are seen as particularly appropriate is the Suppression of Enemy Air Defenses (SEAD). The objective of the SEAD mission is to neutralize, destroy, or temporarily degrade enemy surface based air defenses by destructive and/or disruptive means. The use of a UAV as opposed to a traditional manned system is a particularly attractive alternative due to the extremely high attrition rates typically associated with the SEAD mission. The application of Uninhabited Combat Air Vehicles (UCAV) in a SEAD mission could effectively employ sensors and weapons in high threat areas where vulnerability of manned aircraft represents an unacceptable risk.
Piloted systems are extremely complex, dynamic and full of risk. UAVs add to this complexity with the challenge of remote operations. In addition, a UCAV must be viewed as a collaborative system in that its application will involve coordination among multiple intelligent agents. These agents will include both human operators and automatic control systems. A UCAV system is an example of what Sheridan ( 1997) would call a multiple task telerobotics control system. The remoteness of the operator relative to the battle and the autonomous capability of the UCAV are principle challenges relative to effective collaboration, and thus unique challenges for design. The UCAV is also an example of an adaptive system because it needs to function in a nondeterministic, dynamically changing environment. Combat is not a static encounter, but a dynamic interaction between two competing agents. The environment that exists when a mission is planned