Application of Autonomous Agent Concepts to Robotic Fleet Dynamics.

摘要

The fighting effectiveness and force projection capability of the U. S. Army's Future Force will be significantly increased with deployment of Future Combat Systems (FCS), which include significant numbers of unmanned vehicles and robotic systems to conduct missions ranging from reconnaissance to re-supply to direct assault. Successful application of these robotic systems requires new levels of objective-based autonomous behavior in mixed-fleet scenarios. To meet these requirements for robotic fleet dynamics, designers and integrators should apply autonomous agent concepts and recent research results. As the concepts are demonstrated and improved in challenging environments, they can be further extended to additional applications for the military as well as new ones in the commercial/civil sector. This paper defines the autonomous agent concept and describes its important elements, including goal seeking, resourcefulness, and the optional ability for agents to transfer themselves or clone themselves among physical hosts. Fleet dynamics are discussed as applicable to robot coordination in group maneuvers, complex missions, and rapidly changing conditions and constraints. Multiple aspects of fleet dynamics are considered including distinction of multiple roles, transfer of roles under triggering conditions, and the impact of various strategies of information sharing, whether local, near neighbor, or wide-area broadcast, on fleet performance, robustness, and adaptability. This paper includes a review of promising research results for autonomous agent concepts in academic and military settings and identifies specific methods that appear useful for application to robotic and mixed fleets. Several military and commercial/civil applications are identified for which the autonomous agent concept is especially useful. Technology gaps where additional research and development is justified to overcome current limitations are highlighted.