Space Codesign Systems Inc.

Published purpose

(20STDPL35) The objective of this project is to increase the functionality of our HW/SW codesign methodology by targeting high performance on-board processing platforms that are made of CPU-FPGA architectures and target artificial intelligence applications in space, more specifically based on Deep Neural Network (DNN). We will target for Spacecraft Pose Estimation to demonstrate the potential of these processing platforms. We will concentrate on the inference part of deep learning that can be implemented as a payload application on the satellite. This project aims to develop new techniques allowing the optimal use of the on¬board processing resources offered by CPU-FPGA platforms, while reducing the effort and the time to market. The project is developing technologies in several directions that benefits many Canadian companies.