Swarm Technology at NASA: Building Reslient Systems

The design and development of modern aerospace systems is an inherently complex task driven by standards and safety requirements where complex hardware and sophisticated software must exhibit adequate reliability. Such systems are complex conglomerates of components where control software drives rigid hardware. In addition to careful design and thorough testing, to increase their survivability, NASA has pioneered special self-management features using various AI techniques including swarm computing. Swarm-based spacecraft systems, based on the cooperative nature of a hive culture, provide a new robust way of space exploration. A system of this class is composed of multiple self-organizing and autonomous spacecraft. The design and implementation of such systems requires new engineering approaches. We briefly recap the history of special AI approaches to spacecraft design, particularly swarm computing, review the current status of these efforts at NASA (namely, ANTS), and briefly describe the potential derivative uses in civilian and military applications.