This web page's content and links are no longer actively maintained. It is available for reference purposes only. NASA Official: Cynthia Cheung
 

AI Agent Characteristics

The ANTS architecture is inspired by the success of social insect colonies, a success based on the division of labor within the colonies in two key ways: 1) within their specialties, individual specialists generally outperform generalists, and 2) with sufficiently efficient social interaction and coordination, the group of specialists generally outperforms the group of generalists. Thus systems designed as ANTS are built from potentially very large numbers of highly autonomous, yet socially interactive, elements.

The architecture is self-similar in that elements and sub-elements of the system may also be recursively structured as ANTS on scales ranging from microscopic to interplanetary distances. Let a tetrahedron represent the basic organizational structure, where worker operational nodes are connected to neural heuristic node. The same organizational structure found among worker subsystems connected to a neural nodes to form systems within a spacecraft is also seen as spacecraft specialist workers connected via a heuristic leader to form a subgroup. Subgroups are connected in a similar way to form subswarms or workers, and subswarms to form swarms.

The study of remote multi-body or high surface area targets in inaccessible locations represents challenges unachievable by conventional (direct control of single craft from centralized location) or even evolutionary (direct control of multi-craft from control surrogate) mission design. Such targets include surveys of extreme environments on the Earth, Moon, or Mars, asteroid, comet, or dust popultions. The revolutionary ANTS paradigm makes the achievement of such goals possible through the use of many small, autonomous, reconfigurable, redundant element craft acting as independent or collective agents.

Agent Concept Graphic
Home Page