Researchers at Loughborough University’s Wolfson School of Mechanical, Electrical, and Manufacturing Engineering, in collaboration with the University of Southampton, are making significant strides in improving satellite communication through advanced machine…
Researchers at Loughborough University’s Wolfson School of Mechanical, Electrical, and Manufacturing Engineering, in collaboration with the University of Southampton, are making significant strides in improving satellite communication through advanced machine learning techniques.
The research addresses a major challenge in satellite downlink transmission: the problem of delayed Channel State Information (CSI) due to long space-ground propagation delays. Traditionally, this delay has made it difficult to accurately optimise transmission strategies. To overcome this, the team has developed a novel Deep Reinforcement Learning (DRL)-based approach that dynamically optimises Transmit Precoding (TPC) matrices. By exploiting time-domain correlations in the channel data, the DRL method constructs accurate TPC matrices, maximising data rates while also reducing system complexity compared to conventional methods.
This advancement is particularly impactful for Low Earth Orbit (LEO) satellites serving handheld terrestrial devices. By compensating for delayed CSI more effectively, the DRL-based framework significantly boosts data transmission performance, providing improved global connectivity, especially for remote and underserved regions. Applications of this research include better support for emergency response systems, expanded access to remote education services, and enhanced Iot deployments in areas where conventional networks struggle to reach.
One of the key strengths of the DRL framework is its adaptability. It can respond to dynamic channel conditions and satellite handovers more efficiently than traditional prediction-based methods. The energy-efficient design also promises reduced operational costs, which is vital for large-scale satellite constellations.
Looking ahead, the team, led by Dr Yasaman Omid, plans to extend the DRL framework to more complex multi-satellite collaborative scenarios and to explore integration with air vehicles, paving the way for future space-air-ground integrated networks (SAGIN). Additionally, they aim to investigate new energy-efficient algorithms to further minimise computational complexity, ensuring that the next generation of satellite communication is not only faster and more reliable but also more sustainable.
The research is a part of TITAN Lighthouse Project 5 (LP5): Non-Terrestrial Network Design and Optimisation. The project explores the development and enhancement of communication systems that operate beyond traditional ground-based infrastructure, such as satellite and high-altitude platform networks.
TITAN is part of the Federated Telecoms Hubs (FTH) alongside CHEDDAR, HASC, and JOINER. Together, we conduct world-leading R&D and create a new knowledge-sharing framework, roadmap technology, develop skills, and engage with standards while instilling a commercialisation-focused mindset across academia.
If you have any questions about the research discussed in this article, please contact titan-enquiries@eng.cam.ac.uk.
