Achievability Control Theory for Supervisory Computer-Human Systems

The objective of the proposed project is the development of technology for more efficient and effective human-computer supervision of complex systems. Systems that combine humans and automation in a synergistic or cooperative manner may be termed hybridsystems. Hybrid systems offer advantages over both purely automated systems and purely manual systems in many circumstances. However, future hybrid systems will be even more complex than contemporary ones. This gives rise to a serious need to developmethods for integrating humans more closely-and more efficiently-than is possible now within hybrid systems. We will achieve this by developing Achievability Control Theory (ACT), an innovative extension of Supervisory Control Theory. The ACT approach haspotential to enhance both the efficiency and flexibility of hybrid systems.During Phase I we will provide a proof of concept by developing the formalized control theory necessary to integrate achievability within a supervisory control framework. Specifically, we will consider the special case when a human participates in a hybridsystem. Successful completion of the proposed research will (1) enhance the flexibility and efficiency of future hybrid systems (including battlefield robots), in turn enhancing the mission success rate, robustness, and survivability; (2) support optimalintegration of humans and computer supervisors in future missions; (3) enhance mission planning for hybrid systems, and (4) guide other research and development by identifying problematic areas within particular missions and by identifying problematicfunctions for hybrid systems generally.