NASA Logo, National Aeronautics and Space Administration

+NASA Home

+Ames Home

Autonomous Systems and Operations Project Delivers Machine Learning-Based Display Validation Technology for Orion
Intelligent Systems Division Banner

Autonomous Systems and Operations Project Delivers Machine Learning-Based Display Validation Technology for Orion

The Advanced Exploration System- (AES) funded Autonomous Systems and Operations (ASO) project delivered a machine learning-based Orion spacecraft crew display validation solution to Johnson Space Center’s (JSC) Rapid Prototyping Laboratory (RPL). RPL is responsible for ensuring that Orion’s crew displays, as implemented by Lockheed Martin, meet a wide variety of layout, color, font and other requirements. RPL invests a significant amount of resources in generating and reviewing test artifacts (display reference images) to evaluate the contractor’s displays. The ASO project developed, tested, and delivered a machine learning-based classifier to accelerate the display comparison test process.

The classifier takes as input two display images and classifies them as semantically identical or different. Given machine learning approaches such as deep learning demand, and that the huge amount of data in order to learn higher-level features and the number of available pairs were limited; a feature engineering approach was used as a way of incorporating domain knowledge and reducing the classifier complexity. The engineered features used for this project are related to the characteristics of the pixels in Red, Green, Blue (RGB) channels and designed so that they can separate identical images (that are only different in terms of brightness) from those that are semantically different. A Support Vector Machine (SVM) model trained on these features is very accurate and able to classify 99.9% of pairs correctly. This novel machine-learning solution will allow existing Orion display artifacts to be applicable to a larger scope of test environments. This both reduces the risk associated with performing preliminary testing on a Windows-specific platform, and improves efficiency of RPL efforts by raising the value returned from these artifacts.

BACKGROUND: For over 50 years NASA's crewed missions have been confined to the Earth-Moon system, where speed-of-light communications delays between crew and ground are practically nonexistent. This ground-centered mode of operations, with a large, ground-based support team, is not sustainable for NASA’s future human exploration missions to Mars and other more distant locations in the solar system. Future astronauts will need smarter tools employing Artificial Intelligence (AI) techniques to make decisions without inefficient communication back and forth with ground-based mission control. The Autonomous Systems and Operations (ASO) project develops advanced technology for use on Orion to enable Exploration missions as part of the Artemis program.

NASA PROGRAM FUNDING: Advanced Exploration Systems (AES) Program, Human Exploration Operations Mission Directorate (HEOMD)

TEAM: ARC: Jeremy Frank, Kayla Patterson (Intern), Hamed Valizadegan, and Daniel Weckler; JSC: Jeffery Fox, Lee Morin, and Paul Schauppner

POINT OF CONTACT: Jeremy Frank, jeremy.d.frank@nasa.gov

First Gov logo
NASA Logo - nasa.gov