What is it about?
Deep space missions have limited contact with ground operations teams. It is difficult to direct a spacecraft when we aren't in contact with it, unless the spacecraft can infer details of its surroundings itself. Computer vision models based on machine learning have the potential to perform this inference. But, traditional spaceborne computers do not have enough compute power to run state-of-the-art models. Thus, we benchmark various computer vision models on two new processors optimized for machine-learned models. These processors are deployed onboard the International Space Station. This is a step towards running such models onboard a Mars Rover or satellite.
Featured Image
Read the Original
This page is a summary of: Benchmarking Deep Learning Models on Myriad and Snapdragon Processors for Space Applications, Journal of Aerospace Information Systems, September 2023, American Institute of Aeronautics and Astronautics (AIAA),
DOI: 10.2514/1.i011216.
You can read the full text:
Contributors
The following have contributed to this page







