In order to enable autonomous or resilient satellite constellation operation, immediate decisions and threat mitigation must be executed by satellites in real-time before an enemy is successful in executing their strategies. Ground operators are generally unable to respond to proximity threats instantly in current architectures. Applying edge-computing into a satellite enables autonomous rule-sets to be instantly executed in the new contested and congested space environment. However, without adequate installed hardware, rule-set software has no use. A recent survey of self-driving car applications found that a minimum of 16 TBytes of computational memory is necessary to enable autonomous functions; however, no current small satellite in the US space industrial base approaches that data storage capacity. The objective of this project is to perform applied development, transitioning a best-in-class, already-existing Commercial, Off the Shelf (COTS) NGD Systems ultra-high capacity 8 Terabyte (TB) solid state drive (SSD) targeted for autonomous driving applications to SmallSat application. This COTS SSD will enable in-situ small satellite processing capability with (1) Machine Learning (ML) applications on-orbit, and (2) secure and reliable storage and transmission of data, leading to transformational US Space Force superiority. Deployment path(s) will be demonstrated through Space and Missile Systems Center, Los Angeles Air Force Base (SMC) supply chains and proposed missions. Satellite autonomy is vital to USSF; the project will develop a full-fledged software stack for in-storage ML processing in support of deploying an autonomous 2U CubeSat to the TRL 3 level for an autonomy demonstration, leading the way for a next-step trusted autonomy project.
