Large Scale Distributed Computation for Human Agent Team Coordination in Dynamic Environments Coordination in Dynamic Environments.

摘要

Recent research and technological advances are harbingers of scenarios where large scale crisis response in civilian and military domains can be facilitated by autonomous sensing and effecting entities. To make this a reality scientific research shortfalls must be addressed, notably the challenge of scalability. Current multi-agent coordination techniques do not scale to large numbers of agents Therefore, the goals of this research are (1) to discover ways in which large numbers of autonomously computing entities (including software agents and physical robots) must be coordinated in effective ways, and (2) ways for human operators to supervise and control robotic agents In particular, we want to develop techniques that allow large scale coordination in the execution of complex missions in uncertain and dynamically changing environments To achieve this goal a variety of key research issues will be addressed. Specifically, we will develop (a) tractable automated coordination algorithms, (b) efficient techniques for fusing information, and (c) techniques for allowing effective and flexible control of agent groups by small numbers of human operators.