NASA utilizes Discovery Machine to establish priorities for technology investments.

Easy interface allows experts to simulate and monitor processes as they execute and then make corrections—all without the need for programming.

The U.S. and the world have come to depend more and more upon satellites for communications, geo-positioning and weather data, for example. And meanwhile, exploratory missions to the moon and Mars are not far off. So it is natural that containing the costs of launch systems continues as a priority for the U. S. National Aeronautics and Space Administration (NASA).

The Challenge

Projecting the cost of future launch systems, and their unique payloads, demands highly complex calculations. These are the responsibility of a cadre of experts whose hard-won knowledge is born of long experience. But today’s exponential growth in data—concerning spacecraft, launch systems, payload handling and aeronautics—is almost overwhelming. As funding pressures mount, the accuracy and timeliness of cost estimates is of paramount concern.

The Key Solution

There are a multitude of considerations concerning payloads; many are automated, low-gravity science experiments that must be configured in specific ways. Each payload must be processed before loading, and this processing has an associated cost.

NASA payload experts captured their own strategies for evaluating payload-processing costs using the Discovery Machine Modeler. Its easy-to-use graphical interface allowed the experts to simulate and monitor each process as it executed. Needed changes were identified and corrected quickly and easily without having to re-code.

The DM Modeler utilizes artificial intelligence principles in a novel way, permitting a subject matter expert to directly represent control processes as knowledge that can be interactively stored, shared, reused, reconfigured and edited.

The DM Modeler achieves this by providing a graphical interface to two computing languages for capturing knowledge. The Task-Method-Knowledge (TMK) language encodes strategic, control, or processing knowledge. The graphical front-end to these languages allows non-programmers to easily use them. This capability is immensely useful to process engineers and operators.

The Result

Allowing experts themselves to harness computing power for their specific needs without writing code saves enormous time and effort. The resulting intelligent system helped payload engineers determine resource allocations for new payload-handling technology. Moreover, the intelligent system could be extended, updated and reused without requiring the original expert’s continual vigilance.

Notably, this work was transitioned to the Defense Advanced Research Projects Agency (DARPA), where the project included simulation-based design (SBD) under the Future Warfighting Concepts (WARCON) program with Newport News Shipbuilding.