A report presents updated information about the Remote Agent — a reusable artificial-intelligence software system that was described in “A Remote Agent Prototype for Spacecraft Autonomy” (NPO-19992), NASA Tech Briefs, Vol. 21, No. 3 (March 1997), page 106. This system was conceived to enable spacecraft to operate robustly with minimal human supervision, even in the face of hardware failures or unexpected events. It also is expected to offer similar benefits for communication networks, chemical plants, and other complex systems on Earth.
This work was done by Douglas Bernard; Edward Gamble, Jr.; Benjamin Smith; Nicolas Rouquette; Erann Gat; Yu-Wen Tung; Guy Man; Robert Rasmussen; Steve Chien; Scott Davies; Daniel Dvorak; Greg Rabideau; and David Yan of NASA’s Jet Propulsion Laboratory; P. Pandurang Nayak, Bob Kanefsky, James Kurien, William Millar, Kanna Rajan, Brian Williams, Charles Fry, Gregory Dorais, Barney Pell, Ronald Keesing, Christian Plaunt, William Taylor, Scott Sawyer, Greg Swietek, Greg Whelan, Sunil Mohan, Paul Morris, and Mike Wagner of Ames Research Center; and Reid Simmons of CMU. To obtain a copy of the report, “Remote Agent, autonomous reasoning and control for spacecraft and other complex systems,” access the Technical Support Package (TSP) free on-line at www.nasatech.com/tsp under the Information Sciences category.
This software is available for commercial licensing. Please contact Don Hart of the California Institute of Technology at (818) 393-3425. Refer to NPO-20867.
This Brief includes a Technical Support Package (TSP).
Remote Agent as Applied to the Deep Space 1 Spacecraft.
(reference NPO-20867) is currently available for download from the TSP library.
Don't have an account?
Overview
The document discusses the Remote Agent, a reusable artificial intelligence (AI) software system developed to enhance the autonomy of spacecraft and other complex systems, such as life support systems and chemical plants. The primary goal of the Remote Agent is to enable these systems to operate robustly with minimal human supervision, particularly in the face of hardware failures or unexpected events.
The Remote Agent was tested onboard the Deep Space 1 (DS1) spacecraft in a series of experiments conducted in May 1999. This marked a significant milestone as it was the first AI system to control a spacecraft. The Remote Agent was designed to plan its own mission activities based on goals provided by human operators and to execute these plans by commanding the spacecraft. Notably, it was capable of diagnosing onboard failures and re-planning its mission to ensure that its objectives were met, thereby demonstrating a level of autonomy previously unseen in space missions.
The motivation behind developing the Remote Agent stemmed from the need to increase spacecraft autonomy, reduce development costs, simplify operations, and enable new mission concepts. Traditional spacecraft operations heavily relied on mission controllers on Earth, who would develop detailed sequences of commands for each mission activity. This dependency not only incurred significant costs but also increased the risk of mission loss due to the need for human intervention in critical situations.
The Remote Agent addresses these challenges by automating the analysis and decision-making processes. It can detect and diagnose failures, assess their impact on mission goals, and generate appropriate responses based on a maintained set of models. This capability reduces the amount of pre-mission analysis and software development required, as the system can adapt its behavior in real-time to the conditions it encounters during its mission.
The document also emphasizes the potential applications of the Remote Agent beyond space exploration, suggesting its utility in various complex systems on Earth. The work was carried out by a team from NASA’s Jet Propulsion Laboratory and Ames Research Center, highlighting a collaborative effort in advancing AI technology for autonomous operations. Overall, the Remote Agent represents a significant advancement in spacecraft autonomy, paving the way for more efficient and resilient space missions.
