Part 1: Introduction to Spacecraft Autonomy
This first paper of a two-part series introduces current research on a new spacecraft autonomy architecture, named Remote Agent, which applies artificial intelligence techniques to spacecraft control functions. The necessity for autonomy technology is nowhere greater than in deep space missions. Remote Agent is a model-based autonomous system being developed in a collaborative effort between the National Aeronautics and Space Administration's Jet Propulsion Laboratory and Ames Research Center. This spacecraft control scheme is composed of three parts that work together: 1) the Planner-Scheduler, 2) Mode Identification and Recovery, and 3) the Smart Executive. Models of the spacecraft's components and environment are given to Remote Agent and it figures out the necessary detailed operating procedures on its own. Because autonomy improves mission performance and reduces costs, autonomously controlled spacecraft will become more prominent in future space missions.
The National Aeronautics and Space Administration (NASA) is being confronted with the challenge to perform more frequent and intensive space-exploration missions at greatly reduced cost. Nowhere is this more crucial than among robotic planetary exploration missions that the Jet Propulsion Laboratory (JPL) conducts for NASA. This paper describes current research on a new spacecraft autonomy architecture using artificial intelligence (AI) techniques to challenging computational problems in planning and scheduling, and real-time monitoring and control. Work on this software package combines research from diverse areas of AI such as model-based reasoning, qualitative reasoning, planning and scheduling, and model-based learning, to name a few. NASA views spacecraft autonomy as a major design goal and autonomous spacecraft mission planning and control will become more prominent. Autonomy improves mission performance and relieves ground control operational costs.1
This paper is the first of a two-part series and introduces Remote Agent, a spacecraft software architecture utilizing artificial intelligence and currently being developed by NASA. The second paper describes Remote Agent with more detail of its components and their specific interactions. Before describing what Remote Agent is and who is involved in its development, first presented is a discussion on background information regarding a new generation of control architecture, a system organizational structure, and mission requirements. A perfect place to start is with an explanation of the concept of autonomy.
Autonomous, by definition, is the independence or freedom of one's actions; self-governing. Theoretically, this implies consciousness, logical reasoning, and self-adaption. We could logically define a robot, or automaton, that could be placed in an environment that evolves according to both deterministic and random processes. 2...