Session
Poster Session 2
Location
Salt Palace Convention Center, Salt Lake City, UT
Abstract
SCALES (Spacecraft Compartmentalized Autonomous Learning and Edge-computing System) is a novel hardware/software system for the enhancement of machine learning onboard small spacecraft. Through the first year of development under the NASA University SmallSat Technology Partnerships (USTP) program the team has been able to design and develop the first prototype of the system.
SCALES combines the use of commercial off-the-shelf hardware and the open-source F’ software framework developed by JPL. The system is aimed at coupling radiation tolerant flight computer with the capability of high-performance edge computers. The radiation tolerant flight computer is responsible for the operation of the spacecraft and the monitoring of the edge computer, ensuring that as computationally demanding tasks are sent to the edge computer and risk to the core spacecraft systems is mitigated. Through the expansion of the F’ software framework the system will be able to support the execution and onboard retraining of AI/ML models. The goal is to enable more autonomy onboard small spacecraft by allowing more efficient data processing and system management on orbit.
Document Type
Event
An Updated Architecture for a Modular and Scalable Edge Computing System for Small Spacecraft
Salt Palace Convention Center, Salt Lake City, UT
SCALES (Spacecraft Compartmentalized Autonomous Learning and Edge-computing System) is a novel hardware/software system for the enhancement of machine learning onboard small spacecraft. Through the first year of development under the NASA University SmallSat Technology Partnerships (USTP) program the team has been able to design and develop the first prototype of the system.
SCALES combines the use of commercial off-the-shelf hardware and the open-source F’ software framework developed by JPL. The system is aimed at coupling radiation tolerant flight computer with the capability of high-performance edge computers. The radiation tolerant flight computer is responsible for the operation of the spacecraft and the monitoring of the edge computer, ensuring that as computationally demanding tasks are sent to the edge computer and risk to the core spacecraft systems is mitigated. Through the expansion of the F’ software framework the system will be able to support the execution and onboard retraining of AI/ML models. The goal is to enable more autonomy onboard small spacecraft by allowing more efficient data processing and system management on orbit.
