The recent rapid development of Artificial Intelligence (AI) computing has opened up new opportunities for space exploration and exploitation. However, to leverage the benefits of intelligent processing in the space industry, requirements such as flexibility, radiation tolerance, and reliability must be addressed. We introduce a CGR-AI Engine, a highly parametrized Coarse-Grained Reconfigurable Array (CGRA)-core accelerator designed for space applications. In this paper, we report the results of functionality tests of the CGR-AI Engine which pave the way to implementing the accelerator in UDSM technology. The tests are performed on an FPGA overlay and serve as an intermediate milestone to demonstrate the readiness of the system.
FPGA Prototype of CGR-AI Engine for Space Systems: Step Towards UDSM Implementation
Gabriela Mystkowska;Matteo Monopoli;Pietro Nannipieri;Luca Fanucci;
2025-01-01
Abstract
The recent rapid development of Artificial Intelligence (AI) computing has opened up new opportunities for space exploration and exploitation. However, to leverage the benefits of intelligent processing in the space industry, requirements such as flexibility, radiation tolerance, and reliability must be addressed. We introduce a CGR-AI Engine, a highly parametrized Coarse-Grained Reconfigurable Array (CGRA)-core accelerator designed for space applications. In this paper, we report the results of functionality tests of the CGR-AI Engine which pave the way to implementing the accelerator in UDSM technology. The tests are performed on an FPGA overlay and serve as an intermediate milestone to demonstrate the readiness of the system.I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.


