CTAEMS MDP Optimal Planner is a problem- solving software designed to command a single spacecraft/rover, or a team of spacecraft/ rovers, to perform the best action possible at all times according to an abstract model of the spacecraft/rover and its environment. It also may be useful in solving logistical problems encountered in commercial applications such as shipping and manufacturing.
The planner reasons around uncertainty according to specified probabilities of outcomes using a plan hierarchy to avoid exploring certain kinds of sub- optimal actions. Also, planned actions are calculated as the state-action space is expanded, rather than afterward, to reduce by an order of magnitude the processing time and memory used. The software solves planning problems with actions that can execute concurrently, that have uncertain duration and quality, and that have functional dependencies on others that affect quality. These problems are modeled in a hierarchical planning language called C_TAEMS, a derivative of the TAEMS language for specifying domains for the DARPA Coordinators program.
In realistic environments, actions often have uncertain outcomes and can have complex relationships with other tasks. The planner approaches problems by considering all possible actions that may be taken from any state reachable from a given, initial state, and from within the constraints of a given task hierarchy that specifies what tasks may be performed by which team member.
This program was written by Bradley Clement, Steven Schaffer, and Gregg Rabideau of Caltech for NASA's Jet Propulsion Laboratory.
This software is available for commercial licensing. Please contact Karina Edmonds of the California Institute of Technology at (626) 395-2322. Refer to NPO-44019.
This Brief includes a Technical Support Package (TSP).
Optimal Planning and Problem-Solving
(reference NPO-44019) is currently available for download from the TSP library.
Don't have an account?
Overview
The document is a Technical Support Package from NASA's Jet Propulsion Laboratory (JPL) detailing a software solution for optimal planning and problem-solving, identified as NPO-44019. This software is designed to address complex planning problems where actions can occur concurrently, have uncertain durations and qualities, and exhibit functional dependencies on one another.
The software models these planning challenges as Markov decision processes (MDPs), which allows it to systematically evaluate all possible actions in any state reachable from a given initial state. A key feature of this software is its hierarchical planning language, C_TAEMS, which is a derivative of the TAEMS language used in the DARPA Coordinators program. This hierarchical structure enables the software to avoid exploring certain suboptimal actions, thereby enhancing efficiency.
One of the significant improvements offered by this software is its ability to calculate optimal policies during the expansion of the state-action space, rather than after the expansion is complete. This approach reduces the processing time and memory requirements by an order of magnitude compared to similar software solutions, which typically do not utilize a hierarchical framework.
The document emphasizes the importance of careful planning in realistic environments, particularly when managing teams of spacecraft, robots, or military forces. The software aims to maximize mission success while minimizing expected costs by considering the complex relationships between tasks, such as enabling, disabling, and facilitating actions, along with the probabilities of different outcomes in terms of quality and duration.
Additionally, the software can be applied to individual robots, allowing for optimal action commands to be uploaded to devices like planetary rovers. This capability ensures that the rover can perform the best possible actions at all times based on an abstract model of its environment.
For further inquiries or additional information regarding research and technology in this area, the document provides contact details for the Innovative Technology Assets Management at JPL. It also includes a notice regarding the liability and rights associated with the use of the information contained within the document.
Overall, this Technical Support Package highlights a significant advancement in planning software that has potential applications across various fields, including aerospace, robotics, and military operations.
