Teaming With Self-Integrating Systems

Expecting systems to be able to integrate themselves into an existing or nascent team assumes they can provide a useful function for the team, and changes what kinds of capabilities they must have, in addition to what the team must be able to do, beyond those that make them desired or potentially useful team members. This includes at least operational knowledge about their own capabilities, limitations, and their knowledge thereof. Completely new classes of environment-and self-modeling, self-assessment, and behavior negotiation processes are also needed, along with continual self-and boundarymonitoring processes to ensure that the integration assumptions remain valid. This paper takes one class of example application problem domains (search / surveillance), describes a design process for such a system, and what must be changed in the SISSY (Self-Improving Self-Integrating) system context.