Autonomic Associate for Space Exploration

Greater emphasis at NASA is being given to robotic exploration. The sophistication of robots is increasing while their cost is decreasing in comparison to human controllers. Because of this, not only are robots and spacecraft being given more autonomy, but concepts for multiple robots and spacecraft to work together to perform exploration are being developed. Autonomy alone though, absent autonomicity, will leave spacecraft and other systems vulnerable to the harsh environment in which they work and most likely performance will degrade, the system destroyed or it will not be able to recover from faults. We are working on the development of collaborative autonomic intelligent associates to support the needs of manned and unmanned systems for exploration and other applications. Associate systems are useful in domains with large volumes of data, complex array of actions, critical response times, and limited number of experts available, all of which apply to space exploration missions. This paper describes a high level collaborative associate architecture and gives an overview of the use of concept graphs for modeling, plan goal graphs for planning, and context situated scripts for execution of collaborative associates.