A Shared Intelligence Framework Targeting Flexible Human Agent Cooperative Tasks

We present an innovative solution: a shared intelligence framework targeting flexible human agent cooperative tasks employing both prior arts and the most updated research ideas in multi-agent systems and human agent interactions. By ‘flexible’, we mean that the level of human agent collaboration is not static; instead, it is realized through negotiated task decomposition and allocation among team members, which ensures that human experts will always be updated with an appropriate level of information and could always take control of their task assignments with the collaboration of software agents in coordinated problem solving processes. The following topics are explored in detail: the representation of team members’ roles and capabilities, agents’ tasks and environments, dynamic team formation and task allocation, appropriate agent architecture and multi-agent system architecture, coordination among human experts and software agents through task planning agent (TPA), coordination communication protocols, and task structure analysis for managing interdependencies among agents’ tasks, and finally, the application of the aforementioned general technical solution to a military medical service scenario. The proposed flexible human agent collaboration framework will be well suited to interface with resources and external systems, such as TMIP and information systems of MV-T of future combat systems.