UWE Bristol academic unveils breakthrough in energy-efficient AI at NATO science forum

Dr Jonathan Lancelot, senior lecturer in cyber security at UWE Bristol, has developed a new form of artificial intelligence (AI) that could transform how intelligent machines operate in space, defence, and remote environments where power and computing resources are limited.

His research explores how the next generation of AI systems can think and adapt more efficiently, paving the way for future robots, satellites and autonomous systems that are both smarter and more energy efficient.

In future, this kind of AI could help everything from drones to space probes last longer and make smarter decisions without relying on large data centres or constant communication with Earth. It could also lead to more sustainable smart devices and reduce the environmental footprint of AI, which currently consumes vast amounts of energy worldwide.

Dr Lancelot was invited to present his groundbreaking research at a NATO Science and Technology Organisation (STO) forum – an international platform that brings together scientists, engineers and defence experts from across allied nations. The event was led by the UK’s Defence Science and Technology Laboratory (Dstl) and the NATO alliance, and delivered in partnership with the UK Space Agency.

His research introduces a new binary neural network (BNN) framework – an approach that replaces complex mathematical calculations with simple logical operations. This makes AI faster, more efficient, and capable of running on minimal computing resources without losing performance.

“Traditional AI models depend on heavy, power-hungry computations,” said Dr Lancelot. “This approach proves that intelligent behaviour doesn’t have to come at excessive cost. By combining binary logic with an adaptive learning layer, we can achieve high precision while dramatically cutting energy use.”

The innovation lies in a technique called dynamic fractional stride-based feature capture, which helps the system retain the accuracy usually lost when simplifying AI models into binary form. This means the network can still ‘pinpoint’ by making accurate decisions – even with reduced computational resources.

The result is a breakthrough in energy-efficient AI, with significant implications for aerospace, defence, and other sectors where reliability and endurance are critical.

For example, autonomous satellite operations need smart optical components that have the ability to seek accurately without the loss of detail and succeed in preserving onboard power. The implications for collective defence are significant, as satellites will have the capability to map contested areas in real-time, improve hazard detection and detect attacks against critical GPS or weather satellite constellations in space. 

“Speaking at NATO was a true honour and a chance to share how research at UWE Bristol is helping shape the future of collective security and efficient AI,” said Dr Lancelot. “Our aim is to build intelligent systems that can think and adapt even in the most challenging conditions.”

The NATO forum also highlighted the growing convergence of space and cyber-physical operations. Alongside traditional concerns about protecting satellite constellations, participants examined the threat posed by space weather events that can disrupt the Earth’s critical communication and navigation infrastructure.

“Research at UWE Bristol has the potential to make a significant contribution to collective and UK defence in this area, particularly through the development of adaptive AI technologies capable of supporting resilient, autonomous systems in complex environments,” added Dr Lancelot.

The NATO event gathered experts to discuss emerging advances in cyber security, artificial intelligence and space operations. Dr Lancelot’s contribution highlights UWE Bristol’s growing engagement in international research collaborations that link academic innovation with practical, real-world impact.