Open in App
  • Local
  • U.S.
  • Election
  • Politics
  • Crime
  • Sports
  • Lifestyle
  • Education
  • Real Estate
  • Newsletter
  • Interesting Engineering

    Scientists to ‘shrink’ AI algorithm for smarter, powerful spacecraft

    By Mrigakshi Dixit,

    7 hours ago

    https://img.particlenews.com/image.php?url=001ZJX_0ubuWX0z00

    Scientists at the University of Leicester are working to make future spacecraft smarter than ever before.

    With a £690,000 grant, the team is developing a new method to shrink artificial intelligence algorithms. These advanced tools are currently too huge for the spacecraft’s limited computer space.

    Algorithms are the computational backbone that supports spacecraft operations. From guiding spacecraft to processing immense amounts of data, algorithms could help execute a wide variety of tasks for smooth operations between a distant spacecraft and its ground controllers.

    “Our novel approach to reduce algorithm size considerably while maintaining high accuracy performance provides a disruptive enabling technology poised to unlock a wide range of real-time services from space that previously would not have been possible due to their computational complexity,” said Tanya Vladimirova, the principal investigator.

    New AI algorithm models

    Existing AI algorithms are not suited for spaceflight owing to their size and power consumption. Interestingly, the research aims to create tiny AI models that can operate within spacecraft restrictions, providing additional capabilities.

    A collaborative effort between the university’s computing, physics, and innovation park departments is driving the development.

    This project is known as REALM, which is “Rapid information extraction for environmental remote sensing onboard spacecraft through the application of light Machine Learning models in payload computing systems.”

    They are developing a novel approach called sparse-split-parallelism (SSP) to substantially reduce the size of AI models — without sacrificing performance.  According to the press release, the SSP technique has the potential to lower the size of complicated AI models used to analyze satellite data by at least 45 percent.

    https://img.particlenews.com/image.php?url=1Xx4RL_0ubuWX0z00
    A customised DJI Matrice 300 RTX drone with an AI-enabled multipurpose high-performance computing module, developed by the team as a proof-of-concept. University of Leicester

    Testing on drones

    Once the algorithm has been made, the experts will test it on the ground. The algorithm will be tested on a small, space-grade GPU before being integrated into a drone equipped with multiple payloads. These flying labs will help them refine the technology before it blasts off into space.

    The future of spacecraft will be highly reliant on advanced algorithms for operations. From monitoring climate change to exploring distant planets, these smaller, smarter AI-powered spacecraft will boost challenging space missions.

    This ambitious project is part of a larger £33 million investment by the UK government to boost the nation’s space industry.

    As per the press release , REALM is one of more than 20 national space initiatives set to be unveiled by DSIT Secretary of State Peter Kyle on the first day of the Farnborough International Airshow.

    “These new projects will help kickstart growth, create more high-quality jobs, protect our planet and preserve the space environment for future generations. They go to the heart of what we want to achieve as a national space agency that supports cutting-edge innovation, spreads opportunity across the UK and delivers the benefits of space back to citizens on Earth,” said Dr Paul Bate, CEO of the UK Space Agency.

    The £33 million fund will support a wide range of space initiatives, including in-orbit service, innovative material development, and satellite imagery.

    Expand All
    Comments / 0
    Add a Comment
    YOU MAY ALSO LIKE
    Most Popular newsMost Popular

    Comments / 0