Session
Session IX: Advanced Technologies I
Location
Utah State University, Logan, UT
Abstract
Edge Node is a multi-small satellite free-flying, collaborative testbed for formation flight and rendezvous and proximity operations (RPO) under development by The Aerospace Corporation that is planned for launch in 2026- 2027. Edge Node will develop and advance local space situational awareness sensors for CubeSat-scale platforms and facilitate the development and qualification of autonomous on-board RPO software for CubeSats. Edge Node is intended to serve as a multi-user on-orbit facility through which technologies and algorithms can be tested and validated in the authentic operational low Earth orbital environment. Edge Node leverages recent advances in miniature sensor and computing hardware for terrestrial applications (e.g., NVIDIA Jetson TX2 NX, automotive lidar and radar sensors, microbolometer LWIR camera, and polarization image sensors, among others). The mission will enable close-in observations including pose estimation, feature identification, and dynamic motion model characterization. Edge Node is specifically designed to test the limits of autonomous decision-making to support future missions where control with a ground operator in the loop is not possible. Edge Node builds on the prior success of RPO test facilities (e.g., SPHERES and AstroBee) that were confined to operate within the International Space Station (ISS). By operating independently in low-Earth orbit, experimenters will face realistic lighting, complex background scenes, and orbital dynamics that cannot be accurately replicated via terrestrial testbeds or within the confines of the ISS.
The Edge Node mission’s computing platform, consisting of a cluster of NVIDIA Jetson TX2 NX modules, builds on a foundation of three prior qualification missions in low Earth orbit in addition to proton 50 MeV radiation testing of the target hardware. A radiation-tolerant 32-bit ARM Cortex-M7 processor oversees the TX2 NX cluster and interfaces with the rest of the spacecraft avionics. Dual terabyte NVMe drives running a ZFS filesystem provide enhanced reliability on top of radiation screened industrial NAND Flash storage drives.
Edge Node will utilize a custom Linux environment utilizing Docker to containerize workloads and provide access to hardware including 256 NVIDIA CUDA GPU cores per TX2 NX for acceleration of workloads. The Robot Operating System (ROS2) will be utilized as an infrastructure layer for both intra-satellite and inter-satellite Remote Procedure Calls and data transport, offering useful communication patterns and will be the primary API for hosted applications from experimenters. An on-board autonomy watchdog ensures safety of flight while enabling advanced algorithm demonstrations (e.g., artificial intelligence / machine learning (AI/ML)).
An earlier risk reduction payload, Edge Node Lite, comprising a subset of the sensing and computing hardware will launch as a hosted payload in early 2025. This risk reduction demonstration mission includes multiple advanced image-processing and machine-learning algorithm demonstrations provided by multiple mission partners.
Edge Node: A Multi-User Rendezvous and Proximity Operations On-Orbit Testbed
Utah State University, Logan, UT
Edge Node is a multi-small satellite free-flying, collaborative testbed for formation flight and rendezvous and proximity operations (RPO) under development by The Aerospace Corporation that is planned for launch in 2026- 2027. Edge Node will develop and advance local space situational awareness sensors for CubeSat-scale platforms and facilitate the development and qualification of autonomous on-board RPO software for CubeSats. Edge Node is intended to serve as a multi-user on-orbit facility through which technologies and algorithms can be tested and validated in the authentic operational low Earth orbital environment. Edge Node leverages recent advances in miniature sensor and computing hardware for terrestrial applications (e.g., NVIDIA Jetson TX2 NX, automotive lidar and radar sensors, microbolometer LWIR camera, and polarization image sensors, among others). The mission will enable close-in observations including pose estimation, feature identification, and dynamic motion model characterization. Edge Node is specifically designed to test the limits of autonomous decision-making to support future missions where control with a ground operator in the loop is not possible. Edge Node builds on the prior success of RPO test facilities (e.g., SPHERES and AstroBee) that were confined to operate within the International Space Station (ISS). By operating independently in low-Earth orbit, experimenters will face realistic lighting, complex background scenes, and orbital dynamics that cannot be accurately replicated via terrestrial testbeds or within the confines of the ISS.
The Edge Node mission’s computing platform, consisting of a cluster of NVIDIA Jetson TX2 NX modules, builds on a foundation of three prior qualification missions in low Earth orbit in addition to proton 50 MeV radiation testing of the target hardware. A radiation-tolerant 32-bit ARM Cortex-M7 processor oversees the TX2 NX cluster and interfaces with the rest of the spacecraft avionics. Dual terabyte NVMe drives running a ZFS filesystem provide enhanced reliability on top of radiation screened industrial NAND Flash storage drives.
Edge Node will utilize a custom Linux environment utilizing Docker to containerize workloads and provide access to hardware including 256 NVIDIA CUDA GPU cores per TX2 NX for acceleration of workloads. The Robot Operating System (ROS2) will be utilized as an infrastructure layer for both intra-satellite and inter-satellite Remote Procedure Calls and data transport, offering useful communication patterns and will be the primary API for hosted applications from experimenters. An on-board autonomy watchdog ensures safety of flight while enabling advanced algorithm demonstrations (e.g., artificial intelligence / machine learning (AI/ML)).
An earlier risk reduction payload, Edge Node Lite, comprising a subset of the sensing and computing hardware will launch as a hosted payload in early 2025. This risk reduction demonstration mission includes multiple advanced image-processing and machine-learning algorithm demonstrations provided by multiple mission partners.
