RA (Remote Agent) is a cognitive architecture - the first artificial intelligence (AI) system to fly onboard and control a deep space probe (on the Deep Space One mission), designed as a part of autonomous control software for NASA's deep space missions, by the Remote Agent Team, including Barney Pell, at the NASA Ames Research Center, Moffett Field, California, USA, the Research Institute for Advanced Computer Science (RIACS), Mountain View, CA, USA, as well as Cal Tech's Jet Propulsion Laboratory (JPL).
Когнитивная архитектура RA разработана исследовательской группой в Исследовательском центре НАСА им. Дж. Эймса (NASA Ames Research Center), г. Маунтин-Вью, Калифорния, США и в Институте исследований перспективной вычислительной техники (Research Institute for Advanced Computer Science - RIACS).
Одним из лидеров этой группы был Барни Пелл (Barney Pell).
Архитектура Remote Agent [Pell et al., 1997] была разработана для управления автономным, ориентированным на задания космическим кораблем.
Долгосрочные структуры включают в себя цели задания, возможные действия и ограничения на их исполнение, а также качественные модели компонентов космического корабля, тогда как динамические структуры включают в себя планы действий, которым надо следовать, графики их выполнения и выводы о режимах работы или сбоев.
Эта архитектура включает в себя процессы, которые производят поиск целей высокого уровня, генерируют планы и графики их достижения, исполняют эти графики путем вызова команд низкого уровня, контролируют режимы работы каждого компонента космического корабля и восстанавливают работу в случае сбоя.
From 1993-1998, Dr. Pell worked as a Principal Investigator and Senior Computer Scientist at NASA Ames, where he conducted advanced research and development of autonomous control software for NASA's deep space missions.
Dr. Pell was the Architect for the Remote Agent and the Project Lead for the Executive component of the Remote Agent Experiment (RAX), the first intelligent executive to fly onboard and control a spacecraft (the Deep Space One mission).
Remote Agent is widely considered one of the top achievements in the history of Artificial Intelligence and was awarded NASA's "software of the year" award in 1999.Dr. Pell was also Co-Lead for the Autonomy Integrated Product Development Team for NASA New Millennium Program, responsible for planning and managing technology maturation and demonstration of autonomous systems technology for future use by NASA.
Agent-Based Simulation and Systems (RIACS)
As humans venture out into the cosmos, mission goals will become more ambitious, requiring a complex supporting cast of robots and computer software agents. These astronaut assistants will be an integral part of mission operations and will perform a wide variety of tasks in order to support astronauts during day-to-day activities.
While it is accepted that robots and software agents will play a central role in long-duration space missions, defining what they will do, how they will do it, and how these assistants will work collaboratively with astronauts to complete tasks is an entirely separate challenge.
Research in the area of agent-based simulation and systems seeks to attain a better understanding of work practices and provide responsive computational support for the collective activities of people, robots and software agents collaborating to accomplish a goal. Whether on the moon, on Mars, or in deep space, humans will rely heavily on robots and software agents to enhance their research capabilities. By performing many of the rudimentary but necessary maintenance, logging and tracking tasks, robots and software agents will free up humans to focus on higher level decision-making.
Agent-based simulation and systems research is focused on three main areas. The first is the creation of new software tools and systems that allow for the modeling of work practices. The second is identifying where software agents could be created to make work practice more efficient by simulating and analyzing current and future work practice. The third is designing software agents to control the actions of machines in support of human collaborators.
At NASA, agent-based modeling, simulation and analysis is being used in a variety of domains, including the International Space Station and simulations of a future Martian habitat. While the focus of this research is to prepare for long-duration space missions, it will also advance human-machine collaborations on Earth. Agent-based simulation and systems can lead to greatly improved efficiency in any environment, whether in an office dealing with human-to-human interactions or a complex, automated environment that involves humans and machines.
The Remote Agent Executive: Capabilities to Support Integrated Robotic Agents.
Barney Pell, Research Institute for Advanced Computer Science (U.S.), Gregory A. Dorais, Christian Plaunt, Richard Washington. NASA Ames Research Center, Research Institute for Advanced Computer Science, 1998
The Remote Agent (RA) integrates a broad spectrum of robotic activities, including planning, scheduling, execution, monitoring, failure detection, diagnosis, and recovery.
The RA Executive (EXEC) can be viewed as the core of the agent. EXEC enables software developers to think about the robot at a higher level; it also supports the reuse of knowledge and code across multiple robot applications.
EXEC's capabilities include a high-level procedural action-definition language, services for resource management and configuration management, support for executing flexible closed-loop plans and command sequences, support for replanning, and a framework for specifying fault responses including safe modes and responses to plan failures. We believe that these capabilities are required for most autonomous agents, and that executives and execution capabilities will become more essential as we attempt to develop autonomous agents of increasing capability.
Moreover, we deem modularity, variable autonomy, robustness, and support for tools and testing to be general architectural properties necessary for design, deployment, and operation of executives within a complex mission.
To this end, we are remodularizing the executive by providing each capability as a separate component so that the individual capabilities can be standardized and shared across different agent architectures.
A recognized expert on Autonomous Agents and Human/Agent Interaction, Dr. Pell has published over 30 technical papers on topics related to information retrieval, knowledge management, machine learning, artificial intelligence, and scheduling systems.
Publications of Barney Pell
Pell, B., Bernard, D. E., Chien, S. A., Gat, E., Muscettola, N., & Nayak, P. P. (1997). An autonomous spacecraft agent prototype. In Proceedings of the first international conference on autonomous agents (pp. 253–261). Marina del Rey, CA: ACM Press.
Remote Agent: To Boldly Go Where No AI System Has Gone Before