The UAE-built Rashid rover will test an artificial intelligence (AI)-based machine learning system for the first time on the Moon, where it will be sent in December 2022 along with the Hakuto-R landing module, Space.com reported.
This is the first use of AI beyond Earth and low-Earth orbit.
This AI system, developed by Canada's Mission Control Space Services (MCSS), will use deep learning algorithms to identify geological features of the lunar surface in images taken from the Moon rover, to inform operators of possible hazards, and to assist in determining the Moon rover's route.
MCSS will receive navigational images from the rover through a landing module that will communicate with Earth. MCSS CEO Ewan Reid said that using the Canadian company's deep-learning algorithm, every pixel of the image will be classified as a specific type of terrain. Scientists will use this data to determine the future route of the lunar rover.
Rashid has a lifespan of just one lunar day, 14.5 Earth days. It will reach the Moon in late April 2023. By using AI, scientists will be able to use the lunar rover with maximum efficiency. The AI system will also save a lot of bandwidth, Reid said, and only the data, images and video that scientists really need will be sent back to Earth.
"We see an enormous opportunity for deploying AI on the edge, in space," Reid said. If all goes to plan, the lunar demonstration will allow MCSS "to support other companies and organizations as they work to deploy AI in their missions in the future," Reid said.
Ewan Reid also said that if all goes well, the AI will become an important decision-making tool for the spacecraft, especially for tasks such as finding water on the moon, something NASA also plans to do on the Artemis missions. It could also be used to improve Earth observation.