Modern space applications impose significant challenges to the design of hardware and software platforms. Beyond traditional applications such as avionics, Attitude Orbit Control, and signal/telemetry processing, new developments increasingly leverage Machine Learning models to enhance the autonomy of spacecraft. Such AI-based functionalities promise significant advantages, but require computing power beyond what can be provided by current on-board platforms. At the same time, the challenge of technological sovereignty requires a move towards open hardware and software. To achieve these objectives, within the KDT ISOLDE project started in 2023, we propose the development of a new family of processors for AI-based applications to be deployed on board of satellites. In this paper, we showcase some examples of space applications with their requirements, and highlight the possible solutions as well as the corresponding work that will be carried out in ISOLDE, and the expected results.
RISC-V Processor Technologies for Aerospace Applications in the ISOLDE Project
Fornaciari, William;Reghenzani, Federico;Agosta, Giovanni;Zoni, Davide;Galimberti, Andrea;
2023-01-01
Abstract
Modern space applications impose significant challenges to the design of hardware and software platforms. Beyond traditional applications such as avionics, Attitude Orbit Control, and signal/telemetry processing, new developments increasingly leverage Machine Learning models to enhance the autonomy of spacecraft. Such AI-based functionalities promise significant advantages, but require computing power beyond what can be provided by current on-board platforms. At the same time, the challenge of technological sovereignty requires a move towards open hardware and software. To achieve these objectives, within the KDT ISOLDE project started in 2023, we propose the development of a new family of processors for AI-based applications to be deployed on board of satellites. In this paper, we showcase some examples of space applications with their requirements, and highlight the possible solutions as well as the corresponding work that will be carried out in ISOLDE, and the expected results.File | Dimensione | Formato | |
---|---|---|---|
SAMOS 2023 paper ISOLDEpdf.pdf
accesso aperto
Descrizione: full paper
:
Publisher’s version
Dimensione
774.89 kB
Formato
Adobe PDF
|
774.89 kB | Adobe PDF | Visualizza/Apri |
I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.