The project will explore how generative and agentic AI can reduce the bandwidth and power required to deliver high quality content across integrated terrestrial and non-terrestrial networks.
AWTG has been selected to lead Hybrid Generative Broadcasting & Semantic SATCOM for Satellite Content Delivery in Space-Air-Ground Integrated Networks, working with the University of Surrey. The project is being delivered through the European Space Agency’s Advanced Research in Telecommunication Systems programme, with support from the UK Space Agency. It forms part of the UK Government’s wider £13 million investment in 16 satellite communications projects.
Demand for video, digital services and connected applications continues to grow, but the bandwidth and power available to carry that data remain limited, particularly outside the reach of conventional terrestrial infrastructure and lack of network capacity. Existing compression technologies reduce the amount of data transmitted, but growing demand means networks still need more efficient ways to deliver high quality content.
AWTG brings extensive experience across both sides of this challenge. Its connectivity teams design, integrate and optimise converged telecommunications networks and complex multi-vendor environments. Its AI teams develop generative AI, agentic systems, reinforcement learning, edge intelligence, automation and advanced analytics, alongside AI powered platforms that improve how organisations interact with their customers.
The project will explore how generative AI can go further by identifying the essential information within an image or video and transmitting a compact semantic representation of it. Lightweight AI models at the receiving end will then reconstruct the content. If successful, this would allow comparable content quality to be delivered using less bandwidth and power, enabling the same network capacity to support more users, devices and services.
The system will also use agentic AI and multi agent deep reinforcement learning to respond to changing network conditions. An intelligent scheduler will determine how bandwidth and power should be allocated across terrestrial, airborne and satellite links, based on factors such as content priority, reconstruction quality and the capacity available at that moment.
“This project demonstrates how the UK is applying cutting edge artificial intelligence to address one of the fundamental challenges in satellite communications – delivering more data with less bandwidth and power. By enabling high quality content to be transmitted more efficiently across satellite, airborne and terrestrial networks, AWTG’s work has the potential to strengthen connectivity, improve resilience and support the next generation of digital services in even the most challenging environments.”
Hendy Sands, Head of Satellite Communications, UK Space Agency
“The market is moving towards a future in which terrestrial and non-terrestrial networks operate as one. Customers will not care which network is carrying their service. They will care that it remains available, reliable and consistent.
“This project is about making that convergence more intelligent and commercially viable. By combining AWTG’s capabilities in AI and connectivity, we aim to help operators use their existing capacity more effectively, reduce the cost of delivering services and provide a better experience for customers wherever they are.”
Abbey Alidoosti, CEO of AWTG
The project brings these capabilities together by applying AI to the delivery of information across the network itself. This is important because the customer experience does not begin at the application layer. It depends on the quality, availability and resilience of the underlying connection. As the technology moves towards commercialisation, it could help operators manage capacity, control delivery costs and provide a consistent customer experience across terrestrial and non-terrestrial coverage.
Over the course of the project, AWTG and the University of Surrey will develop and validate a laboratory demonstrator, progressing the technology from Technology Readiness Level 2 to Technology Readiness Level 4. Its performance will be assessed against established video compression approaches, including the quality of the reconstructed content and the bandwidth and power required to deliver it.
The project is being developed at a time when the boundary between terrestrial and non-terrestrial networks is beginning to blur. Satellite messaging on Apple devices and the direct to cell services being introduced by Starlink and mobile operators are early examples of satellite connectivity becoming part of mainstream communications.
For customers, this could mean fewer gaps in coverage, fewer interruptions and more consistent service quality. Potential applications include media delivery, connected transport, maritime communications, industrial site monitoring, emergency response and public services in locations that are currently difficult or expensive to reach.
The commercial benefits are also significant. Better compression could allow operators to carry more traffic over their existing infrastructure, reduce the cost of delivering services and delay the need for additional network capacity. Lower power requirements could further reduce operating costs across satellite and edge environments, where energy is constrained.
These improvements could make previously uneconomical services commercially viable and create new revenue opportunities through wider coverage, resilient connectivity and improved service quality. Following validation, the technology could form the basis of a deployable network capability for satellite operators, mobile operators, broadcasters and digital service providers, creating a potential route from research and development into commercial deployment.