Back to journal
Introducing the ICE Cubes AI-Box, one of our in-house developments, meant to support the performance of a wide range of activities onboard the ISS. The use of such assets will foster the emergence and development of solutions and products for the benefit of current activities on the station as well as in support of future commercial habitats, in LEO and beyond.
With the AI-Box, Artificial Intelligence / Machine Learning (AI/ML) related tasks could be performed directly on board the ISS, offering the possibility to upload and run AI (neural networks) previously trained on ground or to train AI models directly on board.
This small but powerful asset can support S/W demo & validation, Big Science Data processing, in-situ analysis, embedded and edge computing, image classification, object detection, segmentation and speech processing. It can provide the ideal playground for testing crew applications, in assistance and health monitoring, as well as fault detection, analysis and prediction algorithms. The AI-Box can connect with other payloads, internal or external to the ICE Cubes Facility, and can be monitored and commanded in real-time.
Its architecture is based on the NVIDIA Jetson XAVIER NX computer (21 TOPs) and can run modern neural networks in parallel and process data from multiple high-resolution sensors, opening door for embedded and edge computing devices that demand increased performance, but are constrained by size, weight and power budgets.
Moreover, it was designed with a suite of built-in sensors that will generate data to complement onboard activities or to be sent to ground; these sensors can provide readings on relative humidity, atmospheric pressure, temperature, acceleration, rotation, magnetic field, CO2 concentration and particulate matter. The Cube also comes with Advacam’s MiniPIX TPX3 sensor, a miniaturised and low power radiation camera.
Our AI-Box was launched in December 2021 to the station and will be permanently hosted inside the ICE Cubes Facility to be offered on loan to users.
Contact us to get more information about our capabilities and how we could support your research activities and goals.
21 TOPS
NVIDIA Jetson Xavier NX
