Honorable Mention: 1999 NASA Software of the Year Competition
A number of successful applications of automated planning and scheduling of spacecraft operations have recently been reported in the literature. However, these applications have been one-of-a-kind applications that required a substantial amount of development effort. The Artificial Intelligence Group at JPL has been working on a system called ASPEN (Automated Scheduling and Planning ENvironment). Based on AI techniques, ASPEN is a modular, reconfigurable application framework which is capable of supporting a wide variety of planning and scheduling applications. ASPEN provides a set of reusable software components that implement the elements commonly found in complex planning/scheduling systems, including: an expressive modeling language, a resource management system, a temporal reasoning system, and a graphical interface.
- Planning and Scheduling
- Heuristic Search
- Iterative Repair
- Temporal Reasoning
The basic problem is to develop a sequence of commands for a system that achieves the objectives of the user of that system. The user (e.g., scientist) has some high-level objectives, or goals. Typically, the system (e.g., spacecraft) has a low-level command interface. Therefore, the problem becomes translating the high-level goals into a valid sequence of low-level commands. At JPL, some of the primary systems that require commanding are: deep-space probes, planetary rovers, and deep-space communication antennae.
Automated planning/scheduling technologies have great promise in reducing operations cost and increasing the autonomy of aerospace systems. By automating the sequence generation process and by encapsulating the operation specific knowledge, we hope to allow spacecraft commanding by non-operations personnel, hence allowing significant reductions in mission operations workforce with the eventual goal of allowing direct user commanding (e.g., commanding by scientists).
- ASPEN is currently available for external licensing but not export.
- Current and future work includes integrating repair planning with execution. Here, the idea is to continuously replan around updated information coming from execution monitoring. The CASPER project illustrates this concept.
ASPEN is a modular, re-configurable application framework based on Artificial Intelligence techniques, which is capable of supporting a variety of planning and scheduling applications including spacecraft operations planning, planning for mission design, surface rover planning, ground antenna utilization planning, and coordinated multiple rover planning.
As a ground based system, ASPEN uses an internal spacecraft model and set of high level goals to output a sequence of commands to be executed by the spacecraft to achieve those goals. As a flight-based system, ASPEN receives updates on spacecraft or rover state continuously and updates the current plan to reflect environment changes. As an antenna scheduling system, ASPEN has been used to autonomously control a DSN station.
ASPEN contains several innovations that are not available in other planning and scheduling systems in use today. Among those, the following are the most significant:
- Easy to use modeling language - The ASPEN modeling language was designed by non-AI experts to be easy to use. The system does not require any user knowledge in the areas of computer programming, planning, or scheduling. At the same time, the language is flexible enough to support the complex needs of planning multiple spacecraft and resources.
- Re-configurable framework - ASPEN contains a generic architecture that allows the user to choose among several different search engines and propagation algorithms to optimize the planning process.
- Scalable autonomy - ASPEN contains an iterative repair search algorithm that enables the user to interact with the schedule and replan quickly and efficiently.
- Real-time replanning and response - ASPEN allows replanning during plan execution. This feature enables continuous real-time planning for on-board spacecraft and other embedded applications.
- Plan optimization - The plans that ASPEN generates can be optimized for a specific set of goals such as maximizing science data or minimizing power consumption. The optimization goals can be easily and succinctly specified within the modeling language.
These innovations are documented and detailed in the 29 peer-reviewed publications 6 NASA Technology Briefs, 5 NASA Software Awards, and 4 JPL NOVA Technology Awards related to ASPEN technology. Additionally, patent status is pending on several of the Technology Briefs.
- Aspen User's Guide (internal only) describes the ASPEN system architecture components, along with specific syntax and examples of ASPEN inputs. The updated user guide is now only available by license or specific arrangement: please see contacts below.
- Source Documentation (internal only) provides detailed source-level documentation of the ASPEN API.
- Distributed S/C
- Citizen Explorer
- Planetary Rover Operations
- Distributed Rovers
- New Millennium Earth Orbiting 1 (EO1)
- Highly Reusable Space Transportation
- Modified Antarctic Mapping Mission (MAMM)
|ASPEN Technical Contact:||
Gregg.Rabideau at jpl.nasa.gov
|JPL Technical Contact:||
Dr. Steve Chien|
Steve.Chien at jpl.nasa.gov
SponsorsNASA Code S
Autonomy program, Dave Atkinson (JPL) managing.
Also Sponsored By:
Directors Research Discretionary Fund
Jet Propulsion Laboratory
California Institute of Technology